element in HTML, or even, the \<Hn> types. Folks rate their page by the hit rate. OK. Topical indexing. An old and honorable technique, but not really a good indicator. For example, if I want a high hit rate on a page dedicated to selling plumbing tools, all I have to do is put \<title>sex sex sex and more hot wet sex</> in the file. A shoddy trick, yes, but not beyond those who don't give a fig about the users. At least some percentage will indeed be browsing for that topic who need to fix the bathroom sink, so the tactic succeeds often enough to be useful. Now, if I have a \<!doctype plumbingparts ... and a \<catalog>sinks</> element in there, given intelligent searching, wouldn't that be both faster, more accurate, and able to filter the stupid pet trick in the first example? Sure, one could still abuse it, but over time, I suspect the semantic links like \<weekend_boring_tasks id='' reftype= hytime=ilink dtd= > etc. get stronger, that is, the domains would sort themselves out particularly where someone is keeping these in their own standardized link server database? Why can't the server or server owner choose their own categories of interest and simply register their availability? Also, one can specify a kind of protocol (term used loosely) of doctype to doctype matches, eg, when I send a \<!doctype fixsink, I always get back a \<!doctype plumbingparts among others. I admit that I am not thinking this out completely because I've got to get out the door to a gig, but something like that could work. If not, why not? I'm willing to learn. Good to hear from you Tim. Tell the Good Doctor Kiwi hello from the deep south. ;-) len </message> <message id="<480r0m$1gao@usenetp1.news.prodigy.com>" date="3025038806" seqno="11368"> From: XVSK06A@prodigy.com (Wayne Vincent) Newsgroups: comp.text.sgml Subject: Re: sgmls and NAMELEN Date: 11 Nov 1995 00:33:26 GMT Organization: Prodigy Services Company 1-800-PRODIGY Message-ID: <480r0m$1gao@usenetp1.news.prodigy.com> References: <47v99s$m1b@relay.lsl.co.uk> Kevan: ISO 8879 defines a "reference concrete syntax" which lists the default parameters used within an SGML DTD/Instance, such as NAMELEN. Your SGML document should start with an SGML declaration where, by default, ELEMENT names are no longer than 8 characters. To change this, alter NAMELEN in your declaration. The declaration will start with: \<!SGML "ISO 8879-1986" - - - .. within the "QUANTITY" set declaration. Hope this is helpful. Wayne </message> <message id="<480qpa$3q6@crl2.crl.com>" date="3025038570" seqno="11370"> From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: [Q] recursive ELEMENT? Date: 10 Nov 1995 16:29:30 -0800 Organization: Mentally ill venomous purists with strange emotional drives Message-ID: <480qpa$3q6@crl2.crl.com> References: \<BERND.95Nov10190143@harlie.nero.uni-bonn.de> Bernd Kreimeier \<bernd@nero.uni-bonn.de> wrote: >On descriptive markup, it seems to me that elements >like chapter, section, subsection are bound to create >a static and absolute hierarchy. For a couple of >reasons, I'm looking for a DTD that introduces a >structure like this: > [...] >composed of some header (identifier, head line, title, whatever in >a more sophisticated definition), and some (unstructured, >as it doesn't matter here) text. >The important thing is the recursive definition of a folder, >as it might contain folders, textmoduls, or nothing, in >an arbitrary sequence. If I made a mistake: it's just the same >with directories containing directories and files. > >Question: is a recursive definition allowed? Yep. Something like: \<!ELEMENT folder - - (header, (textmodul|folder)*> should do what you want. >Why am I asking? My idea is that parts of a text might change >their importance during writing the text (e.g. what started as >a paragraph, turns out to be a chapter, or a chapter is downgraded >to a simple section). I prefer that approach myself, for exactly that reason. You could use: \<!ELEMENT SECTION - - (title, ((%bodyText;)*, section*))> \ to specify the more common convention that no body material appears between subsections: each section starts with a (mandatory) title, followed by (optional) introductory material, and ends with (optional) subsections. Another model that I like is: \<!ELEMENT section - O (title, ((%bodyText;)*, subsecs))> \<!ELEMENT subsecs - - (section+)> SUBSECS is just a ``dummy'' container element that's there to make \</SECTION> end-tags omissible. This is handy since that way you don't have to keep track of matching \<SECTION> ... \</SECTION> pairs. >Question: are there DTD's defining such a "depth of fold" >counter mechanism, and/or formatters relying on this? Hmm... it looks like the osf-book DTD has a recursive structure. So does IBM ID DOC. HTML 3 allows nested \<DIV>s (but those are more of a general-purpose container than a hierarchical sectioning element.) There was a good article in \<TAG> recently (by Dave Peterson?) that talked about these issues. --Joe English jenglish@crl.com </message> <message id="<9511110154.AA06542@fly.hiwaay.net>" date="3025043641" seqno="11373"> Newsgroups: comp.text.sgml Date: Fri, 10 Nov 1995 19:54:01 -0600 Message-ID: <9511110154.AA06542@fly.hiwaay.net> From: Len Bullard \<cbullard@hiwaay.net> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> \<tbray.10.30A3B6E9@opentext.com> Subject: Re: Solving the BIG problem [Tim Bray] | Don't hold your breath. This problem is too hard for CS types; if | they'd been able to come up with a satisfactory solution, they would | have built the Web. Could be. Sad to think so. My opinion and only that since I was involved at the time was that we were focused on other things. Also, we were all engaged in *heavy cooperation*, that is, standardizing to death a CALS system that cost too much to build. One advantage of the WWW builders is that they only had to deal with a "talknet", not a series of long, bloody interminable and highly politicized meetings in hotels all over the planet. The HyTime group, bless 'em, had to work with every gun in ISO aimed at them and the necessity to satisfy an endless set of requirements just to keep the charter sure complicated the original ideas. ah well.... spilt milk. The work is there and after all of the time invested, I think is quite substantial. Facts just to know: In the mid to late eighties, I remember people (me too) suggesting that we use the techniques and the Internet to build what the Web became. The answers we got as I remember them were: 1. The internet is too loose and the bandwidth is too slow for the kind of structures and security needed to do reliable hypermedia. Try again and please come up with a *complete solution*. (So much for limited realistic goals.) 2. SGML is just a passing thing and a real system would be based on more powerful technology that *we* will design for you if you pay us. 3. Who would use it? People don't even bother to use the PCs they have and when they do, they store recipes on them. The corporations don't care because they have fax machines, and BTW, if you do this "enterprise engineering thing" you'll ruin the game for us. 4. The standards committees, particularly PDES, are much more important and they will give us the solutions we need when we need 'em. What makes you think you are smarter than the best experts money can buy? (That one got me. ever the coward.) 5. Quit going off on these Don Quixote adventures, Len. We have to get 'em pregnant first. (I always hated that one....). So we focused on IETMs which at least were a form of hypermedia that would do some deserving kids some good while a wild one at CERN and some college students who didn't know it couldn't or shouldn't be done did it. They got famous and rich. I got to keep on trying to get the committee types to see the fallacy of using gigantic overbuilt one size fits all DTDs. In other words, I did my job as best I could for those who needed it done. No regrets on that. When I got over my jealousy of the WebHeads, I woke up to discover I was very glad they did. :-) I use the thing daily. And I like it. So much for my whining. | I believe that there has to be some kind of compromise between | the URL and the unrealizable Platonic ideal. I'm just not smart enough | to think it up. Me either, but if we want it and won't accept what_is_as_is like some of us did once before, we can get it to happen. I truly think we are about to. The probability here is that this time it will be done by a corporation, not the government or the committees, and they won't be willing to give it away. But if they don't cooperate, who will use it? I'm willing to ride the tired old horse of my ideals one more time with Sancho. As far as I can see right now, the only barrier is getting a system that can use differentiated link types whose semantic is assignable in combination with stylesheets that control semantic properties. Ilinks are intersection records stored as objects. There is pressure to do this for all of the usual database reasons. Here is one. Maybe the smarter eggs among us can refute this. Right now, as Mark Gaither and his associates point out, a lot of search software on the net index pages based on the content of the \<title> element in HTML, or even, the \<Hn> types. Folks rate their page by the hit rate. OK. Topical indexing. An old and honorable technique, but not really a good indicator. For example, if I want a high hit rate on a page dedicated to selling plumbing tools, all I have to do is put \<title>sex sex sex and more hot wet sex</> in the file. A shoddy trick, yes, but not beyond those who don't give a fig about the users. At least some percentage will indeed be browsing for that topic who need to fix the bathroom sink, so the tactic succeeds often enough to be useful. Now, if I have a \<!doctype plumbingparts ... and a \<catalog>sinks</> element in there, given intelligent searching, wouldn't that be both faster, more accurate, and able to filter the stupid pet trick in the first example? Sure, one could still abuse it, but over time, I suspect the semantic links like \<weekend_boring_tasks id='' reftype= hytime=ilink dtd= > etc. get stronger, that is, the domains would sort themselves out particularly where someone is keeping these in their own standardized link server database? Why can't the server or server owner choose their own categories of interest and simply register their availability? Also, one can specify a kind of protocol (term used loosely) of doctype to doctype matches, eg, when I send a \<!doctype fixsink, I always get back a \<!doctype plumbingparts among others. I admit that I am not thinking this out completely because I've got to get out the door to a gig, but something like that could work. If not, why not? I'm willing to learn. Good to hear from you Tim. Tell the Good Doctor Kiwi hello from the deep south. ;-) len </message> <message id="<DAVID.95Nov11072611@baeda.english.uottawa.ca>" date="3025081571" seqno="11374"> From: david@baeda.english.uottawa.ca (David Megginson) Newsgroups: comp.text.sgml Subject: Re: [Q] recursive ELEMENT? Date: 11 Nov 1995 12:26:11 GMT Organization: University d'/of Ottawa Message-ID: \<DAVID.95Nov11072611@baeda.english.uottawa.ca> References: \<BERND.95Nov10190143@harlie.nero.uni-bonn.de> <480qpa$3q6@crl2.crl.com> In-reply-to: jenglish@crl.com's message of 10 Nov 1995 16:29:30 -0800 The TEI DTDs also allow the recursive elements \<div> (for "division") and \<lg> (for "line group"), but they mar the symmetry by including an attribute "type" which is #CURRENT, so that you are _forced_ to specify it. David -- David Megginson Department of English, University of Ottawa, dmeggins@aix1.uottawa.ca Ottawa, Ontario, CANADA K1N 6N5 ak117@freenet.carleton.ca Phone: (613) 562-5800 ext.1203 WWW: http://www.uottawa.ca/~dmeggins FAX: (613) 562-5990 </message> <message id="<tim-1111952307220001@slug-slip14.wustl.edu>" date="3025141641" seqno="11376"> From: tim@datapage.com (Tim Rand) Newsgroups: comp.text.sgml Subject: SGML & Penta (Data General) Date: Sat, 11 Nov 1995 23:07:21 -0600 Organization: Datapage Technologies International, Inc. Message-ID: \<tim-1111952307220001@slug-slip14.wustl.edu> The firm I work for (Datapage Technologies, St. Peters, MO) performs SGML conversions for a variety of clients. Recently we have been asked to work with files that will come from a Data General, Penta system. While we have experience with a wide variety of typesetting and word processing systems, we aren't sure what to expect from Penta. Can anyone give me advice as to what formats to request data be exported in? What should we not accept? Any additional advice? -- Tim Rand \<tim@datapage.com> Preferred address \<tim@michelob.wustl.edu> Checked regularly All messages will get a response. No black holes here! </message> <message id="<CONNOLLY.95Nov12181709@w3.org>" date="3025207029" seqno="11377"> From: connolly@w3.org (Dan Connolly) Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 12 Nov 1995 23:17:09 GMT Organization: Massachvsetts Institvte of Technology Message-ID: \<CONNOLLY.95Nov12181709@w3.org> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> In-reply-to: Len Bullard's message of 10 Nov 1995 00:18:16 -0600 In article <9511100618.AA15754@fly.HiWAAY.net> Len Bullard \<cbullard@HiWAAY.net> writes: > > I'm told that approximately 4.5% of the extant web docs have a > doctype. I've heard worse numbers. I heard one Q/A dept went to test their stuff against the stuff on the net and started with the Lycos top 200 or something. Out of those 200 web pages, one (1) was a conforming HTML 2.0 document :-{ > 2. The current crop of browsers are about to meet competition from several > products which combine stylesheet and SGML compliant techniques. They > won't be free, Don't be so sure... A pattern I've seen is: * a technology emerges to attack a pretty complex problem, but a problem that at least a few folks need solved at any price * vendors step in and build products. They make compromises and mistakes, and they cut corners to meet deadlines. * Over time, as the products compete, they converge on a core set of features. They aren't generally interoperable, as customers have made use of the non-core features. So brand loyalty stays in effect. * At this point, one of two things happens: (1) the core features are standardized by a standards body (e.g. posix) . Customers that aren't happy for business reasons quit using the proprietary features, and move to another supplier that supports the standard, so some vendors lose. But overall, the market grows as a result of the standard. (2) a free implementation of the core features appears (e.g. gcc). It's too buggy at first, but over time, it matures into a usable product. (sometimes (1) happens, and then (2) -- e.g. Linux) * The thing comes full-circle when commercially support arrives for the free implementations, or when new products implement the standard. Hypertext products have been in the third stage for quite some time now. The web is an example of (2). It's in the 5th stage at this point. SGML/stylesheet products are emerging from stage three as well. The SGML standard doesn't really count as an instance of (1) cuz it standardizes an enabling technology -- just because two products support SGML doesn't mean they do anything similar at all. DSSSL might be a catalyst. We'll see. > We have learned from the Web. We learned HTTP. No offense to Dan or > Raggett, but HTML is simple enough (by design) that there isn't much for an > SGML designer to learn from it. I don't think they had that in mind. Hmmm... I certainly take no offense: I didn't really do much design work at all. I just dotted the i's and crossed the t's on Tim's work, and he got the tags from some IBM/script tagset anyway. On the other hand, I think there's something to learn from Dave's work. His work on Tables incoroprates much of the experience from past systems, tunes it for the interactive processing constraints of the web, and comes up with some novel ideas in the mix. And did you realize that FORMs were his idea? (imagemaps were Tim's idea, by the way: check the www-talk archives in the first part of '93). > The critical difference has been HTTP and that, IMO, is what should remain > stable and effective at all costs. If the protocol goes unstable, the game > is over. But that won't happen, so the future is bright, the fun has just > begun, and we may have jobs for quite a while. "Any sufficiently complex technology is indistinguishable from magic." I think you're just showing a lack of depth of knowledge in the field of protocol design. HTTP has some nifty, novel features, but it's got its warts too. The web will need new protocols over time as well as new data formats. Dan -- Daniel W. Connolly "We believe in the interconnectedness of all things" Research Scientist, MIT/W3C PGP: EDF8 A8E4 F3BB 0F3C FD1B 7BE0 716C FF21 \<connolly@w3.org> http://www.w3.org/hypertext/WWW/People/Connolly/ </message> <message id="<tomw.816216368@ccadfa>" date="3025205168" seqno="11378"> From: tomw@ccadfa.cc.adfa.oz.au (Tom Worthington) Newsgroups: comp.text.sgml Subject: Govt Electronic Document Guidelines at ACS Mtg, Wed, 15th Nov, Canberra Date: 12 Nov 95 22:46:08 GMT Organization: ADFA News Service Message-ID: \<tomw.816216368@ccadfa> ACS Canberra Branch - Monthly Meeting "Australian Government Electronic Document Management Guidelines" by Tom Worthington President-elect of the Australian Computer Society The report "Improving Electronic Document Management - Guidelines for Australian Government Agencies" is now available from the Office of Government Information Technology. These guidelines were the last item of work for the Electronic Data Management Subcommittee of the Information Exchange Steering Committee. Tom Worthington, who chaired the subcommittee which produced the report, will give a short overview of the work of the committee, outline the report, its history and discuss options for the future of electronic document management. FOLLOWED BY: "Principles of documents management for Government agencies" by Dan Daily Information Records Management Representative, Cannon Australia Mr. Daily will discuss Cannon's "exchange" software. It is claimed to be the first product which implements the recommendations of the EDMSC reports. 7pm for 7:30pm Wednesday 15th November ANUtech - ANU Campus, Canberra free for members FOR FURTHER DETAILS: Meeting: E-mail: mmorgan@acslink.net.au or ph: 06 2474830 Report: \<URL:http://acslink.net.au/~tomw/edgrel.html> ACS: E-mail: info@acslink.net.au or \<URL:http://www.acs.org.au/> Speakers: Tom Worthington: \<URL:http://acslink.net.au/~tomw/> Dan Daily: ph 06 2805798, fax: 06 2803933 Posted by Tom Worthington, Director of the Community Affairs Board & President-elect for 1996, Australian Computer Society Inc. E-mail: tomw@acslink.net.au </message> <message id="<CONNOLLY.95Nov12185209@w3.org>" date="3025209129" seqno="11379"> From: connolly@w3.org (Dan Connolly) Newsgroups: comp.text.sgml,comp.infosystems.harvest Subject: Hypermedia Addressing [was: Solving the BIG problem] Date: 12 Nov 1995 23:52:09 GMT Organization: Massachvsetts Institvte of Technology Message-ID: \<CONNOLLY.95Nov12185209@w3.org> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> \<tbray.10.30A3B6E9@opentext.com> In-reply-to: tbray@opentext.com's message of Fri, 10 Nov 1995 20:32:42 GMT In article \<tbray.10.30A3B6E9@opentext.com> tbray@opentext.com (Tim Bray) writes: > > In article <9511100618.AA15754@fly.HiWAAY.net> Len Bullard \<cbullard@HiWAAY.net> writes: > >Now. We still have the issue of URL/URN/URC etc and HyTime to resolve. > >That is a topic worthy of some debate on CTS. Hmmm... possibly. Would you care to start with a statement of the problem as you see it? Try to start with some terms that most folks know, and go from there. Many forumlations of the problem go something like: "URLs are location dependent. URNs are not." That debate never goes anywhere. I'll add my 2 cents below... > > Life for the browser and > >publishing system designers will be hard until this is resolved to > >everyone's satisfaction. I predict life for publishing system designers will always be hard. Publishing -- managing the knowledge and communication of a society -- is just plain hard. But I too suspect there's some technology we can put in place that will make it easier... > Don't hold your breath. This problem is too hard for CS types; if they'd > been able to come up with a satisfactory solution, they would have built > the Web. Since they didn't, CERN/NCSA went ahead and built it anyhow based > on absolute immoveable inflexible unstable unreliable unparseable pointers, > namely URLs. Apply enough spit, glue, and market pressure, and it works. > More than one substantial working group has been disbanded because it > decided it had to solve this problem first. > > Problem is, people have been shooting too high, trying to formalize > abstractions like "author" and "title" and [shudder] "subject". I believe > that there has to be some kind of compromise between the URL and the > unrealizable Platonic ideal. I'm just not smart enough to think it up. Well said. After noodling on the problem for a while, my favorite name for the problem is "Resource Discovery and Reliable Links." In other words: how can we make the most efficient use of the availble computing/network resources to find an information resource? And once I've found it, how can I make a link so that the next guy can follow my footsteps reliably? I've written something on the related issues at: http://www.w3.org/pub/WWW/People/Connolly/drafts/citations.html Search services like DejaNews, Infoseek, and lycos are great, but they don't scale. They're network-traffic-intensive, and as Tim@Opentext will tell you, fulltext search only scales so far. Eventually, precision suffers -- any search returns too much noise. My favorite technoloy to bring to bear on this problem is Harvest: http://rd.cs.colorado.edu/harvest/ About reliable links: my best guess at a solution is that a reliable link looks roughly like a traditional citation, library-catalog-card, business card, etc. It's got one or more URLs -- identifiers that the computer can resolve in sub-second time most of the time for relatively recent resources. But it's also got a collection of constraints, or assertions, or attributes -- a little bit of knowledge. This extra knowledge is used in conjuction with services that make large parts of the web "content addressable" -- like relational databases and fulltext indexes. So 90% of the time, the URL gets you the resource in a reasonable amount of time. In the other cases, your browser submits the contraints/assertions to various services (for free or by subscription) and locates the resource. The catch-22 in knowledge representation is that you have to constrain the problem space to a relatively small knowledge domain in order to succeed with known technology. In fact, Godel's theorem says that no single formalism (aka knowledge representation) will cover all situations, no matter what technology is brought to bear. In the mean time, widely-deployed services that automate knowledge domains such as that spanned by BibTeX would be darned useful. And I think widespread deployment of technology like Harvest, combined with intensive applications of technology like Verity and OpenText would be great. Harvest can supply the standard interfaces and the reference implementation of the gatherer and broker, and then specialized tools can be brought to bear in various domains. For example, Yahoo could live on as a Harvest broker that specializes in popular stuff in net culture -- Melrose Place, new web browsers, business news, etc. Other brokers would spring up for CS literature, Airline reservations, "Computer Shopper," etc. Domain specialists would maintain gatherers on various topics: press releases, tech report collections, etc. And I could set up my own broker that aggregates the gatherers I'm interested in :-) Dan -- Daniel W. Connolly "We believe in the interconnectedness of all things" Research Scientist, MIT/W3C PGP: EDF8 A8E4 F3BB 0F3C FD1B 7BE0 716C FF21 \<connolly@w3.org> http://www.w3.org/hypertext/WWW/People/Connolly/ </message> <message id="<11-12-1995.17755@mbfw451>" date="3025159673" seqno="11389"> Organization: MBF Systems Corporation Subject: Re: Solving the BIG problem References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@news1.mcs.net> <9511100618.AA15754@fly.hiwaay.net> \<tbray.10.30A3B6E9@opentext.com> <9511110154.AA06542@fly.hiwaay.net> Date: Sun, 12 Nov 95 02:07:53 -0800 From: ricday@mbfw451.mlnet.com (Richard S. Day) Newsgroups: comp.text.sgml Message-ID: <11-12-1995.17755@mbfw451> Followup-To: comp.text.sgml In article <9511110154.AA06542@fly.hiwaay.net> Len Bullard \<cbullard@hiwaay.net> writes: >[Tim Bray] > >| Don't hold your breath. This problem is too hard for CS types; if >| they'd been able to come up with a satisfactory solution, they would >| have built the Web. > [ lots of very interesting discussion snipped ] > >| I believe that there has to be some kind of compromise between >| the URL and the unrealizable Platonic ideal. I'm just not smart enough >| to think it up. [ another big snip ] > Why can't the server or server owner choose their own categories >of interest and simply register their availability? Also, one can specify >a kind of protocol (term used loosely) of doctype to doctype matches, eg, >when I send a \<!doctype fixsink, I always get back a \<!doctype >plumbingparts among others. len I suspect the result would be the spamming of categories. Just as we have seen an increasing number of people posting to almost every newsgroup in existence, we can expect to see a nasty number of people "choosing" every possible category for their site. I have already seen a few home pages with \<title>list of tens or even hundreds of keywords running on forever</> Having said that, I agree we need a better solution than the one we have and I like (caveat above) the idea of the server owner registering their own categories of interest. And the notion of some sort of protocol/mechanism for doctype to doctype matches would certainly save me a lot of grey hair while I try to think of synonyms for what I want to find on the current web... Ric Day ricday@mbfw451.mlnet.com </message> <message id="<485g26$mqu$1@mhafm.production.compuserve.com>" date="3025191430" seqno="11376"> From: Eileen Wood <74633.2630@CompuServe.COM> Newsgroups: comp.text.sgml Subject: Roch NY SGML Users Group Date: 12 Nov 1995 18:57:10 GMT Organization: DTD Specialists Message-ID: <485g26$mqu$1@mhafm.production.compuserve.com> Rochester, NY -- The Rochester SGML User�s Group will meet Wednesday, November 15, 1995, at Eastman Kodak Theatre on the Ridge, Building 28, room 155, 200 Ridge Road West (between Lake Ave and Dewey Ave), starting at 6:00 p.m. The meeting is free and open to the public. The presentation for this meeting will be a Case Study by Zane Dick and Eric Auer of Clark Boardman Callaghan, a Legal Speciality Publisher. The Vendor presentation will be by Karyn Lennon of Microstar on the new Near & Far Author for Word. Microstar will raffle off a copy of Near and Far, and will supply refreshments. The Rochester SGML Users Group assists users in the SGML technology for the management of text information. For information, contact Eileen Wood at DTD Specialists phone (716)223-7240, email:eileen_wood@mlstand.com. -- Eileen M. Wood DTD Specialists tel (716) 223-7240 email eileen_wood@mlstand.com </message> <message id="<9511130338.AA28510@fly.HiWAAY.net>" date="3025222690" seqno="11380"> Newsgroups: comp.text.sgml Date: Sun, 12 Nov 1995 21:38:10 -0600 Message-ID: <9511130338.AA28510@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> References: \<tim-1111952307220001@slug-slip14.wustl.edu> Subject: Re: SGML & Penta (Data General) [Dan Connolly] >DSSSL might be a catalyst. We'll see. If it isn't, then a plug-in stylesheet system will be. >"Any sufficiently complex technology is indistinguishable from magic." A. Clark lives. But I studied magick. HTTP isn't that. ;-) >I think you're just showing a lack of depth of knowledge in the field of >protocol design. No doubt. >HTTP has some nifty, novel features, but it's got its warts too. It works for now. So where to go? S-HTTP? HTTP and SSL? What? Who makes those decisions? >The web will need new protocols over time And more plug-ins to support them. Cha-ching! >as well as new data formats. The new data formats are inevitable and even necessary given the kinds of requirements elaborated at NIST. But my worries aren't what happens to the homepages. I am looking at the problems of large distributed enterprises coordinating multiple projects across different domains of skill and expertise. Using SGML, I know how that can be done reasonably well as a focusable compound document architecture is available with which I can reliably handle the multiple data formats. Replacing the innards of the application protocols too often would make this a bit more expensive. Len Bullard </message> <message id="<9511130428.AA31514@fly.HiWAAY.net>" date="3025225720" seqno="11381"> Newsgroups: comp.text.sgml,comp.infosystems.harvest Date: Sun, 12 Nov 1995 22:28:40 -0600 Message-ID: <9511130428.AA31514@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> \<tbray.10.30A3B6E9@opentext.com> \<CONNOLLY.95Nov12185209@w3.org> Subject: Re: Hypermedia Addressing [was: Solving the BIG problem] [Dan Connolly] >Hmmm... possibly. Would you care to start with a statement of the problem >as you see it? Try to start with some terms that most folks know, and go >from there. That paragraph implicitly states the problem pretty well already. 1. Make a statement as you see it. BUT 2. Try to start with terms most folks know. You know what you know but you know you never know. The system has to work with a very specific domain, (what I know) and be mapped to a very general domain (what_i_think_most_folks_know). Tough to do. It's a bit like the Calvin and Hobbes cartoon where in battle with aliens, he tries to fix a broken part. By the time he gets through all the indirection, enters a wrong number, and disambiguates the fault, he is molten aluminum on an asteroid. One can: 1. Explain everything explicitly to everyone all of the time (URL). 2. Only address audiences of experts (semantic link). 3. Provide a system with built-in expertise for use by everyone else (semantic link and scripted control). Semantic links (ilinks) and formal system identifiers combined with a standard query language would be a good start along with entity-based storage objects and standard scripting. Then let each domain create its own types and coordinate communication based on those types. This has always worked for the government, the corporations, and every other document centric enterprise that resisted the temptation to flatten their documents out into one document type with one format and direct links. The military has done this for aeons. Their indexing designs are superb. The military logistics experts do good work when the task involves coordination of very complex performances. There is waste. That is the price of assured success. >I predict life for publishing system designers will always be hard. >Publishing -- managing the knowledge and communication of a society -- is >just plain hard. But I too suspect there's some technology we can put in >place that will make it easier... yep. [Tim Bray] >Problem is, people have been shooting too high, trying to formalize >abstractions like "author" and "title" and [shudder] "subject". I believe >that there has to be some kind of compromise between the URL and the >unrealizable Platonic ideal. I'm just not smart enough to think it up. [Dan Connolly] >Well said. Truth is, we've never had much success at getting everyone to agree on the semantic of specific element types or even their content models at least, if all the relevant parties are consulted. The best I think can be done is to specify the mechanism by which the agreement is made and formalized. For that, doctypes and architectural forms seem adequate. Then registries maintained by communities of interest. [Dan Connolly] >About reliable links: my best guess at a solution is that a reliable link >looks roughly like a traditional citation, library-catalog-card, business >card, etc. Yep. That was suggested on CTS last year. Any service resembling a Librarian helps. A reliable link typically points to a service, not a data object. >The catch-22 in knowledge representation is that you have to constrain the >problem space to a relatively small knowledge domain in order to succeed >with known technology. For which SGML has the concepts of document types and architectural forms. >Domain specialists would maintain gatherers on various topics: press >releases, tech report collections, etc. That's what DoD calls the subject matter experts (SMEs). >And I could set up my own broker that aggregates the gatherers I'm >interested in :-) Cool. But share it with the rest of us in the old fashioned spirit of the World Wide Web. Len Bullard </message> <message id="<1995Nov13.040057.12114@sq.com>" date="3025224057" seqno="11382"> Newsgroups: can.jobs,ont.jobs,comp.text.sgml From: lee@sq.com (Liam R. E. Quin) Subject: Wanted: SGML typists Message-ID: <1995Nov13.040057.12114@sq.com> Organization: SoftQuad Inc., Toronto, Canada Date: Mon, 13 Nov 1995 04:00:57 GMT The SoftQuad Inc. Professional Services Division is looking for experienced SGML typists. You should be familiar with a native SGML authoring package such as Author/Editor, including working with SGML tables. The positions are available immediately. We are prepared to work with home-workers so you don't need to be in Toronto. The work involves a combination of making corrections to SGML documents so that they parse, or so that they match a printed original, and also entering data (mostly tabular) by hand. Please contact Liam Quin Senior Technical Analyst, SoftQuad Inc., Toronto, ONT, CANADA M8X 2W4 lee@sq.com telephone +1 416 239 4801 fax +1 416 234 9188 Attn. Liam Quin (please _don't_ send me a mime-encoded version of your HTML resume'e, or just your URL... I am going to print the replies, take them home, and read them. If you just send your URL, I probably won't get to read it. Thanks.) -- Liam Quin, SoftQuad Inc +1 416 239 4801 lee@sq.com \<URL:http://www.sq.com/> HexSweeper NeWS game;OPEN LOOK+XView+mf-fonts FAQs;lq-text unix text retrieval </message> <message id="<486fge$4ul@ixnews2.ix.netcom.com>" date="3025223630" seqno="11383"> From: klamerus@ix.netcom.com (Mark Klamerus ) Newsgroups: comp.text.sgml Subject: Anyone know of any Diagnostics system written in SGML??? Date: 13 Nov 1995 03:53:50 GMT Organization: Netcom Message-ID: <486fge$4ul@ixnews2.ix.netcom.com> Does anyone know of a vehicle or other diagnostics system which has been written in SGML? Mail especially appreciated. regards, Mark </message> <message id="<486cv6$6ri@pubxfer4.news.psi.net>" date="3025221030" seqno="11384"> From: kimber@passage.com (W. Eliot Kimber) Newsgroups: comp.text.sgml Subject: Re: [Q] recursive ELEMENT? Date: 13 Nov 1995 03:10:30 GMT Organization: Passage Systems, Inc. Message-ID: <486cv6$6ri@pubxfer4.news.psi.net> References: \<BERND.95Nov10190143@harlie.nero.uni-bonn.de> In article \<BERND.95Nov10190143@harlie.nero.uni-bonn.de>, bernd@nero.uni-bonn.de� says... >Question: is a recursive definition allowed? As Joe English pointed out, IBMIDDoc uses recursive divisions, precisely divisions can be re-used in various contexts. Note that it is usually a simple matter (I hesitate to say "trivial", it always gets me in trouble) to transform from explicitly-leveled divisions to recursive divisions, assuming the rest of the content models are the same. Recursive divisions can make processing and style sheets a bit more complicated because the processor has to figure out what the level is since the author didn't make it brain-dead simple for the computer. This is usually more of a problem in style sheets than in batch processors because batch processors can just keep a counter but styles often require deep qualification to distinguish recursive divisions at different levels. >Question: are there DTD's defining such a "depth of fold" >counter mechanism, and/or formatters relying on this? The TAGLVL (tag level) setting in the SGML declaration effectively limits the depth of nesting by limiting the number of elements that can be open at once. For example, the default value of 24 would allow at most about 10 levels of nesting for a DTD with very little extra containment, less if you use some sort of inner container, such as Joe suggested: 1 for document element, 1 for major document divisions (fronmatter, body, etc. in typical DTDs), 2 for each level of division (1 for division, one for leaf elements within the division). Adding another level of containment within a division, say a DivBody to keep paragraphs separate from subdivisions, would limit you to about 5 levels. For formatting, I think most people just account for whatever depth the authors of a set of documents actually use or are likely to use. I've never had unreasonable nesting cause a problem. -- \<Address HyTime=bibloc> W. Eliot Kimber (kimber@passage.com) Systems Analyst and HyTime Consultant Passage Systems, Inc., 2608 Pinewood Terr., Austin TX 78757 (512)339-1400 10596 N. Tantau Ave, Cupertino CA, 95014, (408) 366-0300 "SGML or die" \</Address> </message> <message id="<9511131336.AA02123@fly.HiWAAY.net>" date="3025258599" seqno="11395"> Newsgroups: comp.text.sgml Date: Mon, 13 Nov 1995 07:36:39 -0600 Message-ID: <9511131336.AA02123@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@news1.mcs.net> <9511100618.AA15754@fly.hiwaay.net> \<tbray.10.30A3B6E9@opentext.com> <9511110154.AA06542@fly.hiwaay.net> <11-12-1995.17755@mbfw451> Subject: Re: Solving the BIG problem [Ric Day] | I suspect the result would be the spamming of categories. Just as we | have seen an increasing number of people posting to almost every | newsgroup in existence, we can expect to see a nasty number of people | "choosing" every possible category for their site. I have already seen | a few home pages with | | \<title>list of tens or even hundreds of keywords running on forever</> You are right, Ric. I guess what Tim Bray is saying and I agree with is that we may not make the ideal ("platonic") system, but so what. To me that says we can't keep the system open and get rid of all of the potential abuse. It's the price of freedom and cheap service. We can make it possible for server owners to tighten the system up with some extra description (content modeled DTDs), typed links (see the CaPH proposals for examples) and processing (scripted links). Perhaps that will implicitly redistribute loads by aggregation of interest. We will see more problems as link cascading starts showing up from apps like VRML (which is great fun, and highly recommended for weekend hacking!!) This means we make some choices about how this is done. Dan Connolly points to Harvest technology. I glanced at the pages last night and all the right ideas are in place there. I've not looked at it in enough detail to understand it all, but it appears to be another object-oriented whatis. That is fine for a service. It seems to me that to preserve some of the openness for which the Web was acclaimed, standard addressing, indexing and classification techniques are needed which services can rely on. The IETF groups are working on that as is the WG8 ISO committee responsible for SGML and HyTime. There may be a convergence in the works of these, but anyone who tries to sort them out hits a pretty tough learning curve. Both groups have their own technologies, argot, leadership, very different policies, and openly hostile communities. It ain't pretty and it ain't fun. OTH, that's life at the zoo. By and by, nature will force some solutions. As the reports from my service provider about anomalous Internet behavior indicate, some serious load problems are manifesting as were predicted on CTS. These are directly attributable to Web activity, so the Mothers of Invention are tuning up. You will probably see the advent of server certification services soon. Len Bullard </message> <message id="<487a4u$ld1@news.Belgium.EU.net>" date="3025250910" seqno="11396"> From: Jacques Deseyne \<jad@sema.be> Newsgroups: comp.text.sgml Subject: Re: SGML Windows Editors Date: 13 Nov 1995 11:28:30 GMT Organization: SEMA Group Belgium Message-ID: <487a4u$ld1@news.Belgium.EU.net> References: <46m4ms$1p2@newsbf02.news.aol.com> <46r2nt$ncj@grok.provo.novell.com> \<ucu44rlgvk.fsf@tdb.teleport.com> <47ks6i$fv8@news.Belgium.EU.net> <47tvmc$53ku@tigger.cc.uic.edu> cmsmcq@tigger.cc.uic.edu (C M Sperberg-McQueen) wrote: >Jacques Deseyne (jad@sema.be) wrote: > >: You don't have to modify correct DTDs in order to use them with >: WP SGML Edition. > >Does this mean that Word Perfect have fixed the bugs in the old >Intellitag version of the DTD2LGC program? I certainly hope so - can >anyone say for sure? > The phrase "correct DTDs" was only a repetition of the posting from Russel Senior I reacted to. I actually meant (and should have expressed it that way) : "You don't have to surround your DTDs with the \<!DOCTYPE xyz [...]> delimiters in order to use them with WP SGML Edition" >The bugs I found in the old program were: > > - failure to handle nested marked sections correctly (IGNOREd sections > must be scanned for nested marked sections - other parsers have also > been reported to have this bug, but most have been fixed) This one seems to persist in the versions of DTD2LGCW.EXE I have. > - failure to allow adjustment of capacities, even though the manual > said they could be increased; my recollection is that GRPCNT which > was hard-coded to 32 or something > The documentation gives the following maximal values. I tested out most of them and changes seem to be accepted (e.g., a DTD needing a GRPCNT of at least 64 was accepted). Maximum for WPSGML ATTCNT 80 ATTSPLEN 2048 BSEQLEN 960 (actually IGNORED) DTAGLEN 16 (actually IGNORED) DTEMPLEN 16 (actually IGNORED) ENTLVL 32 GRPCNT 256 GRPGTCNT 512 GRPLVL 32 LITLEN 2048 NAMELEN 100 NORMSEP 2 PILEN 240 (actually IGNORED) TAGLEN 2048 TAGLVL 80 I never wanted to give the impression that the DTD2LGC(W).EXE utility was the ultimate DTD validating utility. Despite these (and other) limits, I think that WP SGML Edition is probably the most advanced among products trying to combine SGML editing functions with a commercial word processor. Let's hope they will be able to continue their efforts in the future ! Best regards, ------------------------------------------------------ Jacques Deseyne \<jad@sema.be> Sema Group Belgium - Stallestraat 96 - B-1180 Brussels ------------------------------------------------------ </message> <message id="<4878vq$4e9@news.cict.fr>" date="3025249722" seqno="11397"> From: mr@piau () Newsgroups: comp.text.sgml Subject: Re: What is SGML? (Language or metalanguage?) Date: 13 Nov 1995 11:08:42 GMT Message-ID: <4878vq$4e9@news.cict.fr> References: <47tdeg$ja0@hopper.acm.org> Reply-To: michel.rodriguez@avions.aerospatiale.fr davep@ACM.ORG wrote: : Most people know what a relational database is. Let me use that : as an analogy for comparison. : \<analogy> : Just to have data in a relational database, one must accept a : data structure format of tables, with some fields linked. : Just to have data in an SMGL system, one must accept a : data structure format of a typed hierarchy whose leaf : nodes are either empty or arbitrary character strings. : To describe the "shape" of a particular relational database, : and to name the tables and the fields therein, one uses a : "data definition language". : To describe the "shape" of a particular typed hierarchy : data base: to name the types and prescribe what types : of nodes (elements) may or must occur within nodes of : each type, and in what orders they may occur, one uses : a "data definition language" (the formal part of a DTD) [interesting explanations... deleted] : Dave Peterson : SGMLWorks! : davep@acm.org -- Could I try a (very different) analogy ? I would think that the SGML model (a hierarchy of nodes, each nodes bearing attributes, one of them being the type). This is equivalent to the tables definition of a relationnal database (the actual schema) Then a DTD represent the integrity constraints on a specific document : which node types are allowed, which attributes etc... Any comments on that ? ------------------------------------------------------------------------ \\\\ \\\\ \\\\ \\\\ \\\\ \\\\ / \\ / \\ / \\ / \\ / \\ / \\ \\ \\ \\ \\ \\ \\ \\ \\(__) \\ \\(__) \\ \\(__) *--> (oo) *--> (oo) *--> (oo) / / \\/ / / \\/ / / \\/ / / / / / / \\ / \\ / \\ / \\ / \\ / \\ / // // // // // // A FLIGHT OF FLYING MOOCHEL ------------------------------------------------------------------------ michel.rodriguez@avions.aerospatiale.fr </message> <message id="<487ehr$ons@rsc3.hermes.si>" date="3025255419" seqno="11398"> From: Bogo Vatovec \<bogo@hermes.si> Newsgroups: comp.text.sgml Subject: Re: FrameMaker -> SGML Conversion Date: 13 Nov 1995 12:43:39 GMT Organization: Hermes SoftLab Message-ID: <487ehr$ons@rsc3.hermes.si> References: <47lmvq$jj7@hpbab.wv> tsmith@wv.mentorg.com (Teresa Smith) wrote: >Does anyone have any recommendations for how to convert >legacy documents from FrameMaker to SGML? > At Hewlett-Packard we use a Frame product called Frame Builder. It is essentially FrameMaker with some structure elements added on top. In a customized version which I believe was developed for HP only, we have SGML export and import possibility. DTD is OpenBook. >For the time being, we'd like to continue authoring in >FrameMaker, so we want to develop some way to continually >export SGML when it's time to ship our documents. > I heard that FrameMaker 5.0 is already out and that it has SGML conversion. Maybe you could check this with your dealer. My 20 cents, Bogo -- ----------------------------------------------------------------------- Bogo Vatovec, HERMES SoftLab, Litijska 51, 61000 Ljubljana, Slovenia Phone: +386 61 1405 004 Fax: +386 61 1405 070 "There is a crack, a crack in everything, that's where the light gets in", Leonard Cohen ----------------------------------------------------------------------- </message> <message id="<816252263snz@sgml-cen.demon.co.uk>" date="3025241063" seqno="11399"> From: Martin Bryan \<mtbryan@sgml-cen.demon.co.uk> Newsgroups: comp.text.sgml,comp.infosystems.harvest Subject: Re: Hypermedia Addressing [was: Solving the BIG problem] Date: Mon, 13 Nov 95 08:44:23 GMT Organization: The SGML Centre Message-ID: <816252263snz@sgml-cen.demon.co.uk> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> \<CONNOLLY.95Nov12185209@w3.org> Reply-To: mtbryan@sgml-cen.demon.co.uk In article \<CONNOLLY.95Nov12185209@w3.org> connolly@w3.org "Dan Connolly" writes: > About reliable links: my best guess at a solution is that a reliable > link looks roughly like a traditional citation, library-catalog-card, > business card, etc. Actually it need to be a bit more like a Cataloguing in Publishing information. It needs a standards classification identifier (Dewey, Library of Congress, etc), some search keywords, a unique identifier (the ISBN), etc. The only thing a libary-catalog-card needs to add to this is a reference to where it is stored in the local system, i.e. its local URL! Let me add a brief plea to consider HyTime topic maps as a way of linking documents with the same keywords, or which have been categorized in the same way. The whole question of finding things is currently postulated on searching, which just does not work in a multilingual environment. In Paris the other day I wanted to find all the pictures in the Louvre with a dog on them. OK, www.culture.fr has an index of paintings in the Louvre. Now search for dog! Nothing found!! No dogs in any pictures? You bet your life not. The problem is that the French for dog is chien. No amount of free text searching in English will find this. You need a multilingual thesaurus to equate keywords! Worldwide searching by search engines are today failing to get sufficient hits for just this reason. Add to this the problem of noise that always appears when free text searching a database with more than 1 million pages in a given subject area and you have a potential for total chaos. -- Martin Bryan @ The SGML Centre, Churchdown, Glos. GL3 2PU, UK (+44 1452 714029) </message> <message id="<487kcp$dsp@cliffy.lfwc.lockheed.com>" date="3025261400" seqno="11400"> From: mcclellantj@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: Hardcopy to SGML conversion metrics Date: 13 Nov 1995 14:23:20 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <487kcp$dsp@cliffy.lfwc.lockheed.com> References: <47th1j$1icc@tetsuo.communique.net> Reply-To: mcclellantj@lfwc.lockheed.com Jefferson Spear (jefferson_spear@psislidell.com) wrote: : Anybody have any numbers that estimate the time it takes to convert a : certain number of hardcopy pages to be scanned or OCRed, proofed, then : converted into SGML? ^^^^^^^^^^^^^^^^^^^ This part would be _wildly_ variable. It depends on the complexity of your markup language (defined in the DTD) and the rigor applied to structure (consistency) of the documents in the original hardcopy. i.e. inconsistent hardcopy structure will greatly increase the cost of converting to SGML (and probably cannot be done by a machine). -- Tad McClellan, Logistics Specialist (IETMs and SGML guy) email: mcclellantj@lfwc.lockheed.com All I want, is a little more than I'll ever get. </message> <message id="<487l7o$dsp@cliffy.lfwc.lockheed.com>" date="3025262264" seqno="11401"> From: mcclellantj@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: [Q] recursive ELEMENT? Date: 13 Nov 1995 14:37:44 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <487l7o$dsp@cliffy.lfwc.lockheed.com> References: \<BERND.95Nov10190143@harlie.nero.uni-bonn.de> Reply-To: mcclellantj@lfwc.lockheed.com Bernd Kreimeier (bernd@nero.uni-bonn.de) wrote: : Question: are there DTD's defining such a "depth of fold" ^^^^^^^^^^^^^ : counter mechanism, and/or formatters relying on this? I believe the maximum depth of fold is specified in the SGML declaration TAGLVL quantity (defaults to 24). From ISO-8879 para 7.3.3: "The number of open elements cannot exceed the "TAGLVL" quantity" -- Tad McClellan, Logistics Specialist (IETMs and SGML guy) email: mcclellantj@lfwc.lockheed.com Just be thankful we don't get all the government we pay for! </message> <message id="<487lfp$eku@newsbf02.news.aol.com>" date="3025262521" seqno="11402"> From: epsgroup@aol.com (Epsgroup) Newsgroups: comp.text.sgml Subject: Re: SGML & Penta (Data General) Date: 13 Nov 1995 09:42:01 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <487lfp$eku@newsbf02.news.aol.com> References: \<tim-1111952307220001@slug-slip14.wustl.edu> Reply-To: epsgroup@aol.com (Epsgroup) I have over 5 years experience converting Penta to SGML and it is not an easy task. Penta is a very open coding system, there are many ways to do the same thing (and most users use all of them). There is also no document structure and it is all dependent on how the formats (or style sheets) are set up. The best way to start is ask for ASCII exported files with coding in place. Take a look at these files and you will see what I mean. If you need any assistance let me know, I am relatively cheap as far as conversion services go and I can give a good turn around. Please respond to mtorr@gnn.com and not epsgroup. Thanks Mike Torrence Technical Director Electronic Publishing Services Epsgroup@aol.com </message> <message id="<816274368.97@moston.demon.co.uk>" date="3025291991" seqno="11403"> From: david@moston.demon.co.uk (David Page) Newsgroups: comp.text.sgml Subject: PRESS RELEASE - MILES 33 Date: Mon, 13 Nov 1995 22:53:11 GMT Message-ID: <816274368.97@moston.demon.co.uk> Miles 33 International will be at the GCA SGML '95 Conference and Expo, Sheraton Boston Hotel and Towers, Boston, December 4-7. Please stop at the Miles 33 booth and see Genera SGML Pagination running with live data from Deere and Company. In Europe, Pixellence, Miles 33's French distributor, will be demonstrating Genera SGML Pagination at Documation '95 Conference and Expo, Hotel Meridien Etoile, Paris, France, December 12 and 13. </message> <message id="<487g9m$6fq@naurouze.cert.fr>" date="3025257206" seqno="11404"> From: Martin Hollender \<Martin.Hollender@onecert.fr> Newsgroups: comp.text.sgml Subject: Re: Anyone know of any Diagnostics system written in SGML??? Date: 13 Nov 1995 13:13:26 GMT Organization: EURISCO Message-ID: <487g9m$6fq@naurouze.cert.fr> References: <486fge$4ul@ixnews2.ix.netcom.com> To: klamerus@ix.netcom.com I'm noot shure how you define diagnostics system, but if you mean trouble-shooting systems than you should have a look at CALS (Computer-aided acquisition and logistic support) of the US DoD http://navysgml.dt.navy.mil/home.html MIL-D-87269 Interactive electronic technical manuals is pointing to DTDs for trouble-shooting Martin </message> <message id="<DHzr3L.4xs.0.-s@cs.vu.nl>" date="3025270641" seqno="11405"> Newsgroups: comp.text.sgml From: jrvosse@cs.vu.nl (JR van Ossenbruggen) Subject: Re: Hytime Question References: <309E569D.6F28@ecrc.gmu.edu> <47qett$fes@pubxfer5.news.psi.net> Sender: news@cs.vu.nl Organization: Fac. Wiskunde & Informatica, VU, Amsterdam Date: Mon, 13 Nov 1995 16:57:21 GMT Message-ID: \<DHzr3L.4xs.0.-s@cs.vu.nl> W. Eliot Kimber (kimber@passage.com) wrote: : The CMIF team has submitted a paper to the Hypertext '96 conference : to be held in the US in May (Boston?) and I assume it will be accepted. : They've demonstrated being able to represent sophisticated hypermedia : presentations using HyTime event schedules. They've also done a good job : of evaluating the design of event schedules and have provided some useful : insights into their use and possible refinement. I urge you to look for : the paper when the proceedings of the conference are published. They've : provided the DTD for the system at : http://www.cs.vu.nl/~jrvosse/HyTime/cmif.dtd. Eliot, our paper has not been accepted. People interested in the draft version can obtain a copy from ftp://ftp.cwi.nl/pub/mmpapers/CMIF-HyTime.ps.gz Jacco van Ossenbruggen </message> <message id="<4880d1$slo@bell.maths.tcd.ie>" date="3025273697" seqno="11406"> From: tim@maths.tcd.ie (Timothy Murphy) Newsgroups: comp.text.sgml Subject: Re: HTML as SGML dtd Date: 13 Nov 1995 17:48:17 -0000 Organization: Dept. of Maths, Trinity College, Dublin, Ireland. Message-ID: <4880d1$slo@bell.maths.tcd.ie> References: <4768a8$qbo@lanczos.maths.tcd.ie> <476e6a$1fl@crl2.crl.com> <47ivdn$bu5@bell.maths.tcd.ie> <47ld2j$cs4@crl9.crl.com> Keywords: SGML HTML jenglish@crl.com (Joe English) writes: >Use the one in RFC 1866 (formerly draft-ietf-html-spec-06.txt), >which is the same as the one at > \<URL:http://www.w3.org/hypertext/WWW/MarkUp/html-spec/html-pubtext.html> I wasn't able to access this URL -- it seemed to be read-protected. -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland </message> <message id="<jfritchDHzuJD.DDs@netcom.com>" date="3025275096" seqno="11418"> Newsgroups: comp.text.sgml From: jfritch@netcom.com (Jeanne Fritch) Subject: SGML and Page Composition Message-ID: \<jfritchDHzuJD.DDs@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Mon, 13 Nov 1995 18:11:36 GMT Sender: jfritch@netcom14.netcom.com I am looking for people using SGML and still have important Page Layout requirements. The questions I have are: 1. Describe the process from SGML to output. 2. What tools ae you using? 3. Is it a batch process or does a lot of work need to be done on the information at publishing time. 4. What was the process of setting up the environment? 5. Do you deal with A-pages and how? 6. Tips you can suggest for someone starting the setup process. I appreciate any information you can send me. Jeanne ------------------------------------------------------- Jeanne Fritch JF Consulting email: jfritch@netcom.com 17014 W. 16th Ave. phone: (303) 384-3969 Golden, CO 80401 "SGML Training and Support. Helping out in your SGML needs." </message> <message id="<4885ts$b91@ernie.idi.oclc.org>" date="3025279356" seqno="11420"> From: magnus salgo \<magnus@ids.se> Newsgroups: comp.text.sgml Subject: PI - and RE Date: 13 Nov 1995 19:22:36 GMT Organization: Information Dimensions, Inc. Message-ID: <4885ts$b91@ernie.idi.oclc.org> Could someone help me to read the Handbook. I am trying to find out what is the right way of handling Process Instructions and RE. The process instruction I will use does not return data. It is a switch in the environment for the receiving viewer system (compare the Handbook page 340 - Clause 8) I have the following examples with a process instruction called PROC a) RE before Process intructions Example ==================================== RSthisRE RS\<?PROC>textRE ==================================== This is equal to RSthis\<?PROC>textRE b) RE after a Process Instruction Example ==================================== RStext before\<?PROC>RE RSthis text includes a process instructionRE ==================================== This is equal to RStext before\<?PROC> this text includes a process instructionRE My conclusion is that to get a space in a) between the words then I have to add a space at the nend of the first line or a space at the start of line 2. Any comments ?? Regards Magnus Salgo magnus@ids.se http://www.ids.se/personal/magnus </message> <message id="<487tql$mus@dove.nist.gov>" date="3025271061" seqno="11421"> From: bagwill@kangaroo.ncsl.nist.gov (Robert Bagwill) Newsgroups: comp.text.sgml,comp.infosystems.www.authoring.html Subject: Re: Slight flame on SGML Date: 13 Nov 1995 17:04:21 GMT Organization: NIST Message-ID: <487tql$mus@dove.nist.gov> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> In article <47tn8b$m7r@news.iastate.edu>, choward@iastate.edu (Chris Howard) writes: >HTML is a pointy stick SGML is Ada, HTML is MS-DOS batch files. >The beauty of pointy sticks is that they're handy and easy to use. Nostradamus predicts: In a few more months, almost no HTML will be hand-generated. Almost everyone will be using their favorite WYSIWYG word processor, or one of the new web-oriented ones. Artists, who are currently bitching about not being able to use HTML to do the kind of layout they want to do, will be using some tool that will give them output close enough to what they want. It may be HTML 4.0, it may be PDF+, it may be Netscape HTML\<tm>. They will never have to learn what a doctype or DTD is. Serious Web publishers, on the other hand, will be going nuts because it will be almost impossible to give their publications a consistent look-and-feel, since every artist will be using their own incompatible tool, and because it will be impossible to automatically generate tables-of-content, indexes, bibliographies, updated links, etc. So publishers who ignored SGML the first time around will either rediscover it, or will buy into proprietary tools that will let them control the production of thousands of complex documents, or both (since interoperability and interchange seem to be "a custom More honoured in the breach than the observance." in the SGML world. That's why what authors think about HTML or SGML or whatever is irrelevant. -- Bob Bagwill \<rbagwill@nist.gov> </message> <message id="<48806c$o7t@news2.isp.net>" date="3025273646" seqno="11422"> From: eberg@slip.net (Eric Berg) Newsgroups: comp.text.sgml Subject: Re: [Q] SGML for Linux? Date: Mon, 13 Nov 1995 17:47:26 GMT Organization: Mazama Software Labs, Incorporated Message-ID: <48806c$o7t@news2.isp.net> References: \<BERND.95Nov8184954@harlie.nero.uni-bonn.de> \<CONNOLLY.95Nov9004740@w3.org> \<evpwf2niam.fsf@haukugle.ii.uib.no> Reply-To: eberg@mazama.com magnus@haukugle.ii.uib.no (Magnus Y Alvestad) wrote: >[Bernd Kreimeier] > >| 3) HTML files (dunno about EPSF, possibly conversion to GIF?, dunno >| about formulae, possibly to LaTex to DVI to GIF?) > >[Dan Connolly] > >| QWERTZ: NO (?) >| linuxdoc: YES >| SNAFU/GF: YES > >Regarding qwertz; it's trivial to just plug the HTML support in >linuxdoc into qwertz. After all, linuxdoc is based on qwertz. The problem with this is that, at least for LinuxDoc/SGML it is restricted to the LDP DTD, thus not really useful for other DTD's. -E. X-------------- Eric D. Berg Mazama Software Labs, Inc. -- \<URL:http://www.mazama.com/> eberg@mazama.com -- eberg@slip.net -- 71172.43@compuserve.com X-------------- </message> <message id="<488i93$8kf@franklin.cc.utas.edu.au>" date="3025292003" seqno="11426"> From: John Lamp \<John.Lamp@cs.utas.edu.au> Newsgroups: comp.text.sgml Subject: Re: SGML & Penta (Data General) Date: 13 Nov 1995 22:53:23 GMT Organization: University of Tasmania, Department of Computer Science Message-ID: <488i93$8kf@franklin.cc.utas.edu.au> References: \<tim-1111952307220001@slug-slip14.wustl.edu> <9511130338.AA28510@fly.HiWAAY.net> Len Bullard \<cbullard@HiWAAY.net> wrote: >[Dan Connolly] > >>"Any sufficiently complex technology is indistinguishable from magic." > >A. Clark lives. But I studied magick. HTTP isn't that. ;-) > >>I think you're just showing a lack of depth of knowledge in the field of >>protocol design. > >No doubt. > >>HTTP has some nifty, novel features, but it's got its warts too. > >It works for now. So where to go? S-HTTP? HTTP and SSL? What? Who >makes those decisions? > >>The web will need new protocols over time HTTP is a simplistic, limited protocol, no doubt. I do doubt that the "web" (whatever that means) needs new protocols. It would have been better if it had used one of the more robust existing protocols. My choice to replace the HTTP/HTML modalities of the web is Z39.50/SGML. Version 3 of Z39.50 moves it out of librarianship camp, providing authentication, access control, extensibility, costing -- the work= s. Tie that in with SGML and you're talking power applications. Have a look at http://lamp.cs.utas.edu.au/net.html for Z and SGML re= sources and papers. Cheers John -- _--_|\\ John Lamp, originating in Hobart, Tasmania / \\ Phone: 002 20 2375 - Fax: 002 20 2913 \\_.--._/ email: John.Lamp@cs.utas.edu.au v <--<< http://lamp.cs.utas.edu.au/jw_lamp.html Information Systems Group, Dept of Comp Science, Uni of Tasmania </message> <message id="<488v9v$3se@crl2.crl.com>" date="3025305343" seqno="11425"> From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: sgmls and NAMELEN Date: 13 Nov 1995 18:35:43 -0800 Organization: Mentally ill venomous purists with strange emotional drives Message-ID: <488v9v$3se@crl2.crl.com> References: <47v99s$m1b@relay.lsl.co.uk> <480d3v$sse@crl2.crl.com> <488s82$69@bell.maths.tcd.ie> Keywords: sgmls,NAMELEN Timothy Murphy \<tim@maths.tcd.ie> wrote: >jenglish@crl.com (Joe English) writes: >>SGML declarations tend to be rather long and not terribly >>informative to human beings, so it's more convenient to keep >>a single copy in a separate file and pass this on the 'sgmls' >>command line, e.g.: > >> sgmls [ other flags ] sgml.decl mydoc.sgml > >I found this out by painful experience. >It seems completely undocumented in the SGML Handbook. It's in there; see productions 1 and 2: | [1] SGML document (6.1, 294:5) = | ( SGML document entity [2], | *( SGML subdocument entity [3] | | SGML text entity [4] | | character data entity [5.1] | | specific character data entity [5.2] | | non-SGML data entity [6] ) ) | | [2] SGML document entity (6.2, 295:13) = | ( *s [5], | SGML declaration [171], <== here | prolog [7], | document instance set [10], | Ee ) (Sorry, no page number, my copy's at home.) The Handbook doesn't say anything about 'sgmls', naturally, though the sgmls man page says: | DESCRIPTION | Sgmls parses and validates the SGML document entity in | filename... [...] | Note that the document entity may be spread amongst sev- | eral files; for example, the SGML declaration, document | type declaration and document instance set could each be | in a separate file. >Why don't .sgml files specify the .decl they expect? The answer to that is in productions 171 through 204: the blasted things are way too long -- the SGML declaration for HTML is 93 lines -- and there's no way to include them by reference. (This is one of SGML's most painful design flaws, IMO.) In practice, there tends to be a many-to-one mapping from document types to SGML declarations; an SGML declaration is usually supplied with each DTD or DTD family. (An extension to the TR9401 catalog file format allows this mapping to be performed automatically in cases where the document character set can be intuited far enough to read the initial \<!DOCTYPE ... > declaration.) In fact, I've managed to get by with a single, system-wide SGML declaration for every document I've ever written or received, regardless of the DTD. Whenever a DTD exceeds a quantity, or a document exceeds a capacity, I just increase the corresponding number and stop worrying about it. (This practice will get me into trouble if I ever need to deal with a character set that isn't ASCII or ISO 8859-1, of course...) --Joe English jenglish@crl.com </message> <message id="<488r3b$ruf@bell.maths.tcd.ie>" date="3025301035" seqno="11428"> From: tim@maths.tcd.ie (Timothy Murphy) Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 14 Nov 1995 01:23:55 -0000 Organization: Dept. of Maths, Trinity College, Dublin, Ireland. Message-ID: <488r3b$ruf@bell.maths.tcd.ie> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com> jbottoms@world.std.com (John W Bottoms) writes: >I disagree. SGML is designed to be just that! A meta-grammar. We haven't >had one until now and it will take a while for people to understand that >one is needed and how it is used. We have had plenty of grammars before >but they didn't fulfill needs that exist. There is a well-defined mathematical notion of "grammar", as described eg in M. Davis & E. Weyuker, "Computability, Complexity and Languages", and in many other similar works. A grammar in this sense is defined by a set of "productions" of terminals and variables, eg S -> abT. It seems to me that SGML is definitely a grammar in this sense, and in fact probably belongs to the sub-class of context-free grammars. -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland </message> <message id="<488s82$69@bell.maths.tcd.ie>" date="3025302210" seqno="11429"> From: tim@maths.tcd.ie (Timothy Murphy) Newsgroups: comp.text.sgml Subject: Re: sgmls and NAMELEN Date: 14 Nov 1995 01:43:30 -0000 Organization: Dept. of Maths, Trinity College, Dublin, Ireland. Message-ID: <488s82$69@bell.maths.tcd.ie> References: <47v99s$m1b@relay.lsl.co.uk> <480d3v$sse@crl2.crl.com> Keywords: sgmls,NAMELEN jenglish@crl.com (Joe English) writes: >SGML declarations tend to be rather long and not terribly >informative to human beings, so it's more convenient to keep >a single copy in a separate file and pass this on the 'sgmls' >command line, e.g.: > sgmls [ other flags ] sgml.decl mydoc.sgml I found this out by painful experience. It seems completely undocumented in the SGML Handbook. Why don't .sgml files specify the .decl they expect? This seems to me like a C program not specifying what header files are required, or a LaTeX program not specifying the style it is using, but leaving it to you to guess. I just compiled "sp" on my Linux system -- it was 10MB in size! Admittedly when stripped it reduced to 550k -- but that is still twice the size of Linux itself. Is it really that big a task? (Incidentally, sp did not compile under SunOS or Mips RiscOS. The problem could be trivial -- I didn't spend time on it. It seemed to be in the C++ code. I was using gcc/g++.) SGML seems to me a very nice idea, very badly implemented. -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland </message> <message id="<489p7e$gdi@gidora.kralizec.net.au>" date="3025371448" seqno="11430"> From: jon@zeta.org.au Newsgroups: comp.text.sgml Subject: SGML + Lotus Notes Date: Tue, 14 Nov 95 20:57:28 Organization: Kralizec Dialup Internet Sydney, +61-2-837-1183 V.32bis Message-ID: <489p7e$gdi@gidora.kralizec.net.au> Has anyone seen any good examples of a hybrid Lotus Notes/SGML system or any toolkits that can be used to help build such systems? jon. </message> <message id="<489r37$m8l@news.Belgium.EU.net>" date="3025333807" seqno="11431"> From: stevenn@kluwer.be (Steven Noels) Newsgroups: comp.text.sgml Subject: How to recrute SGML-typists? Date: Tue, 14 Nov 1995 10:30:07 GMT Organization: Kluwer Legal Publishers Belgium Message-ID: <489r37$m8l@news.Belgium.EU.net> Beloved c.t.s.-ers, I lurk, therefore Iam. Although we're using SGML since a few years in our company, we are still searching for the definitive, ideal test for SGML-inputters. We know they should have a fast type-rate, they should be friendly & nice, and so on... We're just wondering if anybody in this group has already compiled a concise, one hour, ArborText ADEPT-Editor based typing test, and some "scientific" way of evaluating SGML-ability on-the-spot. I'm willing to compile the huge amount of answers :) in the definitive, never before published and soon legendary "how-to-recrute-SGML-typists-FAQ", to be posted later on. TIA, \</Steven> stevenn@kluwer.be, key technical SGML/EPP dept. Kluwer Legal Publishers Belgium -------------------------------------------------------------------- Steven Noels @ Kluwer Legal Publishers Belgium \<stevenn@kluwer.be> Four11-homepage: \<a href="http://www.wp.com/StevenN"> Tech. Support SGML Dept. tel. +32/3/360.02.92 fax +32/3/360.04.74 snmail: Santvoortbeeklaan 21-25, B-2100 Deurne, BELGIUM -------------------------------------------------------------------- </message> <message id="<9511141759.AA10033@fly.HiWAAY.net>" date="3025360769" seqno="11431"> Newsgroups: comp.text.sgml Date: Tue, 14 Nov 1995 11:59:29 -0600 Message-ID: <9511141759.AA10033@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> Subject: Re: External Cameras This article is cross-posted from www-vrml@wired.com to comp-text-sgml because the issues of link and location design are also critical in SGML and HTML, and because we must consider the problems of compound document architectures. Those who work with many applications of SGML must begin to understand how these problems are handled in other notations and if the current proposals in HyTime and DSSSL can solve these for those whose systems are targeted to the integration of these in a single system. The initial posting dealt with the problems of accessing a VRML file to insert a camera view not originally provided by the author. [Michael Pichler] | Look at Hyper-G: here *all* links are kept outside the documents and | kept consistent by the server. VRML documents will be treated the same | way: we merge the links into the document when it is accessed via the | http protocol, so they look like ordinary links to the outside. Yes. I reviewed Hyper-G materials this year as part of a technical report to DoD pursuant to the MID project. What advantages do you obtain through this approach? My first guesses would be enhanced maintenance, reliable persistent links, capability to handle location ladders where required, dynamic views, etc. \<phil>The issues of simplicity (HTML) and robustness (SGML/HyTime) revolve around the user and their requirements. I am of the opinion that sacrificing system flexibility to the cause of preserving author's intent is a critically flawed design decision for many current systems although it appears to be a principled one. For most industrial applications of hypermedia to enterprise wide document construction and management, the artistry of the author is a hindrance, not an advantage. Documents which can be individually tailored are not conducive to concurrent integrated product development and often do not meet the stringent requirements for deliverables which are subject to review, validation, auditing, and occasional penalties for failure to comply. Now, ask me about what I care to put on an album I am producing, and my answer is the exact opposite, but in this, maintenance of technical information is not at issue. It seems to me that the questions unanswered too often in the discussions of the VRML language are who are the customers, and , what do the customers want to do with the language. Some systems must be designed for the heavy lifting problems of document management, while others are designed for the highly interactive environment of the world wide web. It is not likely that given the costs, these can be the same systems. If however, the VRML community understands that the language can be applied to both and will be, then it can set its architectural goals accordingly. Preservation of author's intent should not be the sine qua non of these goals. Portability and reusabiity are, IMO, more important. VRML is a notation that must work within a compound document architecture. Overall, this is a different ideal than that of the WWW applications although not in conflict with it. A critical mistake made in the second phase of some designs is to lose sight of who will use the language. It then contains many features which only an experienced programmer can successfully apply. The only caveat is that these features can be disregarded in profiles of the language. So, as VRML progresses, I hope the currently simple features are kept and that changes made to it extend these features without sacrificing simplicity. Otherwise, we get yet-another-object-whatsis that thrills the engineer while sacrificing the artists. A firm set of strongly stated and focused requirements could aid the development of VRML, but that does not seem to be the way this is going and the results thus far have been satisfactory. So I watch and learn. At this time, my preference is for systems like WebFX that support native browser behaviors that can be easily understood and applied. Systems that rely exclusively on applet-like or server-centric systems for behavior may be technically more extensible, but will be at an initial market disadvantage unless a very capable library is provided that is invisible. Look at the Netscape ads in Wired for an indicator of things to come. The user community is being deliberately set against the technical community by exploiting fairly base paranoias and stereotypes. It is opportunistic, but predatory and sets a tone for the future.\</phil> \<joke>Ask the gods for power, and they give wisdom. Ask the gods for wisdom, and they give power. Ask the gods for money, and they skip town. ;-)\</joke> I apologize for rambling on. Thanks for your information. Len Bullard </message> <message id="<30A8B306.50BB@pratique.fr>" date="3025351046" seqno="11432"> From: Dave Slosberg \<slos@pratique.fr> Newsgroups: comp.text.sgml Subject: Internet Assistant for MAC Date: Tue, 14 Nov 1995 16:17:26 +0100 Organization: Internet Way Message-ID: <30A8B306.50BB@pratique.fr> Can anyone tell me if the Micosoft Word Internet Assistant exists for the MAC ?? Thanx, Dave </message> <message id="<DI1ILn.JsA@world.std.com>" date="3025352938" seqno="11432"> Newsgroups: comp.text.sgml From: jbottoms@world.std.com (John W Bottoms) Subject: Re: Slight flame on SGML Message-ID: \<DI1ILn.JsA@world.std.com> Organization: The World Public Access UNIX, Brookline, MA References: <47d5ro$o3j@hamilton.maths.tcd.ie> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com> <488r3b$ruf@bell.maths.tcd.ie> Date: Tue, 14 Nov 1995 15:48:58 GMT In article <488r3b$ruf@bell.maths.tcd.ie>, Timothy Murphy \<tim@maths.tcd.ie> wrote: >jbottoms@world.std.com (John W Bottoms) writes: >>I disagree. SGML is designed to be just that! A meta-grammar. We haven't >>had one until now and it will take a while for people to understand that >>one is needed and how it is used. We have had plenty of grammars before >>but they didn't fulfill needs that exist. > >There is a well-defined mathematical notion of "grammar", >as described eg in M. Davis & E. Weyuker, >"Computability, Complexity and Languages", >and in many other similar works. >A grammar in this sense is defined by a set of "productions" >of terminals and variables, eg S -> abT. > >It seems to me that SGML is definitely a grammar in this sense, >and in fact probably belongs to the sub-class of context-free grammars. >-- >Timothy Murphy >e-mail: tim@maths.tcd.ie >tel: +353-1-2842366 >s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland My understanding is that all meta-grammars must be grammars. And the distinguishing difference between the two is that a meta-grammar contains rules for extending the set of base production rules (is this the concrete syntax->abstract syntax?). Are all context-free grammars also meta-grammars? I don't believe so. But some are. And by definition a meta-grammar must be mathematically proveable. I realize there have been earlier, proprietary meta-grammars but SGML is the first "ISO blessed" meta-grammar. -jb </message> <message id="<48adfg$bvi@hopper.acm.org>" date="3025352624" seqno="11433"> From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: sgmls and NAMELEN Date: 14 Nov 1995 15:43:44 GMT Organization: ACM Network Services Message-ID: <48adfg$bvi@hopper.acm.org> References: <47v99s$m1b@relay.lsl.co.uk> <480d3v$sse@crl2.crl.com> <488s82$69@bell.maths.tcd.ie>,<488v9v$3se@crl2.crl.com> Reply-To: davep@ACM.ORG In article <488v9v$3se@crl2.crl.com>, jenglish@crl.com (Joe English) writes: >Timothy Murphy \<tim@maths.tcd.ie> wrote: >>jenglish@crl.com (Joe English) writes: >>>SGML declarations tend to be rather long and not terribly >>>informative to human beings, so it's more convenient to keep >>>a single copy in a separate file [...] >>I found this out by painful experience. >>It seems completely undocumented in the SGML Handbook. As Joe pointed out; the Handbook isn't about any one entity manager's way of handling/storing entities. SGML just specifies that you should always send the SGML declaration as part of the document entity. >>Why don't .sgml files specify the .decl they expect? The ".sgml file" you're using is not a complete document; if the person who sent it to you didn't tell you what SGML declaration it's meant to work with, don't blame SGML-- blame the sender. If you only send a fragment, you need to send some form of environment information for that fragment, and the ISO standard doesn't address fragments. I understand that the new HyTime amendment does address fragments; SGML Open is also working the problem. Meanwhile, it's necessarily ad hoc. >The answer to that is in productions 171 through 204: >the blasted things are way too long -- the SGML declaration >for HTML is 93 lines -- and there's no way to include them >by reference. (This is one of SGML's most painful design >flaws, IMO.) I'm confident the forthcoming 8879 revision will address the problem of incorporating SGML declarations by reference, and modifying small parts thereof without having to rewrite the whole thing. (And don't ask "When?" and don't bitch that it isn't out yet. You probably haven't done anything substantive to help get it out.) Dave Peterson SGMLWorks! Principal, ANSI X3V1 US Technical Expert, ISO/IEC JTC1/SC18/WG8 davep@acm.org </message> <message id="<48advl$bvi@hopper.acm.org>" date="3025353141" seqno="11434"> From: davep@ACM.ORG Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 14 Nov 1995 15:52:21 GMT Organization: ACM Network Services Message-ID: <48advl$bvi@hopper.acm.org> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>,<488r3b$ruf@bell.maths.tcd.ie> Reply-To: davep@ACM.ORG In article <488r3b$ruf@bell.maths.tcd.ie>, tim@maths.tcd.ie (Timothy Murphy) writes: >It seems to me that SGML is definitely a grammar in this sense, >and in fact probably belongs to the sub-class of context-free grammars. SGML is certainly more than just a grammar, though certainly a grammar is used in its definition. That grammar definitely is not context-free. Simple example: how you parse a net ("/") depends on whether you are currently within an element started by a net-enabling start-tag. Dave Peterson SGMLWorks! davep@acm.org </message> <message id="<30A8CBC9.3B6@falch.no>" date="3025357385" seqno="11427"> From: Morten Gr�sby \<morteng@falch.no> Newsgroups: comp.text.sgml Subject: The whirlwind guide to SGML tools (was: Re: SGML Windows Editors) Date: Tue, 14 Nov 1995 09:03:05 -0800 Organization: Falch Infotek as Message-ID: <30A8CBC9.3B6@falch.no> References: <46m4ms$1p2@newsbf02.news.aol.com> \<LC05BCOW@news.uqam.ca> > if you go to http://www.falsh.no/~pepper/sgml-tools/, you'll find the > whirlwind guide. This fine guide is a good reference to different The correct URL is: http://www.falch.no/~pepper/sgmltool/ *--------------------E-mail: morteng@falch.no--------------------* Morten Gresby,SNAIL: Dagalivn.26,0387 Oslo, Norway @Falch Infotek as, Stansevn.21,PB130 Kalbakken,0902 Oslo,Norway. *--------------------http://www.falch.no/~morteng/----------------* </message> <message id="<48b8rc$qpd@manuel.anu.edu.au>" date="3025380652" seqno="11435"> From: Gavin Mercer \<gavin@sanger.anu.edu.au> Newsgroups: comp.text.sgml Subject: Program to take SGML source and produce a DTD Date: 14 Nov 1995 23:30:52 GMT Organization: Australian National University Message-ID: <48b8rc$qpd@manuel.anu.edu.au> Like it says does anyone have a program that generates a DTD (doesn't have to be brilliant ) from an SGMl document source ? PS I have ominmark -- --------------------- Australian National University --------------------- Gavin Mercer Email: Gavin.Mercer@anu.edu.au Australian National Phone: (06) 249 0476 Dictionary Centre Fax: (06) 249 0475 Centre for Networked Phone: (06) 279 8341 Information Publishing Fax: (06) 279 8120 </message> <message id="<48beml$n7p@peg.apc.org>" date="3025386645" seqno="11436"> From: carolp@ozlink.com (Carol Perkins) Newsgroups: comp.text,comp.text.desktop,comp.text.frame,comp.text.interleaf,comp.text.pdf,comp.text.sgml,comp.text.tex,comp.theory,comp.theory.cell-automata,comp.theory.dynamic-sys Subject: Miracle Herbal Smoking Cure 100% guaranteed to work Date: 15 Nov 1995 01:10:45 GMT Organization: Pegasus Networks Sender: interbiz@peg.apc.org Message-ID: <48beml$n7p@peg.apc.org> The Yin & Yang Anti-Smoking Programme. The only stop-smoking programme in the world which is guaranteed 100% to eliminate nicotine addiction. More Info At: http://www.ozlink.com/yinyang/index.htm </message> <message id="<48boj9$p7@maroon.tc.umn.edu>" date="3025396777" seqno="11437"> From: milor001@maroon.tc.umn.edu (R A Milowski) Newsgroups: comp.text.sgml Subject: Re: sgmls and NAMELEN Date: 14 Nov 1995 21:59:37 -0600 Organization: University of Minnesota Message-ID: <48boj9$p7@maroon.tc.umn.edu> References: <47v99s$m1b@relay.lsl.co.uk> <480d3v$sse@crl2.crl.com> <488s82$69@bell.maths.tcd.ie> Keywords: sgmls,NAMELEN In article <488s82$69@bell.maths.tcd.ie>, Timothy Murphy \<tim@maths.tcd.ie> wrote: >jenglish@crl.com (Joe English) writes: > [ lines deleted] >I just compiled "sp" on my Linux system -- >it was 10MB in size! >Admittedly when stripped it reduced to 550k -- >but that is still twice the size of Linux itself. >Is it really that big a task? >(Incidentally, sp did not compile under SunOS or Mips RiscOS. >The problem could be trivial -- I didn't spend time on it. >It seemed to be in the C++ code. I was using gcc/g++.) I believe that by default the makefiles are setup for debug builds in the distribution. Thus, if you don't turn off debugging, you will get enormous object code. SP uses quite a bit of templates which seems to cause the debugging information to be quite large. When you stripped the executible, the debugging information was removed. Hence, the difference in size. >SGML seems to me a very nice idea, very badly implemented. Actually, SP is quite well implemented. For very simple SGML, SP may have quite a few features that are not needed. On the other hand, it supports almost all of the SGML standard and handles mutiple coding systems quite well. It is the industrial strength SGML toolkit. I have uses SP quite extensively and it is *really* fast and complete. I'd give it a second chance. ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu -or- sgml@winternet.com Copernican Solutions Incorporated (612) 825 - 4132 -- ============================================================================== R. Alexander Milowski milor001@maroon.tc.umn.edu Copernican Solutions Incorporated (612) 825 - 4132 </message> <message id="<48c52q$kc1@news.Belgium.EU.net>" date="3025409574" seqno="11438"> From: stevenn@kluwer.be (Steven Noels) Newsgroups: comp.text.sgml Subject: Re: Miracle Herbal Smoking Cure 100% guaranteed to work Date: Wed, 15 Nov 1995 07:32:54 GMT Organization: Kluwer Legal Publishers Belgium Message-ID: <48c52q$kc1@news.Belgium.EU.net> References: <48beml$n7p@peg.apc.org> carolp@ozlink.com (Carol Perkins) wrote: >The Yin & Yang Anti-Smoking Programme. The only stop-smoking programme >in the world which is guaranteed 100% to eliminate nicotine addiction. >More Info At: http://www.ozlink.com/yinyang/index.htm Thanks, Carol Perkins. We appreciate your concern in our health. Is there some item in your produkt-line that takes care of major SGML/HTML/HyTime addiction? This could be of higher :) interest for this newsgroup. Bandwith? Who cares: after all, MSN, powered by Ham95, will give us the opportunity to go where we wanted today. \</Steven> ----------------------------------------------------------------------- Steven Noels @ Kluwer Legal Publishers Belgium \<stevenn@kluwer.be> Key Technical SGML/EPP Dept. tel. +32/3/360.02.92 fax +32/3/360.04.74 snmail: Santvoortbeeklaan 21-25, B-2100 Deurne, BELGIUM ----------------------------------------------------------------------- </message> <message id="<30A967E3.3CA8@intext.cpsg.com.au>" date="3025397347" seqno="11439"> From: Craig Fisher \<fisherc@intext.cpsg.com.au> Newsgroups: comp.text.sgml Subject: ANPA 1312 and IPTC 7901 formats? Date: Wed, 15 Nov 1995 14:09:07 +1000 Organization: InTEXT Systems Message-ID: <30A967E3.3CA8@intext.cpsg.com.au> Not directly related to SGML, but I couldn't find a more appropriate group. Can anyone tell me where/how I can get hold of the specifications for the IPTC 7901 (International Press Telecommunications Council) and ANPA 1312 (American Newspaper Publishers Association) text formats? I understand that these are to be superceded by the UTF SGML standard. Also, are any major newspaper publishers currently using UTF for distributing news? If so, who? Thanks very much, Craig -- Craig Fisher (fisherc@intext.cpsg.com.au) Product Development Manager InTEXT Systems Division CP Software Group Canberra, Australia </message> <message id="<48c58m$8ah@doc.jmu.edu>" date="3025409750" seqno="11440"> From: LSBUMGAR@vax1.acs.jmu.edu (lee s. bumgarner) Newsgroups: comp.text.sgml,comp.infosystems.www.advocacy Subject: Re: HTML The Beautiful Date: 15 Nov 1995 07:35:50 GMT Organization: Undertoad Hypermedia Publishing Co. Sender: LSBUMGAR@jmu.edu Message-ID: <48c58m$8ah@doc.jmu.edu> References: <9510261615.AA17590@fly.HiWAAY.net> <470ohr$f1s@toads.pgh.pa.us> <19951030T045926Z@naggum.no> In-Reply-To: erik@naggum.no's message of 30 Oct 1995 04:59:26 +0000 > > somebody will have to start thinking seriously about ways to decrease the > bandwidth costs of their ideas of what it takes to communicate effectively, > because for the first time in mass marketing history, the _more_ money you > pour into your high-profile graphics and imagery, the _less_ efficient your > communication becomes. if nothing else, I hope I have blown some > manager-droids' fuses with _that_ scenario. yes.....but > > fact is, there isn't enough bandwidth on this here Internet to allow all > the marketing-droids the visual control they were taught to want in the TV > and the newspaper/magazine eras of advertising. some serious reinvention > of their industry is about to happen. they just don't know it, yet. maybe > even the guys at the Ad Council will get the idea. No, there isn't enough bandwidth...but it is coming very very soon. Think of it like this. You want a paved parking lot for your local elementry school (kewl WWW goodies) but since no one can see the advantage of it (lack of bandwidth) it goes nowhere. You then change tactics and get the money and get the paved parking lot (using the current lack of bandwith to show kewl WWW goodies.) People suddenly go "Gee, that was good idea even though we are in debt." So the fork over the money to pay of the debt (ATM, ISDN, cable modems, and that new technology that allows copper wire to give almost as much bandwidth as a cable modem.) TA DA, problem solved. The question is, of course, what happens when the Net REALLY goes mainstream and you have every Tom Dick and Bubba using it. What happens to Net.Culture? Will the Net be able to handle 100 million people using it? I suspect besides pure greed, THAT fear is the reason why the <$1000 "Internet Appliance" is taking so long to create. The network people need time to create enough bandwidth for all those people. -l --- -----> Undertoad (under construction) http://falcon.jmu.edu/~bumgarls/ <-------- REALITY.SYS corrupted. Reboot universe (Y/N/Q)? | "Usenet is Frosty The Snowman committing suicide with a flame thrower."-- Kibo |Somebody visited a.r.k and all I got was this lousy .sig | M$'s Blackbird is EVIL, do not support it! | ##30## </message> <message id="<48bv00$79@doc.jmu.edu>" date="3025403328" seqno="11441"> From: LSBUMGAR@vax1.acs.jmu.edu (lee s. bumgarner) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 15 Nov 1995 05:48:48 GMT Organization: Undertoad Hypermedia Publishing Co. Sender: LSBUMGAR@jmu.edu Message-ID: <48bv00$79@doc.jmu.edu> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> In-Reply-To: bagwill@kangaroo.ncsl.nist.gov's message of 13 Nov 1995 17:04:21 GMT In <487tql$mus@dove.nist.gov> bagwill@kangaroo.ncsl.nist.gov writes: > In article <47tn8b$m7r@news.iastate.edu>, > choward@iastate.edu (Chris Howard) writes: > >HTML is a pointy stick SGML is Ada, HTML is MS-DOS batch files. > >The beauty of pointy sticks is that they're handy and easy to use. > > Nostradamus predicts: In a few more months, almost no HTML will be > hand-generated. Almost everyone will be using their favorite WYSIWYG > word processor, or one of the new web-oriented ones. Artists, who are > currently bitching about not being able to use HTML to do the kind of > layout they want to do, will be using some tool that will give them > output close enough to what they want. It may be HTML 4.0, it may be > PDF+, it may be Netscape HTML\<tm>. They will never have to learn what a > doctype or DTD is. Of course, it could be Blackbird, at which point I seriously think about how to do a EPS near M$ HQ 8-). What I think is going to happen is Adobe's "Amber" is really going to take off, thus preventing Blackbird from doing any serious damnage to the Web. > > Serious Web publishers, on the other hand, will be going nuts because it > will be almost impossible to give their publications a consistent > look-and-feel, since every artist will be using their own incompatible > tool, and because it will be impossible to automatically generate > tables-of-content, indexes, bibliographies, updated links, etc. So > publishers who ignored SGML the first time around will either rediscover > it, or will buy into proprietary tools that will let them control the > production of thousands of complex documents, or both (since > interoperability and interchange seem to be "a custom More honoured in > the breach than the observance." in the SGML world. Let me ask a question I've asked before, but try to be as clear and spacific as possible to make my question understandable: Currently, the Interchange online service is SGML based...From what I've seen of it, it looks a great deal like a Web page (given that SGML is a superset of HTML that make sense) So, how likely is it that in the future we would have a totally SGML/Ambur based web...or is HTML too intrenched to go away. I am aware of the fact that SGML is a superset of HTML, but my question remains the same. -l --- -----> Undertoad (under construction) http://falcon.jmu.edu/~bumgarls/ <-------- REALITY.SYS corrupted. Reboot universe (Y/N/Q)? | "Usenet is Frosty The Snowman committing suicide with a flame thrower."-- Kibo |Somebody visited a.r.k and all I got was this lousy .sig | M$'s Blackbird is EVIL, do not support it! | ##30## </message> <message id="<48c5k3$8ah@doc.jmu.edu>" date="3025410115" seqno="11442"> From: LSBUMGAR@vax1.acs.jmu.edu (lee s. bumgarner) Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 15 Nov 1995 07:41:55 GMT Organization: Undertoad Hypermedia Publishing Co. Sender: LSBUMGAR@jmu.edu Message-ID: <48c5k3$8ah@doc.jmu.edu> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> In-Reply-To: cbullard@HiWAAY.net's message of 10 Nov 1995 00:18:16 -0600 > > Reality check: they are a very long ways from having the clout to take over > Microsoft's position in the food chain. I'm not so sure...We are entering a new age of computing and M$ might be slow enough for Netscape to get pretty big. Esp. if cheap networked computers are created.... -l --- -----> Undertoad (under construction) http://falcon.jmu.edu/~bumgarls/ <-------- REALITY.SYS corrupted. Reboot universe (Y/N/Q)? | "Usenet is Frosty The Snowman committing suicide with a flame thrower."-- Kibo |Somebody visited a.r.k and all I got was this lousy .sig | M$'s Blackbird is EVIL, do not support it! | ##30## </message> <message id="<48cm3q$b8c@news.xs4all.nl>" date="3025427002" seqno="11443"> From: jang@xs4all.nl (Jan Grootenhuis) Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 15 Nov 1995 12:23:22 GMT Organization: XS4ALL, networking for the masses Message-ID: <48cm3q$b8c@news.xs4all.nl> References: <48b8rc$qpd@manuel.anu.edu.au> Gavin Mercer \<gavin@sanger.anu.edu.au> writes: >Like it says does anyone have a program that generates a DTD (doesn't have to >be brilliant ) from an SGMl document source ? It's called EASY-DTD or something close. As always. you can find it in ftp.ifi.uio.no Cheers, -- Jan Grootenhuis Kralenbeek 1873 1104 KJ AMSTERDAM The Netherlands Tel/fax (+31)20-699.89.66 Internet jang@xs4all.nl </message> <message id="<truly.1850.0008C7C7@lunemere.com>" date="3025414006" seqno="11444"> From: truly@lunemere.com (Truly Donovan) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: Wed, 15 Nov 1995 08:46:46 Organization: La Lunemere Message-ID: \<truly.1850.0008C7C7@lunemere.com> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> In article <48bv00$79@doc.jmu.edu> LSBUMGAR@vax1.acs.jmu.edu (lee s. bumgarner) writes: >Let me ask a question I've asked before, but try to be as clear and spacific >as possible to make my question understandable: Currently, the Interchange >online service is SGML based...From what I've seen of it, it looks a great >deal like a Web page (given that SGML is a superset of HTML that make sense) >So, how likely is it that in the future we would have a totally SGML/Ambur >based web...or is HTML too intrenched to go away. I am aware of the fact that >SGML is a superset of HTML, but my question remains the same. SGML is no more a "superset" of HTML than COBOL is a superset of the payroll program. HTML is an *application* of SGML. Maybe that's why your question isn't understandable. Truly Donovan </message> <message id="<48dc8c$7np@crl5.crl.com>" date="3025449676" seqno="11445"> From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: PI - and RE Date: 15 Nov 1995 10:41:16 -0800 Organization: Mentally ill venomous purists with strange emotional drives Message-ID: <48dc8c$7np@crl5.crl.com> References: <4885ts$b91@ernie.idi.oclc.org> magnus salgo \<magnus@ids.se> wrote: > >I am trying to find out what is the right way of handling >Process Instructions and RE. [...] > >a) RE before Process intructions >==================================== >RSthisRE >RS\<?PROC>textRE >==================================== >This is equal to >RSthis\<?PROC>textRE > >b) RE after a Process Instruction >==================================== >RStext before\<?PROC>RE >RSthis text includes a process instructionRE >==================================== >This is equal to >RStext before\<?PROC> this text includes a process instructionRE Those aren't the results I get with SGMLS: \<!DOCTYPE test [ \<!element test - - (L+)> \<!element L - - (#PCDATA)> ]> \<test> \<l>RE before PI: \<?pi>blah\</l> \<l>PI before RE:\<?pi> blah\</l> \</test> parses as: (TEST (L -RE before PI: ?pi -\\nblah )L (L -PI before RE: ?pi -\\nblah )L )TEST C In other words, both cases are returned as: text\<?PI>&#RE;text with the PI before the RE, regardless of the original order. The reason for this is in 7.6.1, "Record Boundaries" (p. 322): "Within content, an *RE* [...] is treated as data unless its presence can be attributed solely to markup. That is: a) [...] b) The last *RE* in an element is ignored if no data or proper subelement follows it. c) [...]" And on p. 324: "An *RE* is deemed to occur immediately prior to the first data or proper subelement that follows it (that is, after any intervening markup declaration, processing instruction, or included subelement." I don't quite understand the rationale behind the first rule, but the second rule is to make sure that PIs and included subelements -- those which are legal due to an inclusion exception -- etc., can be inserted anywhere in the document without affecting the element structure, and without requiring infinite lookahead. By the first rule, \<FOO> blah \</FOO> is equivalent to \<FOO>blah\</FOO> and by the second rule: \<FOO> blah \<?PI>\</FOO> is *also* equivalent (PIs and included subelements aren't counted as "true" content, in a sense). When a parser sees an RE, it doesn't know if it's the last RE or not until it sees data, a subelement, or the end-tag (possibly implied) for the current element. If the very next thing is a PI, etc., it "remembers" that an RE was present and will reinsert it as soon as it can tell that it was not the "last" one. This has the effect of moving REs after all PIs and included subelements that follow it. >My conclusion is that to get a space in a) between the words >then I have to add a space at the nend of the first line >or a space at the start of line 2. That shouldn't be the case -- the RE might get *moved*, but it won't get *deleted* when you insert a PI. Spaces and other separator characters are not affected at all by these rules. Are there any short reference maps in effect? What parser are you using? --Joe English jenglish@crl.com </message> <message id="<48da3n$ib2@netline-fddi.jpl.nasa.gov>" date="3025447479" seqno="11447"> From: pjs@euclid.JPL.NASA.GOV (Peter J. Scott) Newsgroups: comp.text.sgml Subject: Fixing up instance when changing DTDs Date: 15 Nov 1995 18:04:39 GMT Organization: Jet Propulsion Laboratory, NASA/Caltech Message-ID: <48da3n$ib2@netline-fddi.jpl.nasa.gov> Reply-To: Peter.J.Scott@jpl.nasa.gov We're doing a lot of work at the moment refining some DTDs in an iterative process that unfortunately requires us to have a valid instance as an example each time we iterate. So we have been modifying the instance by hand. Generally the changes are not substantial - adding an attribute to an element, adding an element to a content model, etc, so the edits are fairly simple if tedious. Are there any tools out there which might automate this given some simple description of how the DTD has changed? I realize this is asking a lot - I have an appreciation of what would be required to do this in general and it is mind-boggling - but it is worth asking. -- This is news. This is your | Peter Scott, NASA/JPL/Caltech brain on news. Any questions? | (Peter.J.Scott@jpl.nasa.gov) </message> <message id="<CONNOLLY.95Nov15185945@w3.org>" date="3025468785" seqno="11448"> From: connolly@w3.org (Dan Connolly) Newsgroups: comp.text.sgml,comp.infosystems.harvest Subject: Re: Hypermedia Addressing [was: Solving the BIG problem] Date: 15 Nov 1995 23:59:45 GMT Organization: Massachvsetts Institvte of Technology Message-ID: \<CONNOLLY.95Nov15185945@w3.org> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> \<CONNOLLY.95Nov12185209@w3.org> <816252263snz@sgml-cen.demon.co.uk> In-reply-to: Martin Bryan's message of Mon, 13 Nov 95 08:44:23 GMT In article <816252263snz@sgml-cen.demon.co.uk> Martin Bryan \<mtbryan@sgml-cen.demon.co.uk> writes: > > Let me add a brief plea to consider HyTime topic maps as a way of linking > documents with the same keywords, or which have been categorized in the > same way. Sounds interesting. Could you elaborate? Anybody else want to give an example etc.? Pointers to relavent reading material? Dan -- Daniel W. Connolly "We believe in the interconnectedness of all things" Research Scientist, MIT/W3C PGP: EDF8 A8E4 F3BB 0F3C FD1B 7BE0 716C FF21 \<connolly@w3.org> http://www.w3.org/hypertext/WWW/People/Connolly/ </message> <message id="<19951116T071452Z@naggum.no>" date="3025494892" seqno="11446"> From: Erik Naggum \<erik@naggum.no> Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 16 Nov 1995 07:14:52 +0000 Organization: Naggum Software; +47 2295 0313 Message-ID: <19951116T071452Z@naggum.no> References: <48b8rc$qpd@manuel.anu.edu.au> cc: Sunniva Solstrand \<sunnivas@falch.no> [Gavin Mercer] | Like it says does anyone have a program that generates a DTD (doesn't | have to be brilliant) from an SGMl document source? this is an interesting challenge, and Sunniva Solstrand \<sunnivas@falch.no> researched the topic for her Master's degree. written in Common Lisp, the system successfully generated good content models for the Wittgenstein Archives in Bergen, which was quite messy. unfortunately, for this to work at all, it must be brilliant. fortunately, her work is brilliant. the thesis itself is also written in an elite language, Norwegian. #\<Erik 3025494892> -- it is much more important to do it right than do get it done. </message> <message id="<tim-1511952305350001@slug-ppp17.wustl.edu>" date="3025487134" seqno="11449"> From: tim@michelob.wustl.edu (Tim Rand) Newsgroups: comp.text.sgml Subject: Searching for SEMA/OSTER Date: Wed, 15 Nov 1995 23:05:34 -0600 Organization: St. Louis Unix Users Group Message-ID: \<tim-1511952305350001@slug-ppp17.wustl.edu> I am searching for information on the product MARK-IT authored by SEMA in the UK. It was marketed by OSTER in the US. Can anyone tell me how to get info on this firm/product? Thanks in advance Tim Rand -- Tim Rand \<tim@michelob.wustl.edu> Checked regularly All messages will get a response. No black holes here! </message> <message id="<B3DlqvV.liora@delphi.com>" date="3025476717" seqno="11450"> From: liora@delphi.com Newsgroups: comp.text.sgml Subject: ABCD... SGML: A New Book Date: Wed, 15 Nov 95 21:11:57 -0500 Organization: Delphi (info@delphi.com email, 800-695-4005 voice) Message-ID: \<B3DlqvV.liora@delphi.com> I would like to announce the publication of my long-in- the-works book ABCD... SGML, published by International Thomson Computer Press. I have read this discussion group sporadically during the years I have worked on the book and it has always been my intention to create a resource for those who ask, periodically, Whoa, what is this all about? Where can I turn for a simple explanation and to figure out how it relates to my needs? I hope that I have provided a tool for these readers. As it turned out, it took me a few more pages than I anticipated to do this (my apologies to trees everywhere), but then, the subject deserves it. The book is 432 pp (well, exactly 100 are consumed in resource guides, appendices and front and back matter) and costs $39.95 (US). It includes a disk with a special build of SoftQuad's Panorama FREE that will load under Windows 3.1 with the book's SGML Resource Guide even if the user does not have a Web connection. With a Web connection, the Guide has live links to SGML-related Web resources. The disk includes two style sheets and two navigators which present a neat method of showing how presentation applications are independent of data coding. It is *not*, as a recent poster requested, "SGML for Dummies." I don't believe in writing for dummies, regardless of the subject. It is, I hope, a first book for all those interested in SGML -- programmers, writers, pubs managers, MIS. It is, I hope, an accessible book that considers the workplace changes that must be concurrent with the introduction of any new technology and creates a conceptual framework for understanding SGML production processes that will have value beyond this volume and the current round of tools and data designs. A writer well known to readers of this list is preparing a review of the book for comp.text.sgml. What follows in the publisher's Press Release. Excerpts are available at http://www.thomson.com/itcp under Samples. Soon -- how soon depends on Thomson -- the SGML and Panorama files from the disk that comes with the book will be posted as Thomson Web Extras at the same address. My appreciation to all those readers who helped me in the process of producing the book -- you are many -- I tried to mention each of you in the print volume, apologies to any who were inadvertently left out. I look forward to feedback on the volume -- please feel free to contact me directly at liora@delphi.com or through my publisher. Readers who would like to review the book for other electronic or print publications should contact the publisher (Kathleen Raftery kraftery@thomson.com or at the address below.) Best wishes, Liora Liora Alschuler The Word Electric POB 177 East Thetford, VT 05043 802/785-2623 liora@delphi.com =================================================== FOR IMMEDIATE RELEASE CONTACT: USA OUTSIDE THE USA: Kathleen Raftery Josette Garcia Tel. 617-348-8184 Tel. 44-171-497-1422 Fax 617-695-1615 Fax 44-171-497-1426 kraftery@thomson.com Josette.Garcia@itpuk.co.uk International Thomson Computer Press Announces the Release of ABCD... SGML A User's Guide to Structured Information by Liora Alschuler ISBN 1-850-32197 Boston, MA - International Thomson Computer Press announced today the publication of ABCD... SGML A User's Guide to Structured Information, the first book for managers, writers, and programmers that takes a system-wide approach to the application of SGML (Standard Generalized Markup Language) to publishing, information management and the Internet. This is the first SGML book that considers the collateral changes to the organization of work that accompany every successful transition to new technology. The book includes a disk that demonstrates SGML in action. The book provides an overview of the tools and processes that put SGML to work on real-world applications and describes over a dozen case studies of SGML in use in financial analysis; judicial administration; multimedia entertainment; scientific and academic research; electronic, print and technical publishing; and, not least of all, on the World Wide Web. It explains how to do a needs analysis that will tell you if SGML is appropriate for your situation. The book introduces SGML-based data design and system design and covers the transition to SGML and work in an SGML-based environment. ABCD... SGML explains SGML and SGML-based technology in such a way that anyone can read it and understand how to use SGML to: * create high quality electronic books and online databases * spin-off new products from current resources * eliminate redundant coding * add value to your information resources independent of any single application, vendor, or platform and insure against future data conversion costs * automate new areas of your enterprise not suitable for conventional database technology * comply with industry standards for information interchange ABCD... SGML places SGML in context as an open standard for data encoding that increases the power of current publishing and information management technology. The volume published by ITCP demonstrates SGML with a floppy disk containing ABCD... SGML Resource Guide A: Keeping Pace with SGML, a guide to SGML contacts that is coded in SGML and is bundled with SoftQuad's Panorama FREE browser. The files can be viewed as plain ASCII text or locally in Panorama for Windows or, with a Web connection, in Panorama with hot links to all the Web resources listed in the Guide. These resources include the best sites for information about SGML and the best sites for information in SGML. Special Note on SGML and HTML SGML is the parent language to HTML -- hypertext markup language -- the core data encoding system of the World Wide Web. That is, HTML is just one example of what you can do with SGML. HTML was designed as an easy- to-learn, lowest-common-denominator and it works well for short, simple documents on the World Wide Web. SGML allows you to put more information in your markup so that it describes all types of data with precision and flexibility. For more information on HTML and SGML and extracts from ABCD... SGML, including: * an annotated Table of Contents, * Preface and * Case Study of Microsoft's Cinemania, see our Web site at http://www.thomson.com/itcp.html In addition to the book extracts, you will soon find Keeping Pace with SGML in SGML, as a Thomson WebExtra. Liora Alschuler is a writer and consultant in SGML and technical publishing who has been writing about electronic text delivery systems since 1987. A frequent lecturer and trainer, she teaches an introductory course on SGML for the University of Wisconsin, Engineering Professional Development program. She is a Senior Member of the Society for Technical Communications and Associate Editor of Technical Communication for Computers and Communication where she recently edited two special sections on SGML. ******************************** International Thomson Computer Press is an innovative, leading publishing group focused globally on the needs of professionals. Our expert authors come from all corners of the world and work with us to bring new products and ideas to a community keeping pace with the latest technologies. ITCP titles can be purchased through major bookstore chains, independent book stores, college bookstores, over the World Wide Web, or at our Distribution Center at Tel. 800-842-3636. </message> <message id="<SCHULZE.95Nov16065012@amarant.ims.uni-stuttgart.de>" date="3025489812" seqno="11451"> From: schulze@amarant.ims.uni-stuttgart.de (Bruno Maximilian Schulze) Newsgroups: comp.text.sgml Subject: Re: Fixing up instance when changing DTDs Date: 16 Nov 1995 05:50:12 GMT Organization: IMS, University of Stuttgart, Germany Message-ID: \<SCHULZE.95Nov16065012@amarant.ims.uni-stuttgart.de> References: <48da3n$ib2@netline-fddi.jpl.nasa.gov> Reply-To: schulze@ims.uni-stuttgart.de In-reply-to: pjs@euclid.JPL.NASA.GOV's message of 15 Nov 1995 18:04:39 GMT >>>>> "Peter" == Peter J Scott \<pjs@euclid.JPL.NASA.GOV> writes: Peter> We're doing a lot of work at the moment refining some DTDs Peter> in an iterative process that unfortunately requires us to Peter> have a valid instance as an example each time we iterate. Peter> So we have been modifying the instance by hand. Generally Peter> the changes are not substantial - adding an attribute to an Peter> element, adding an element to a content model, etc, so the Peter> edits are fairly simple if tedious. Are there any tools out Peter> there which might automate this given some simple Peter> description of how the DTD has changed? I realize this is Peter> asking a lot - I have an appreciation of what would be Peter> required to do this in general and it is mind-boggling - Peter> but it is worth asking. This seems to be a task for the Revision Control System (RCS). It is available as part of GNU, so don't care about costs. It should be available on many ftp servers. If you have problems to find it, please let me know. Here are some lines of the rcsintro man page: The Revision Control System (RCS) manages multiple revisions of files. RCS automates the storing, retrieval, logging, identification, and merging of revisions. RCS is useful for text that is revised frequently, for example programs, docu- mentation, graphics, papers, and form letters. Hope this helps. Best, Max -- ------------------------------------------------------------------------------ Bruno Maximilian Schulze | Institute of Natural Language Processing Tel: +49-711-121-1367 | University of Stuttgart Fax: +49-711-121-1366 | Azenbergstr. 12 email:schulze@ims.uni-stuttgart.de | D-70174 Stuttgart, Germany URL: http://www.ims.uni-stuttgart.de/IMS.html ------------------------------------------------------------------------------ </message> <message id="<48e3s9$33u@ni1.ni.net>" date="3025473865" seqno="11452"> From: km Brooks \<kbrooks@jouve.com> Newsgroups: comp.text.sgml Subject: JDM SGML Specialist Position Description Date: 16 Nov 1995 01:24:25 GMT Organization: Jouve Data Management Message-ID: <48e3s9$33u@ni1.ni.net> Jouve Data Management has developed several innovative text/graphic retrieval systems for technical documentation, utilizing CD-ROM and online technology. We are currently seeking a systems analyst with SGML experience. I have attached a copy of the position description. If you are interested, please contact me via email, fax (714/474-4201), telephone 717/474-4200 ex 227 or mail: Jouve Data Management, 17671 Cowan Ave., Suite 200, Irvine, CA 92714. POSITION TITLE: SYSTEMS ANALYST DEPARTMENT: INFORMATION SYSTEMS GENERAL: Systems Analysts are responsible for application design and specification. Applications include SGML conversion for text files, graphic conversion, database preparation programs and procedures, data verification, and other projects as assigned. Individuals represent the department in meetings with marketing, customers and suppliers on product definition, improvements and resolution of problems. They also act as technical project managers, keeping track of project tasks and milestones. PRINCIPAL DUTIES AND RESPONSIBILITIES: =B7 Collect and analyze technical documentation for text and graphic source files. =B7 Identify specific customer requirements which may impact database and interface design. =B7 Provide and meet schedule estimates on design and development efforts. Provide timely updates on progress for management and tracking of progress. =B7 Design database layout, including DTDs and record presentation. =B7 Develop specifications in accordance with JDM standards. =B7 Develop test specifications for quality control and data verification. =B7 Represent the department in meetings with Product Development, customers and suppliers on product definition, improvements and resolution of problems =B7 Support Production Services staff in trouble-shooting and management of production process. CHARACTERISTICS: =B7 Work with general supervision and review. =B7 Act as technical project managers, keeping track of project tasks and milestones. =B7 Good analytical and logical thinking skills; ability to communicate complex ideas to non-technical and technical staff; good judgment and organizational ability. =B7 Operates with some latitude for unreviewed action or decision. QUALIFICATIONS: =B7 B.S. in Computer Science, Information Science or equivalent education. =B7 Effective communication and technical writing skills, directed to technical and non-technical audiences. =B7 2-5 years experience in data analysis and database design and specification. =B7 Experience with information retrieval system and/or database system data preparation. =B7 Demonstrable knowledge and application of SGML, including DTD development. SALARY RANGE $40,000 - 50,000 (negotiable, depending upon experience and expertise) </message> <message id="<48fb2d$sr5@news.Belgium.EU.net>" date="3025513997" seqno="11453"> From: Jacques Deseyne \<jad@sema.be> Newsgroups: comp.text.sgml Subject: Re: Searching for SEMA/OSTER Date: 16 Nov 1995 12:33:17 GMT Organization: SEMA Group Belgium Message-ID: <48fb2d$sr5@news.Belgium.EU.net> References: \<tim-1511952305350001@slug-ppp17.wustl.edu> tim@michelob.wustl.edu (Tim Rand) wrote: >I am searching for information on the product MARK-IT authored by SEMA in >the UK. It was marketed by OSTER in the US. Can anyone tell me how to get >info on this firm/product? > >Thanks in advance I should probably be able to do that. Mark-It was developed by Sema Group Belgium (at that time, Sobemap). Its first version existed as far ago as 1986. It was the first product (and I believe, still the only one) which supports ALL features of ISO 8879. A number of major vendors use either source code or object code from Mark-It inside their products. The distributor agreement with Oster was not continued We are finalizing a new release at this very moment. Instead of plugging our products here in this newsgroup, I will mail you a full description. You may be interested to learn that academic and research institutes can have a free license. Other possibly interested readers of CTS may direct their inquiries using the pointers given below. Best regards, ------------------------------------------------------ Jacques Deseyne \<jad@sema.be> Sema Group Belgium - Stallestraat 96 - B-1180 Brussels info-markit@sema.be info-sgml@sema.be ------------------------------------------------------ </message> <message id="<48fer6$9mu@cliffy.lfwc.lockheed.com>" date="3025517862" seqno="11454"> From: mcclellantj@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: sgmls and NAMELEN Date: 16 Nov 1995 13:37:42 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <48fer6$9mu@cliffy.lfwc.lockheed.com> References: <47v99s$m1b@relay.lsl.co.uk> <480d3v$sse@crl2.crl.com> <488s82$69@bell.maths.tcd.ie> Reply-To: mcclellantj@lfwc.lockheed.com Timothy Murphy (tim@maths.tcd.ie) wrote: : jenglish@crl.com (Joe English) writes: : >SGML declarations tend to be rather long and not terribly : >informative to human beings, so it's more convenient to keep : >a single copy in a separate file and pass this on the 'sgmls' : >command line, e.g.: : > sgmls [ other flags ] sgml.decl mydoc.sgml : I found this out by painful experience. : It seems completely undocumented in the SGML Handbook. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ The fact that an SGML declaration is _required_ is evident in the ISO-8879 itself (and therefore in the "SGML Handbook", pp 294-295): [1] SGML_document= SGML_document_entity, \ ( SGML_subdocument_entity | SGML_text_entity | character_data_entity | specific_character_data_entity| non-SGML_data_entity )* [2] SGML_document_entity= s*, SGML_declaration, \ prolog, document_instance_set, Ee : Why don't .sgml files specify the .decl they expect? As '.sgml' files aren't commonly defined, let me rephrase the question: Q: Why doesn't the DTD specify the SGML Declaration it expects? Because it may expect different decls for the _same_ DTD. For example, a workstation may have enough resources to allow a TAGLVL ( number of open tags allowed ) of 100, but a PC may only be able to handle a TAGLVL of 24. This may not be common, but it is possible. The specification should be made as general as possible, so it should handle this case (as it does). : This seems to me like a C program not specifying : what header files are required, This seems to me like a C program not specifying the maximum depth of the stack (which it does not specify). -- Tad McClellan, Logistics Specialist (IETMs and SGML guy) email: mcclellantj@lfwc.lockheed.com All I want, is a little more than I'll ever get. </message> <message id="<48fjhm$5mp@idefix.eunet.fi>" date="3025522388" seqno="11455"> From: kukkonen@index.fi (Petri Kukkonen) Newsgroups: comp.text.sgml Subject: Re: How to recrute SGML-typists? Date: Thu, 16 Nov 1995 14:53:08 GMT Organization: Index IT Oy Message-ID: <48fjhm$5mp@idefix.eunet.fi> References: <489r37$m8l@news.Belgium.EU.net> stevenn@kluwer.be (Steven Noels) wrote: >Beloved c.t.s.-ers, >I lurk, therefore Iam. Although we're using SGML since a few years in >our company, we are still searching for the definitive, ideal test for >SGML-inputters. We know they should have a fast type-rate, they should >be friendly & nice, and so on... >We're just wondering if anybody in this group has already compiled a >concise, one hour, ArborText ADEPT-Editor based typing test, and some >"scientific" way of evaluating SGML-ability on-the-spot. >I'm willing to compile the huge amount of answers :) in the >definitive, never before published and soon legendary >"how-to-recrute-SGML-typists-FAQ", to be posted later on. >TIA, >\</Steven> >stevenn@kluwer.be, key technical SGML/EPP dept. >Kluwer Legal Publishers Belgium >-------------------------------------------------------------------- >Steven Noels @ Kluwer Legal Publishers Belgium \<stevenn@kluwer.be> >Four11-homepage: \<a href="http://www.wp.com/StevenN"> >Tech. Support SGML Dept. tel. +32/3/360.02.92 fax +32/3/360.04.74 >snmail: Santvoortbeeklaan 21-25, B-2100 Deurne, BELGIUM >-------------------------------------------------------------------- Try FrameMaker+SGML... Programmers release is already out there. Petri Kukkonen ; Internet: kukkonen@index.fi Index IT Oy ; Phone : +358-0-461977 Otsolahdentie 8 D ; Telefax : +358-0-462849 FIN-02110 Espoo, Finland ; Mobile : +358-49-485946 ; WWW : www.index.fi \\|||/-----------------------------------------------------------> o - v </message> <message id="<9511161626.AA13700@fly.HiWAAY.net>" date="3025527988" seqno="11456"> Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Date: 16 Nov 1995 10:26:28 -0600 Message-ID: <9511161626.AA13700@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> <48c5k3$8ah@doc.jmu.edu> Subject: Re: Slight flame on SGML [Len Bullard] | Reality check: they are a very long ways from having the clout to take | over Microsoft's position in the food chain. [Lee Bumgarner] | I'm not so sure...We are entering a new age of computing and M$ might | be slow enough for Netscape to get pretty big. Esp. if cheap networked | computers are created.... Could be right, but here are some other thoughts to consider. 1. They rely quite heavily on the Microsoft technology, so I think the relationship is mutually supportive and only competitive at the application level. Microsoft is adept at keeping this kind of relationship going. I think the big Netscape stock boom is based on a bubble of hype and misperception, but I wish I had some of that stock right now. ;-) 2. Others with clout, market share, and expertise are entering the game. They also will innovate and gain loyalty. Neither Netscape nor TimBL invented the notion of neutral format document-centric computing. It's been with us and others have notions about how that can work better. So, expect some more seachanges. No, HTML will not go away, but SGML-complete systems will process it just like any other DTD. We may see a split-out of table types, forms, etc. into subdocuments as a more component-based approach is considered in which some of the components of a document are also negotiated similar to the MIME types. 3. The Netscape technology at present is based on a limited metaphor and implementation. It has not yet shown how it can support compound document architectures other than with plug-ins. This is a valid approach though. 4. We entered the new age of computing awhile back. It is called client-server computing on a global network. The Web is a manifestation of that, but not the only one. What happened was a new market and the marriage of two technologies client/server and SGML, not a new technology. Yes I know about JAVA and such. One question is, is this too hard for the target customer to master or even dangerous? An article in Scientific American labels it "the best virus ever invented". 5. The idea that people will "rent the software" of the future is interesting, but I wonder if that is similar to the "info is online; junk the printer" notion which has proved to be untrue. 6. Just as the cheap network computer has arrived, the painful limits of the Internet are beginning to show. A better infrastructure will appear, no doubt, but I just don't think it will be free. This is a "wait and see" position. Cheers, Len Bullard </message> <message id="<BERND.95Nov16160057@harlie.NeRo.Uni-Bonn.DE>" date="3025522857" seqno="11457"> From: bernd@NeRo.Uni-Bonn.DE (Bernd Kreimeier) Newsgroups: comp.text.sgml Subject: Re: Fixing up instance when changing DTDs Date: 16 Nov 1995 15:00:57 GMT Organization: University of Bonn, Dept. of Comp. Sc. VI, Germany Message-ID: \<BERND.95Nov16160057@harlie.NeRo.Uni-Bonn.DE> References: <48da3n$ib2@netline-fddi.jpl.nasa.gov> \<SCHULZE.95Nov16065012@amarant.ims.uni-stuttgart.de> In-reply-to: schulze@amarant.ims.uni-stuttgart.de's message of 16 Nov 1995 05:50:12 GMT In article \<SCHULZE.95Nov16065012@amarant.ims.uni-stuttgart.de> schulze@amarant.ims.uni-stuttgart.de (Bruno Maximilian Schulze) writes: > From: schulze@amarant.ims.uni-stuttgart.de (Bruno Maximilian Schulze) > > >>>>> "Peter" == Peter J Scott \<pjs@euc > Peter> edits are fairly simple if tedious. Are there any tools out > Peter> there which might automate this given some simple > Peter> description of how the DTD has changed? I realize this is > Peter> asking a lot - I have an appreciation of what would be > Peter> required to do this in general and it is mind-boggling - > Peter> but it is worth asking. > > This seems to be a task for the Revision Control System (RCS). It is > available as part of GNU, so don't care about costs. It should be > available on many ftp servers. If you have problems to find it, please You might want to look for CVS and PCL-CVS (Emacs frontend), which are using RCS but provide additional mechanisms. GNU as well, and AFAIR there's renewed support for CVS (read: somebody is reported to be working on a new version). > Hope this helps. Too. B. </message> <message id="<816536693snz@sgml-cen.demon.co.uk>" date="3025525493" seqno="11458"> Newsgroups: comp.text.sgml,comp.infosystems.harvest Subject: Re: Hypermedia Addressing [was: Solving the BIG problem] Message-ID: <816536693snz@sgml-cen.demon.co.uk> From: Martin Bryan \<mtbryan@sgml-cen.demon.co.uk> Date: Thu, 16 Nov 95 15:44:53 GMT Reply-To: mtbryan@sgml-cen.demon.co.uk References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> \<CONNOLLY.95Nov15185945@w3.org> Organization: The SGML Centre In article \<CONNOLLY.95Nov15185945@w3.org> connolly@w3.org "Dan Connolly" writes: > > Let me add a brief plea to consider HyTime topic maps as a way of linking > > documents with the same keywords, or which have been categorized in the > > same way. > > Sounds interesting. Could you elaborate? Anybody else want to give an > example etc.? Pointers to relavent reading material? The official contact point for this group is through subscribing to the newsgroup run from caph@techno.com. The real expert in Topic Maps is Michel Biezunski at High Text in Paris (michel@hightext.com). Alternatively attend one of the HyTime sessions organized by GCA. -- Martin Bryan @ The SGML Centre, Churchdown, Glos. GL3 2PU, UK (+44 1452 714029) </message> <message id="<30AB90BE.62B9@tdiinc.com>" date="3025538878" seqno="11459"> From: Bob Steele \<bsteele@tdiinc.com> Newsgroups: comp.text.sgml Subject: Transporting SGML fragments Date: Thu, 16 Nov 1995 11:27:58 -0800 Organization: Technology Deployment International Message-ID: <30AB90BE.62B9@tdiinc.com> at the risk of repeating a previous thread, i've heard of a reference describing the transport of an sgml fragment. can anyone help me find it? thanks in advance, bob -- Bob Steele, TDI Inc., 4100 Moorpark Ave, Suit 116, San Jose, CA, 95117 (408) 345 9157 mailto:bsteele@tdiinc.com http://www.tdiinc.com </message> <message id="<48g54e$8d3@news.INbe.net>" date="3025540686" seqno="11460"> From: "M.T. Carrasco Benitez" \<carrasco@innet.lu> Newsgroups: comp.text.sgml Subject: SGML -> HTML Date: 16 Nov 1995 19:58:06 GMT Organization: INnet NV (post doesn't reflect views of INnet NV) Message-ID: <48g54e$8d3@news.INbe.net> I am looking for SGML->HTML filter, or a pointer to the FAQ where this kind of answer could be. Mail prefered. Regards, Tomas </message> <message id="<48gns4$frm@bell.maths.tcd.ie>" date="3025559876" seqno="11461"> From: tim@maths.tcd.ie (Timothy Murphy) Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 17 Nov 1995 01:17:56 -0000 Organization: Dept. of Maths, Trinity College, Dublin, Ireland. Message-ID: <48gns4$frm@bell.maths.tcd.ie> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>,<488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> davep@ACM.ORG writes: >SGML is certainly more than just a grammar, though certainly a grammar >is used in its definition. That grammar definitely is not context-free. >Simple example: how you parse a net ("/") depends on whether you are >currently within an element started by a net-enabling start-tag. That certainly does not show that SGML is not a context-free grammar (although I now doubt if it is). A context-free grammar can be defined by a pushdown automaton, which can be any one of a number of states when examining a particular input token. However, the basic point is that SGML seems to be a grammar, and as such could be formally defined in a very small space. I don't understand why this wasn't done. -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland </message> <message id="<DI625t.F6M@curia.ucc.ie>" date="3025565094" seqno="11462"> Newsgroups: comp.text.sgml From: pflynn@www.ucc.ie (Peter Flynn) Subject: Re: SGML book? Message-ID: \<DI625t.F6M@curia.ucc.ie> Sender: usenet@curia.ucc.ie Reply-To: pflynn@www.ucc.ie Organization: UCC References: <480i5v$hsp@usenet.ucs.indiana.edu> Date: Fri, 17 Nov 1995 02:44:54 GMT Dixit pyachnes@ucs.indiana.edu (paul alan yachnes): > Can someone recommend the best book for learning SGML? The best one I can recommend is "ABCD... SGML: A User's Guide to Structured Information" by Liora Alschuler (ISBN: 1-850-32197-3) just published by ITCP. Details at http://www.thomson.com/itcp/liora/LA_HOME.HTM ///Peter </message> <message id="<vkeenan-1611952306040001@vern_ppp.lanminds.com>" date="3025577164" seqno="11463"> From: vkeenan@us.oracle.com (Vernon Keenan) Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: Thu, 16 Nov 1995 23:06:04 -0700 Organization: Oracle Corporation Message-ID: \<vkeenan-1611952306040001@vern_ppp.lanminds.com> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> <48c5k3$8ah@doc.jmu.edu> <9511161626.AA13700@fly.HiWAAY.net> In article <9511161626.AA13700@fly.HiWAAY.net>, Len Bullard \<cbullard@HiWAAY.net> wrote: > [Len Bullard] > > | Reality check: they are a very long ways from having the clout to take > | over Microsoft's position in the food chain. > > [Lee Bumgarner] > > | I'm not so sure...We are entering a new age of computing and M$ might > | be slow enough for Netscape to get pretty big. Esp. if cheap networked > | computers are created.... > > Could be right, but here are some other thoughts to consider. > \<snip> > 5. The idea that people will "rent the software" of the future is > interesting, but I wonder if that is similar to the "info is online; > junk the printer" notion which has proved to be untrue. > > 6. Just as the cheap network computer has arrived, the painful limits of > the Internet are beginning to show. A better infrastructure will > appear, no doubt, but I just don't think it will be free. This is a > "wait and see" position. > > Cheers, > > Len Bullard But, isn't it interesting that Oracle, along with IBM, Netscape, Sun and others, have made the challenge to Microsoft? Not everyone has signed up with Larry Ellison's Network Computer or George Gilder's "photons will be free" ideas, yet, but I think these inspirational thoughts have the potential to generate a lot of momentum in the general population and the worldwide economy. The evidence is increasing that these ideas will drive the next generation of consumer computer devices and have a great effect on the daily lives of everyone living in a technological society. Vern -- +----------------------------------------------------------+ | Vernon Keenan | 500 Oracle Parkway | | Product Marketing Manager | Web/Workgroup Division | | Oracle Corporation | Redwood Shores, CA 94065 | | E-mail: vkeenan@us.oracle.com | Phone: (415) 506-4651 | | WWW: http://www.oracle.com/ | +----------------------------------------------------------+ </message> <message id="<48hi3i$1bo@news.Austria.EU.net>" date="3025586700" seqno="11464"> From: gscheder@MANZ.co.at (Georg Scheder) Newsgroups: comp.text.sgml Subject: Stellenanzeige fuer SGML-Fachkraft Date: Fri, 17 Nov 1995 08:45:00 GMT Organization: MANZ Verlags- und Universitaetsbuchhandlung GmbH. Message-ID: <48hi3i$1bo@news.Austria.EU.net> Wir sind ein fuehrender oesterreichischer Fachverlag und suchen zum baldigen Eintritt *** eine(n) Mitarbeiter(in) fuer Aufbau, Pflege und Weiterentwicklung einer SGML-Textdatenbasis ***. Sie sind geuebt im Umgang mit einem PC und eine konzentrierte, genaue Arbeitsweise liegt Ihnen im Blut. Idealerweise verfuegen Sie au�erdem ueber folgende Kenntnisse: * PC-Betriebssysteme (DOS und Windows) * SGML/HTML (DTDs, Kodierung und Editoren) * Programmierung (Perl, ev. TCL/TK) Wenn Sie an einer langfristig ausbaubaren Stellung als SGML-Fachkraft in einem kleinen, initiativen Team interessiert sind, senden Sie Ihre aussagekraeftigen Bewerbungsunterlagen an MANZ Verlag, DI Georg Scheder, Kohlmarkt 16, 1014 Wien. </message> <message id="<199511171112.MAA06733@cygne.ais.berger-levrault.fr>" date="3025595534" seqno="11465"> Newsgroups: comp.text.sgml Date: 17 Nov 1995 12:12:14 +0100 From: "St�phane Bornnerand" \<sbor@Berger-Levrault.fr> References: <48da3n$ib2@netline-fddi.jpl.nasa.gov> Message-ID: <199511171112.MAA06733@cygne.ais.berger-levrault.fr> Subject: Re: Fixing up instance when changing DTDs [Peter J. Scott] > We're doing a lot of work at the moment refining some DTDs in an > iterative process that unfortunately requires us to have a valid instance > as an example each time we iterate. So we have been modifying the > instance by hand. Generally the changes are not substantial - adding an > attribute to an element, adding an element to a content model, etc, so > the edits are fairly simple if tedious. Are there any tools out there > which might automate this given some simple description of how the DTD > has changed? I realize this is asking a lot - I have an appreciation of > what would be required to do this in general and it is mind-boggling - > but it is worth asking. You could use software to automatically transform your old SGML instance into the new one. In fact, the software provides you a powerful and easy way to specify what to add or what to delete and by default echoing the rest of the SGML instance from the source file (old instance) to the target (new instance). AIS is developping BALISE(TM) and is shipping Release 3 of BALISE(TM) product which is an SGML document processing language with usual features to handle text and specialized features to handle SGML structures under various representations such as events and trees. The main characteristics of Balise could be listed as follows: � Compiled/interpreted language � Automatic memory management � Portable (source and binary level) � Modular and Extensible Architecture (5 set of APIs) � C-like syntax with object-oriented declaration � General purpose with SGML knowledge buit-in at the core � Mixed SGML document processing mode (event-driven, Tree navigation) � SGML Context handling based on both backward and look-ahead � Source level debugger To illustrate how BALISE could address your needs, I enclose the simplest Balise programe. It shows how it is easy to iteratively upgrade an SGML instance with few modifications at each step from this exemple which is used to just copy an SGML instance. default { on start cout << echo; // start tag with attrs on end cout << echo; // end start if not empty } content (CDATA|SDATA|PI) { cout << echo; // text content } and if you want to insert an attribute STATUS with the value PARTIAL to an element COMPONENT, you should add the next lines: // Quick and dirty solution element COMPONENT { on start cout << echo.replace (0,">$","STATUS = PARTIAL"); on end cout << echo; } or // Structured and readable solution element COMPONENT { on start { cout << format("<%1", elemName()); // copy the name COMPONENT for i,v in attr[] do { cout << format(" %1 = %2", i, v); // copy each pair attribute = value } cout << "STATUS = PARTIAL" + ">"; // add a new pair STATUS = PARTIAL } on end cout << echo; } Hope to reply to your request. For further information, contact Isabelle Bornier: phone: [+33] 1 46 40 84 15 FAX: [+33] 1 46 40 84 10 email: ibor@ais.berger-levrault.fr Regards, Stephane Bornerand. _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/ Stephane Bornerand phone: [+33] 1 46 40 84 06 _/ _/ AIS S.A. FAX: [+33] 1 46 40 84 10 _/ _/ 35, Rue du Pont email: sbor@ais.berger-levrault.fr _/ _/ 92200, Neuilly, FRANCE WWW: http://ais.berger-levrault.fr _/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ </message> <message id="<199511171131.LAA02593@sun1.cup.cam.ac.uk>" date="3025596196" seqno="11466"> From: Paul Treadaway \<pt100@cup.cam.ac.uk> Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: Fri, 17 Nov 95 11:23:16 GMT Organization: Cambridge University Press Message-ID: <199511171131.LAA02593@sun1.cup.cam.ac.uk> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>, <488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> In-Reply-To: <48advl$bvi@hopper.acm.org> In article <48advl$bvi@hopper.acm.org>, davep@ACM.ORG says... >In article <488r3b$ruf@bell.maths.tcd.ie>, > tim@maths.tcd.ie (Timothy Murphy) writes: >>It seems to me that SGML is definitely a grammar in this sense, >>and in fact probably belongs to the sub-class of context-free grammars. > >SGML is certainly more than just a grammar, though certainly a grammar >is used in its definition. That grammar definitely is not context-free. >Simple example: how you parse a net ("/") depends on whether you are >currently within an element started by a net-enabling start-tag. Are you sure? Surely if SGML couldn't be represented as a CFG then it couldn't be deterministically parsed in polynomial time? This is all rather beside the point - SGML is a grammar for writing grammars really. A DTD is a grammar, but need not be extensible. However SGML also provides a universal syntax for both DTD's and instances, so the boundary between metagrammar and grammar is blurred. The problem is that these are very difficult things to explain to people who have no background in mathematics or computer science. And I'm not convinced that it will ever be easy. People can cope with concepts like the internet by analogy, they can understand at some level the idea of a file format, and even tagged data, but even if you could explain a grammar in the Chomsky heirarchy sense, I think SGML is just too hard to convey to non-specialists, which is why non- technical issues are so frequently what people fall back on to make the case for SGML. </message> <message id="<48i0qh$16e@news.Belgium.EU.net>" date="3025601809" seqno="11467"> From: Jacques Deseyne \<jad@sema.be> Newsgroups: comp.text.sgml Subject: Re: SGML -> HTML Date: 17 Nov 1995 12:56:49 GMT Organization: SEMA Group Belgium Message-ID: <48i0qh$16e@news.Belgium.EU.net> References: <48g54e$8d3@news.INbe.net> "M.T. Carrasco Benitez" \<carrasco@innet.lu> wrote: >I am looking for SGML->HTML filter, or a pointer to the FAQ where this >kind of answer could be. Mail prefered. > >Regards, Tomas > A quite good way is to do it through EXPLICIT LINKs. In general, the cross-conversion needs are quite simple and it takes not too many LINK sets to achieve a remarkably good result. The advantage is that you use mechanisms which are defined within ISO 8879 itself and not one or another proprietary application language. Alas, only a few tools support LINKs. Best regards, ------------------------------------------------------ Jacques Deseyne \<jad@sema.be> Sema Group Belgium - Stallestraat 96 - B-1180 Brussels ------------------------------------------------------ </message> <message id="<48i20d$hca@hamilton.maths.tcd.ie>" date="3025603021" seqno="11468"> From: tim@maths.tcd.ie (Timothy Murphy) Newsgroups: comp.text.sgml Subject: Re: sgmls and NAMELEN Date: 17 Nov 1995 13:17:01 -0000 Organization: Dept. of Maths, Trinity College, Dublin, Ireland. Message-ID: <48i20d$hca@hamilton.maths.tcd.ie> References: <47v99s$m1b@relay.lsl.co.uk> <480d3v$sse@crl2.crl.com> <488s82$69@bell.maths.tcd.ie> <488v9v$3se@crl2.crl.com> Keywords: sgmls,NAMELEN jenglish@crl.com (Joe English) writes: >>Why don't .sgml files specify the .decl they expect? >The answer to that is in productions 171 through 204: >the blasted things are way too long -- the SGML declaration >for HTML is 93 lines -- and there's no way to include them >by reference. (This is one of SGML's most painful design >flaws, IMO.) Thanks for your generally helpful replies. But in this case all I was looking for was a _comment_ in the DTD to specify what declaration file was expected, and hopefully where one can get it. Surely that would not be too much to ask? -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland </message> <message id="<9511171331.AA12385@fly.HiWAAY.net>" date="3025603908" seqno="11469"> Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Date: 17 Nov 1995 07:31:48 -0600 Message-ID: <9511171331.AA12385@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> Subject: Re: Slight flame on SGML [Vernon Keenan] | But, isn't it interesting that Oracle, along with IBM, Netscape, Sun | and others, have made the challenge to Microsoft? In what form has the challenge been issued? In what form has it been answered? It's not that I don't believe it, it's just that I'm awfully leery of such as it seems every year I hear that someone is "laying down the gauntlet" with some initiative or another, yet they fade. So, what is this challenge? | Not everyone has signed up with Larry Ellison's Network Computer or | George Gilder's "photons will be free" ideas, yet, but I think these | inspirational thoughts have the potential to generate a lot of momentum | in the general population and the worldwide economy. Again, I am a skeptic. I watched the "inspirations" of Steve Jobs, etc, drive the industry down some dead ends and into some good technology. Not too much changed until the technology finally provided functionality that people really did want to use. The numbers of people in the US signed up to the Internet still remains a low percentage although growing daily. My guess is until the set top box arrives and the cable is the data hose, those numbers will peak. | The evidence is increasing that these ideas will drive the next | generation of consumer computer devices and have a great effect on the | daily lives of everyone living in a technological society. I agree that a change is afoot. I don't think the ideas of Ellison and Gilder are driving them though. I think certain technologies are converging and their potential for crystallizing markets is being realized. I am truly wary of the hyped prophecies though because they seem to make better press than sales, and it is the sale of systems that keeps the technology train on the track. Perception does drive the user into the store. On this we agree completely. Thanks. Len Bullard </message> <message id="<48inf9$cu1@newsbf02.news.aol.com>" date="3025625001" seqno="11470"> From: isipres@aol.com (ISI Pres) Newsgroups: comp.text.sgml Subject: Re: Transporting SGML fragments Date: 17 Nov 1995 14:23:21 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <48inf9$cu1@newsbf02.news.aol.com> References: <30AB90BE.62B9@tdiinc.com> Reply-To: isipres@aol.com (ISI Pres) Hi Bob; I'm not aware of the specific reference to which you refer but we have a fair amount of experience in using SGML fragments. In general, the problem is this: There is, strictly speaking, no such thing as a valid SGML fragment. This is because the requirements for a valid SGML instance include the doctype, front matter and full, unbroken, context from top level to whatever fragment one desires to transport. Thus, the fragment, while it is isolated from its full instance, is just data with some arcane tags in it. That this is structurally invalid does not mean that it can't or shouldn't be done, just that it occurs outside the conceptual world of SGML. The key to this transport is that the fragment, like a puzzle piece, must have a structurally valid home at its destination. In the puzzle analogy, the home would be another copy of the same puzzle into which the wayward piece could be reinserted. Only in this way could the piece (fragment) regain its full meaning and context. With SGML fragments, the challenge is the same. There must be a fully valid structural SGML instance on the receiving end into which the fragment can be reinserted. Of course, if the data inside the fragment itself can be assumed to have standalone value, it can be treated as tagged data and processed as desired; it is in this instance, no longer SGML. In most cases one can visualize, there would not always be a full instance on the receiving end, but there must be a full DTD with which the fragment is compatible (the original or a version of same). In such a case, the key to the value of the fragment is the software on the receving end. If the fragment is all you have, the software must be capable of inferring its context so that the rules of the DTD can be applied. Some software, editors from Arbor Text, Softquad for instance, can do this and work quite nicely. This accomplished, the fragment can reassume it full character and be processed as if it were still inside its home instance. The one major caveat to this capability is the fact that the context of a fragment must be unambiguous if the software is to handle it properly. For example, if the fragment is a paragraph and paragraph can exist in more than one brance of a DTD, then the receiving software will not be able to properly infer its context without more information. In such a case, the fragment cannot be transported and used in a manner similar to its inclusion in its source instance. This problem can be ameliorated to some extent if the same software is used on both ends. For example, if Arbor Text is the creating and receiving software, it can be set up to create a "processing instruction" containing the proper context of the fragment and to read in and act upon that PI at the receiving end. The problem here, of course, is that the flexibility of the fragment is reduced and the environment becomes application specific. Hope this helps, Best regards, Barry Schaeffer INFORMATION STRATEGIES, INC. </message> <message id="<DI79G4.B42@udcf.gla.ac.uk>" date="3025621011" seqno="11471"> Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Message-ID: \<DI79G4.B42@udcf.gla.ac.uk> From: flavell@v2.ph.gla.ac.uk (Alan J. Flavell) Date: Fri, 17 Nov 1995 18:16:51 GMT Reply-To: flavell@v2.ph.gla.ac.uk Sender: news@udcf.gla.ac.uk (News) References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov>,<48bv00$79@doc.jmu.edu> Organization: High Energy Physics - Glasgow University - UK >In <487tql$mus@dove.nist.gov> bagwill@kangaroo.ncsl.nist.gov writes: >> >> Nostradamus predicts: In a few more months, almost no HTML will be >> hand-generated. Almost everyone will be using their favorite WYSIWYG >> word processor, or one of the new web-oriented ones. Artists, who are >> currently bitching about not being able to use HTML to do the kind of >> layout they want to do, will be using some tool that will give them >> output close enough to what they want. I put it to you that you are peering into the wrong end of the telescope. It does not in the least matter what the document looks like to its author. Neither when the HTML is hand coded, nor when it is generated by some kind of "what you see won't be what they get" interface. Remember, users read those documents by means of browsers. On many different platforms. I assure you, what they see is not what you see. Except in a small proportion of cases. { Unless, of course, you are writing for a captive audience, for whom you can enforce the capabilities of their platform, in regard to hardware, software, and configuration. But that is not the World Wide Web, where you must take your readers as you find them. } >> ... it will be impossible to automatically generate >> tables-of-content, indexes, bibliographies, updated links, etc. So >> publishers who ignored SGML the first time around will either rediscover >> it, or will buy into proprietary tools that will let them control the >> production of thousands of complex documents, or both (since >> interoperability and interchange seem to be "a custom More honoured in >> the breach than the observance." in the SGML world. Well, either that or rediscover the original intentions of HTML, i.e as marking up the structure of information. --- Alan | http://ppewww.ph.gla.ac.uk/%7Eflavell/ "Aberglaube bringt Unglueck" </message> <message id="<48ihmj$bge@newsbf02.news.aol.com>" date="3025619091" seqno="11472"> From: isipres@aol.com (ISI Pres) Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 17 Nov 1995 12:44:51 -0500 Organization: America Online, Inc. (1-800-827-6364) Sender: root@newsbf02.news.aol.com Message-ID: <48ihmj$bge@newsbf02.news.aol.com> References: <19951116T071452Z@naggum.no> Reply-To: isipres@aol.com (ISI Pres) I may be missing something but, given the fact that SGML content is supposed to be structured according to a DTD, how would one come into the possession of "SGML Source" for which a DTD must be created? In the rare instance in which data is available but the DTD is no longer still around, one might try to reverse engineer one, but the caveat would be that such inferrence would, at best, create a very tenuous picture of the intent of the original creators. Barry Schaeffer INFORMATION STRATEGIES, INC. </message> <message id="<CONNOLLY.95Nov17175945@w3.org>" date="3025637985" seqno="11473"> From: connolly@w3.org (Dan Connolly) Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 17 Nov 1995 22:59:45 GMT Organization: Massachvsetts Institvte of Technology Message-ID: \<CONNOLLY.95Nov17175945@w3.org> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>,<488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> <48gns4$frm@bell.maths.tcd.ie> In-reply-to: tim@maths.tcd.ie's message of 17 Nov 1995 01:17:56 -0000 In article <48gns4$frm@bell.maths.tcd.ie> tim@maths.tcd.ie (Timothy Murphy) writes: > However, the basic point is that SGML seems to be a grammar, > and as such could be formally defined in a very small space. > I don't understand why this wasn't done. Some definitions: A is an alphabet iff A is a finite set. w is a string iff w is a finite sequence over an alphabet. L is a language iff L is a set of strings G is a grammar iff: G is a tuple (start, symbols, terminals, productions) start is a non-terminal symbol (it's in the set-difference of symbols and terminals) symbols is an alphabet terminals is a subset of symbols productions is a set of pairs (rhs, lhs) where rhs and lhs are strings over symbols (usually written rhs -> lhs) xSy derives xwy in G one step iff s -> w is a production in G x derives w in G iff there are strings y1 ... yn such that x derives y1 in one step in G y1 derives y2 in one step in G y2 derives y3 in one step in G ... yn derives w in one step in G w is in L(G), the language generated by G, iff s derives W in G where s is the (string consisting of the) start symbol in G. G is a context-free grammar iff the rhs of all productions are length-one strings consisting of non-terminals. L is a context-free language iff L is the languge generated by some context-free grammar. By those definitions, the grammar given in the SGML spec is context-free. A transcription of the grammar is available at: ftp://ftp.ifi.uio.no/pub/SGML/productions (thanks to Erik Naggum) Here's an example SGML document and its derivation: Document: \<!doctype simple> \<simple>abc\&tm;def\</simple> Derivation: SGML document -> SGML document entity SGML document entity -> SGML declaration, prologue, document instance set SGML declaration -> prologue -> document type declaration, other prolog document type declaration -> mdo ("<!"), "DOCTYPE", ps, document type name, mdc (">") ps -> s s -> SPACE (" ") document type name -> generic identifier generic identifier -> name name -> name start character, name character*5 name start character -> LCLetter ("s") name character*5 -> LCNMCHAR*5 ("i", "m", "p", "l", "e") other prolog -> s s -> RE ("\\n") document instance set -> base document element base document element -> document element document element -> element element -> start-tag, content, end-tag start-tag -> stago ("<"), document type specification, generic identifier specification, attribute specification list, tagc ("<") document type specification -> generic identifier specification -> generic identifier generic identifier -> name name -> name start character, name character*5 name start character -> LCLetter ("s") name character*5 -> LCNMCHAR*5 ("i", "m", "p", "l", "e") attribute specificaton list -> content -> mixed content mixed content -> data character*3, other content, data character*3 data character*3 -> SGML character SGML character*3 -> DATACHAR ("a", "b", "c") other content -> general entity reference general entity reference -> ero ("&"), name, reference end name -> name start character, name character*1 name start character -> LCLetter ("t") name character -> LCNMCHAR ("m") reference end -> refc (";") data character*3 -> SGML character SGML character*3 -> DATACHAR ("d", "e", "f") end-tag -> etago ("</"), document type specification, generic identifier specification, tagc ("<") document type specification -> generic identifier specification -> generic identifier generic identifier -> name name -> name start character, name character*5 name start character -> LCLetter ("s") name character -> LCNMCHAR ("i", "m", "p", "l", "e") Note that the terminals in the grammar are not characters, but lexical tokens: the grammar refers to stago, which is the lexical token corresponding to the string "<". The "meta" in "SGML is a meta-grammar" means that an SGML document itself contains a grammar. For example, here's a document with its DTD included explicitly, rather than by reference: \<doctype Memo [ \<element Memo - - (Salutation, P*, Closing?)> \<element Salutation O O (Date & To & Address?)> \<element (P|Closing|Date|To|Address) - O (#PCDATA)> ]> \<Memo> \<Date>June 12, 1994 \<To>Third Floor \<p>Please limit coffee breaks to 10 minutes. \<Closing>The Management \</Memo> A DTD is somewhat more complex than the definition of grammar I gave above -- its right-hand-sides can include operators like *, ?, &, etc. But those operators can be eliminated by introducing new rules, yielding a context-free grammar. For example, x -> y* becomes x -> z z -> /* empty */ z -> y z [Hmmm... inclusion and exclusion exceptions complicate this part... I'm not sure if they extend the expressive capability of a DTD beyond the realm of context-free grammars or not. Anybody else remeber? Anyway... ] The set of terminals in the DTD-grammar includes data-character, plus \<x> and \</x> for each element in the DTD. (entities and marked sections are taken care of by the lexical analyzer) So here's a derivation of the instance of the document with respect to the grammar given in the DTD: Memo -> \<Memo>, Salutation, P, Closing, \</Memo> Salutation -> Date, To Date -> \<Date>, "June 12, 1994" To -> \<To>, "Third Floor" P -> \<P>, "Please limit coffee breaks to 10 minutes." Closing -> \<Closing>, "The Management" Now: the language generated by that grammar is not _exactly_ the same as the set of instances that conform to the DTD: the tag inference rules create exceptions, and I've simplified away attributes (which interact with the grammar in the case of CONREF). Finally, SGML is not exactly the language generated by the grammar consisting of the productions in the standard. For example, the following string (well... the lexical tokens that it represents) is in that grammar, but it is not an SGML document: \<!doctype x [ mdo DOCTYPE name DSO \<!element x - O ANY> mdo ELEMENT name ??? ??? name mdc ]> msc mdc \<y>\</y> stago name tagc etago name tagc Then one might ask whether SGML is context-free at all; i.e. is there _any_ context-free grammar that generates SGML? No. For a proof, consider the above example and consult the proof in Louis & Papadimitriou using the "pumping theorem" to show that the language consisting of strings of the form x-x isn't context-free. The same proof shows that the C programming language isn't context-free, even though there's a handy context-free grammar that nearly describes it. Dan -- Daniel W. Connolly "We believe in the interconnectedness of all things" Research Scientist, MIT/W3C PGP: EDF8 A8E4 F3BB 0F3C FD1B 7BE0 716C FF21 \<connolly@w3.org> http://www.w3.org/hypertext/WWW/People/Connolly/ </message> <message id="<30ACEA76.630@cbvcp.com>" date="3025627382" seqno="11477"> From: Susan Schmaltz \<schmaltz@cbvcp.com> Newsgroups: comp.text.sgml Subject: SGML '95 DTD Date: Fri, 17 Nov 1995 12:03:02 -0800 Organization: CompuTech Message-ID: <30ACEA76.630@cbvcp.com> We are looking at the SGML 95 DTD which we got from GCA (we think) and we find it somewhat cryptic. We want to know if anyone can provide a version which defines the elements more fully? Any pointers or help would be greatly appreciated. </message> <message id="<48jgka$fb5@ixnews2.ix.netcom.com>" date="3025661725" seqno="11474"> From: klamerus@ix.netcom.com (Mark Klamerus) Newsgroups: comp.text.sgml Subject: Re: SGML -> HTML Date: Sat, 18 Nov 1995 05:35:25 GMT Organization: Netcom Message-ID: <48jgka$fb5@ixnews2.ix.netcom.com> References: <48g54e$8d3@news.INbe.net> <48i0qh$16e@news.Belgium.EU.net> Jacques Deseyne \<jad@sema.be> wrote: >"M.T. Carrasco Benitez" \<carrasco@innet.lu> wrote: >>I am looking for SGML->HTML filter, or a pointer to the FAQ where this >>kind of answer could be. Mail prefered. >> >>Regards, Tomas >> >A quite good way is to do it through EXPLICIT LINKs. In general, >the cross-conversion needs are quite simple and it takes not too >many LINK sets to achieve a remarkably good result. >The advantage is that you use mechanisms which are defined >within ISO 8879 itself and not one or another proprietary >application language. Alas, only a few tools support LINKs. >Best regards, >------------------------------------------------------ >Jacques Deseyne \<jad@sema.be> >Sema Group Belgium - Stallestraat 96 - B-1180 Brussels >------------------------------------------------------ Actually making the required filters in UNIX scripts is very, very easy. There is no "generic" filter that I know of since every DTD is different, but if you're looking for a "tool" you might consider Exoterica, which is excellent. </message> <message id="<48jgo3$fb5@ixnews2.ix.netcom.com>" date="3025661846" seqno="11475"> From: klamerus@ix.netcom.com (Mark Klamerus) Newsgroups: comp.text.sgml Subject: Re: Searching for SEMA/OSTER Date: Sat, 18 Nov 1995 05:37:26 GMT Organization: Netcom Message-ID: <48jgo3$fb5@ixnews2.ix.netcom.com> References: \<tim-1511952305350001@slug-ppp17.wustl.edu> <48fb2d$sr5@news.Belgium.EU.net> >We are finalizing a new release at this very moment. Instead of plugging >our products here in this newsgroup, I will mail you a full description. >You may be interested to learn that academic and research institutes >can have a free license. Go ahead, plug away. I think the readers of this group are educated enough, and interested enough, to want to know everything that's available in this area, especially since someone asked. </message> <message id="<48jgvc$fb5@ixnews2.ix.netcom.com>" date="3025662079" seqno="11476"> From: klamerus@ix.netcom.com (Mark Klamerus) Newsgroups: comp.text.sgml Subject: Decent Source of "Commented" IETM DTD Date: Sat, 18 Nov 1995 05:41:19 GMT Organization: Netcom Message-ID: <48jgvc$fb5@ixnews2.ix.netcom.com> I am trying to locate a "well-commented" copy of the CALS DTD for IETMs. Does anyone know of a good source for this? The only DTDs I've seen on-line are uncommented (and even hypertexted) and I can't follow my usual source of ANSI and ISO standards (I don't think). thanks, Mark </message> <message id="<48jb3u$k1i@mailgate.bender.com>" date="3025645118" seqno="11478"> From: "Jay.Wood@bender.com" \<Jay.Wood@bender.com> Newsgroups: comp.text.sgml Subject: [Q] system declaration for MS SGML Author for Word? Date: 18 Nov 1995 00:58:38 GMT Organization: Matthew Bender & Co Message-ID: <48jb3u$k1i@mailgate.bender.com> Does anybody have or know where I might get a system declaration for Microsoft's SGML Author for Word? I have tried repeatedly to get this from Microsoft's sales and support lines and web site, with no success. Thanks in advance... ---------------------------------------------------------------------- Jay Wood Matthew Bender & Company 212/216-8151 (voice) 11 Penn Plaza 212/244-3188 (fax) New York, NY 10001 </message> <message id="<19951118T103550Z@naggum.no>" date="3025679750" seqno="11479"> From: Erik Naggum \<erik@naggum.no> Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 18 Nov 1995 10:35:50 +0000 Organization: Naggum Software; +47 2295 0313 Message-ID: <19951118T103550Z@naggum.no> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>, <488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> <199511171131.LAA02593@sun1.cup.cam.ac.uk> [Dave Peterson] | SGML is certainly more than just a grammar, though certainly a grammar | is used in its definition. That grammar definitely is not | context-free. Simple example: how you parse a net ("/") depends on | whether you are currently within an element started by a net-enabling | start-tag. [Paul Treadaway] | Are you sure? Surely if SGML couldn't be represented as a CFG then it | couldn't be deterministically parsed in polynomial time? I don't see how that follows. but let me first address your first question. a context-free grammar is one where the left-hand side of all productions are single meta-syntactic tokens, such that anywhere the left-hand side occurs, the right-hand side may replace it. SGML cannot be expressed this way, for several reasons. e.g., a end-tag can be expressed thus: end-tag = ETAGO, gi, TAGC but with the null end-tag, we don't have this simple rule, we have start-tag contents end-tag = net-enabling-start-tag contents NET it is the multi-token left-hand-side that defines the context in the production. note that no language is without context at some level of interpretation or another. in Lisp, for instance, the semantics of a binding is defined by whether the symbol name is declared special or not. however, this is only a semantic constraint, and has no influence on any syntactic processing. similarly, the type of a variable provides a semantic context for the compiler that may return violations of the semantics. for a syntactic modification, we may consider another context-sensitive grammar, namely that of C++, where a class definition introduces a new keyword into the language. this means that a given statement may be a syntax error if the proper class declaration is not in scope. in SGML, we also have the deprecated (by me) CDATA and RCDATA "declared contents": start-tag contents end-tag = cdata-start-tag character-data end-tag start-tag contents end-tag = rcdata-start-tag replaceable-character-data end-tag now, one might argue that the NET is of the same category as the string syntax in other languages, since it has a special syntax to identify both start and end. however, it also affects elements inside it, and that's not so good. to your point about polynomial time, there are several interesting parsing technologies around. at SGML '95, Ron Turner will present a paper on parsing with Petri nets. the most common parsing technology in the Unix world, partly because lex and yacc are so easy to use, is lexically context sensitive tokenization with regular expressions (lex) and parsing by finite state automatons built from LALR grammars with yacc. if the grammar is LL(k), one may instead use recursive descent parsing, wherein which tokens are looked for by functions that call other functions to handle whatever it finds. LL-parsing is polynomial in the number of tokens and productions, whereas LR-parsing is polynomial in the number of state transitions, which is not necessarily polynomial in the number of tokens and productions. (at least that's how I have understood it, I haven't proven this myself.) | SGML also provides a universal syntax for both DTD's and instances, so | the boundary between metagrammar and grammar is blurred. this is an important observation. this is why I sometimes explain SGML as like a programming language where SGML is the actual language, and the elements have the roles of functions and variables _names_. if one has been exposed to a sufficiently powerful syntax in a programming language (e.g., Lisp) it is easier to explain SGML than if all you have is a criminally impotent syntax (e.g., C). of course, the semantics of SGML is closer to Lisp's nested expressions that it is to C's linear progression of statements, too. since "everybody" believe C (or its evil half-brother C++) is the only language worth learning these days, your following observation is also important: | People can cope with concepts like the internet by analogy, they can | understand at some level the idea of a file format, and even tagged | data, but even if you could explain a grammar in the Chomsky heirarchy | sense, I think SGML is just too hard to convey to non-specialists, | which is why non-technical issues are so frequently what people fall | back on to make the case for SGML. I don't think it would have been this bad if people had been taught that computers can be used to process more than the largely irrelevant bits. I find that people in general are terribly confused on the issue of value vs representation of value. (this also holds for money and time.) also, it wouldn't have been this bad if people had not been subjected to these gross lies about how "easy" computers are to use. they aren't, except in the most trivial of cases. if people had been taught that computers are hard to use, but _some_ tasks are simple and easy, imagine how much further we could have come! of course, such things aren't even possible to discuss in a society that awards people who first are clumsy enough to burn themselves on hot coffee at fast-food joints and then stupid enough to sue, with tons of money. who wants to sell hot coffee to such people? I don't. #\<Erik 3025679750> -- "if you know what you want and you don't go out and do it yourself, you're basically a loser, you know." -- Bj�rk. </message> <message id="<19951118T115840Z@naggum.no>" date="3025684721" seqno="11480"> From: Erik Naggum \<erik@naggum.no> Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 18 Nov 1995 11:58:41 +0000 Organization: Naggum Software; +47 2295 0313 Message-ID: <19951118T115840Z@naggum.no> References: <19951116T071452Z@naggum.no> <48ihmj$bge@newsbf02.news.aol.com> [Barry Schaeffer] | I may be missing something but, given the fact that SGML content is | supposed to be structured according to a DTD, how would one come into | the possession of "SGML Source" for which a DTD must be created? In | the rare instance in which data is available but the DTD is no longer | still around, one might try to reverse engineer one, but the caveat | would be that such inferrence would, at best, create a very tenuous | picture of the intent of the original creators. please do not mail me copies of posted news articles without telling me that they are copies of posted news articles. not only does it give a false sense of private communication, I hate wasting my time replying to what appears to be a personal message but isn't. the following message wasn't intended for public posting, but I don't have time to rewrite it. Date: 17 Nov 1995 16:54:14 UT From: Erik Naggum \<erik@naggum.no> Organization: Naggum Software; +47 2295 0313 Message-Id: <19951117.14766@naggum.no> To: Barry Schaeffer \<isipres@aol.com> In-Reply-To: <9511171546.AA08452@newsbf02.news.aol.com> Subject: Re: Program to take SGML source and produce a DTD | I may be missing something but, given the fact that SGML content is | supposed to be structured according to a DTD, how would one come into | the possession of "SGML Source" for which a DTD must be created? ah, but there is no logical connection between "content is supposed be structured according to a DTD" and a chronological requirement as to which of them is first. suppose you like to create documents from legacy data and the "supposed to be" is rather hard to intuit until after you have converted all of it. or put it this was: SGML favors a certain rigidity in both content and DTD (in fact too much for its own good), and many interpret this to mean that the DTD is the rigid form into which the content is supposed to fit. but that may not be so. suppose the content is rigid and your DTD is fluid. that is, you know what you want, but you don't know precisely how that will turn out in the DTD. as I see it, a DTD is poured over the data and slowly sets into a form that future data is supposed to fit. this may happen in the heads of the people who design this stuff, or it may happen in the data, as a "latent DTD". in the latter case, it is very useful to be able to automate the abstraction process, and fix those things that turn out "wrong" in the DTD. | In the rare instance in which data is available but the DTD is no | longer still around, one might try to reverse engineer one, but the | caveat would be that such inferrence would, at best, create a very | tenuous picture of the intent of the original creators. again, this presumes an a priori DTD, while an a posteriori DTD is at least as valid a concept. in fact, the creators' intent may be broken by a DTD, and the DTD must change. this happens all the time, in my experience, and the question for tools that can automatically realign extant documents to changes in the DTD is very interesting, indeed. I have been thinking about those problems for at least four years -- it is generally intractable, but there must be special cases that are common enough to warrant implementation. #\<Erik 3025616616> #\<Erik 3025684720> -- "if you know what you want and you don't go out and do it yourself, you're basically a loser, you know." -- Bj�rk. </message> <message id="<19951118T151733Z@naggum.no>" date="3025696653" seqno="11481"> From: Erik Naggum \<erik@naggum.no> Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 18 Nov 1995 15:17:33 +0000 Organization: Naggum Software; +47 2295 0313 Message-ID: <19951118T151733Z@naggum.no> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>,<488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> <48gns4$frm@bell.maths.tcd.ie> \<CONNOLLY.95Nov17175945@w3.org> [Dan Connolly] | By those definitions, the grammar given in the SGML spec is | context-free. wrong. the productions alone do not define the grammar of SGML. why is it that so many are fooled by appearance and do not look into the matter? one need only read minor parts of the standard to see that the productions are for purposes of illustration, not rigorous definition. | Finally, SGML is not exactly the language generated by the grammar | consisting of the productions in the standard. [semantic example] the requirements of the standard are the real definition of SGML. the syntactic productions are simplifications, sometimes gross, of the real grammar. thus, any reasoning from the productions that fails to recognize the informal requirements of the standard is irrelevant. the "recognition modes" of the syntax delimiters is proof positive that the syntax is not context-free. the fact that the syntactic productions are modified by recognition modes specified in the informal requirements according to previously established context is proof that the productions are insufficient to reason about SGML. | The same proof shows that the C programming language isn't | context-free, even though there's a handy context-free grammar that | nearly describes it. there is a fundamental difference between syntactic and semantic context. when the syntax of the language can be reduced without regard to the context of the lexical tokens, the syntax is context-free. C's syntax is context-free. SGML's is not. SGML's syntax is driven by the semantics of the language, as in CDATA/RCDATA elements, as in CONREF attributes, as in short reference delimiters. SGML cannot be context-free syntactically. any language that expects to represent anything interesting will, however, have _semantic_ context. all C programs need a "main", all functions must be defined, all variables must be declared in their scope, the type of a variable and the values held by it must match, etc. these semantic constraints are not "contexts" in any useful definition of the term, and it is useless to talk about "context-free" in the semantic sense because the languages that are context-free in this sense do not communicate anything on their own. if you want to look into _complete_ syntax productions for SGML, you will find that the left-hand-side (it appears that you confuse rhs with lhs) needs to specify the context if the productions are to stand on their own. #\<Erik 3025696653> -- "if you know what you want and you don't go out and do it yourself, you're basically a loser, you know." -- Bj�rk. </message> <message id="<mvulpe-1711951920500001@ri4i001.i4i.org>" date="3025642802" seqno="11482"> From: mvulpe@i4i.org (Michel Vulpe) Newsgroups: comp.text.sgml Subject: S4-Browser and Composer (SGML Browser Product Announcement) Date: 18 Nov 1995 00:20:02 GMT Organization: Infrastructures for Information Inc. Message-ID: \<mvulpe-1711951920500001@ri4i001.i4i.org> S4-Browser�: Product Announcement Infrastructures for Information Inc. is proud to announce the S4-Browser�, one of the Standard SGML Support System�, or S4�, suite of SGML tools. The S4-Browser, through its API, provides SGML browsing, navigation, and content services to calling applications. These applications can use BASIC, 4GLs, macros, or C/C++ to access the S4-Browser services. The S4-Browser packages includes: Infrastructures Structure Viewer� that provides a full-function browser GUI for the S4-Browser a set of MS-Word 6.x� Word-Basic macros that support the compostion of valid SGML documents encoded with the HTML DTD. The design philosophy of the S4-Browser allows the user to take advantage of their existing desktop technology and investment in personnel training. Rather than duplicating the functionality of existing document applications such as MS-Word and Adobe Acrobat�, the S4-Browser leverages them and limits itself to what it does best, the provision of SGML browsing, navigation, content and electronic review services. Infrastructures configuration of the S4-Browser for MS-Word 6.x (a version for Adobe Acrobat is also available) provides. Import and validation of an arbitrary valid SGML file Formatting of the SGML in MS-Word 6.x using MS-Word macros. Formatting includes tables and graphics placement Ability to save and view the formatted file as an MS-Word 6.x file without loss of any SGML data or functionality. Viewing and browsing of the SGML tags and their attributes. Search for SGML tags. Search for SGML attribute values. Highlighting in Word of SGML tag contents. Automated navigation of ID and IDREFs. Electronic annotation (e-review) linked to SGML tags. Printing of SGML the structure and related annotation Concurrent views of SGML tags (requires S4-Editor� to built concurrency). Technical Specifications: Windows 3.1 (or higher) with DLL support for your Window's application's development environment. (contact Infrastructures for Mac and Unix implementations) Approximately one megabyte of disk space inclusive of S4-Browser, the Structure Viewer, documentation and the sample MS-Word 6.x macros. S4-Browser has over 30 calls and callbacks divided into service classes as follows: � DTD � Data Insertion � Navigation � Annotation � Data Extraction � Content Model S4-Browser HTML� for the HTML DTDs is limited to non-commercial applications and is available as freeware. For a copy of S4-Browser HTML mail us your request with your return e-mail and postal address to sas@i4i.org or check the Infrastructures WWW site at www.i4i.org under FREE STUFF. Commercial versions of S4-Browser with support for arbitrary DTD's can be licensed from Infrastructures for Information for $149.00 US per seat. For information on the commercial implementation please mail us your request with your return e-mail and postal address to sas@i4i.org or check our WWW site at www.i4i.org or write us at Infrastructures for Information Inc. 330 Dupont, Suite 302 Toronto, Ont., M5R 1V9 416.920.6489 (voice) 416.920.6493 (fax) 990 Hayman Place Los Altos CA 415.969.7363 (voice) 415.969.7003 (fax) sales@i4i.org http://www.i4i.org -- Infrastructures for Information Inc 330 Dupont, #302 Toronto,Ont., M5R 1V9 416.920.6489 </message> <message id="<JBW.95Nov18152744@bigbird.bu.edu>" date="3025715264" seqno="11483"> From: jbw@bigbird.bu.edu (Joe Wells) Newsgroups: comp.text.sgml Subject: parsing SGML with Petri nets? (was: Slight flame on SGML) Date: 18 Nov 95 20:27:44 GMT Organization: Boston University Computer Science Department Message-ID: \<JBW.95Nov18152744@bigbird.bu.edu> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>, <488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> <199511171131.LAA02593@sun1.cup.cam.ac.uk> <19951118T103550Z@naggum.no> In-reply-to: Erik Naggum's message of 18 Nov 1995 10:35:50 +0000 In article <19951118T103550Z@naggum.no> Erik Naggum \<erik@naggum.no> writes: at SGML '95, Ron Turner will present a paper on parsing with Petri nets. Pointer to on-line version of this paper? -- Joe Wells \<jbw@cs.bu.edu> </message> <message id="<48l8jq$im2@pubxfer4.news.psi.net>" date="3025708090" seqno="11484"> From: kimber@passage.com (W. Eliot Kimber) Newsgroups: comp.text.sgml,comp.infosystems.harvest Subject: Re: Hypermedia Addressing [was: Solving the BIG problem] Date: 18 Nov 1995 18:28:10 GMT Organization: Passage Systems, Inc. Message-ID: <48l8jq$im2@pubxfer4.news.psi.net> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> \<CONNOLLY.95Nov12185209@w3.org> In article \<CONNOLLY.95Nov12185209@w3.org>, connolly@w3.orgc says... ... >After noodling on the problem for a while, my favorite name for the >problem is "Resource Discovery and Reliable Links." In other words: >how can we make the most efficient use of the availble >computing/network resources to find an information resource? And once >I've found it, how can I make a link so that the next guy can follow >my footsteps reliably? I've written something on the related issues at: ... >About reliable links: my best guess at a solution is that a reliable >link looks roughly like a traditional citation, library-catalog-card, >business card, etc. It's got one or more URLs -- identifiers that the >computer can resolve in sub-second time most of the time for >relatively recent resources. But it's also got a collection of >constraints, or assertions, or attributes -- a little bit of >knowledge. This extra knowledge is used in conjuction with services >that make large parts of the web "content addressable" -- like >relational databases and fulltext indexes. I generally agree with Dan that for addressing to be reliable it has to use the semantics of the data or objects addressed, not simply their location at the moment. I've often said that explicit linking is impossible at interesting scales (i.e., 10 libraries of 10 documents where each document has on the order of 100 links from itself to the other books the libraries.) However, addressing needs also vary at different levels of granularity. If my goal is to find a book on otters, a catalog card search is good enough. However, if my goal is to take you to the specific sentence about otters in a particular book, then card-catalog accuracy won't do. At this point you have three choices: 1. Use an explicit address, such as a URL with an ID (http://www.otterplace.com/books/otterbook.htm#para2) 2. Combine semantic addressing with explicit addressing (use the card catalog to find the otterbook, then address para2 in that book) 3. Add sufficient semantic description to the paragraph to enable finding it by a query (which, at a certain level of specificity, is identical to a unique identifier). We are agreed that option 1 is generally unsatisfactory. Option 2 is one form of indirect and multi-part addressing defined by HyTime. You can choose exactly how explicit your addresses are, and therefore how much effort goes into creating and maintaining them. Option 3 is essentially the CApH Topic Map approach, which requires adding additional semantic information to the data to be located (again, in the CApH design, using HyTime linking and location addressing mechanisms to do the basic addressing, the association of additional semantic description with data objects, and the representation of the typed links themselves). In other words, there isn't just one answer to the addressing problem. You need a range of options to accomodate a range of requirements and a range of scales and levels of granularity. HyTime provides such a range by defining a general framework of addressing mechanisms that can be combined in an infinite variety of ways to suit the needs of the moment. Because HyTime is built on SGML, it automatically includes the basic semantic descriptive power of SGML and the additional addressing indirection of entities, which allows you to use any storage-system-specific location method to locate storage objects (whose contents will then be located using other methods). When you teach HyTime you have to focus on the explicit addressing methods first, because they're simplere and more famililar, so most people seem to think of HyTime as a way to do cross-document ID addressing, but in fact, it's is the semantic addressing methods that make HyTime really powerful. It is important to remember that HyTime provides ways to integrate any query method into the larger HyTime addressing framework. This means that you should be able to get the best of search and retrieval and the best of explicit addressing in a HyTime system. Note too that HyTime is just the framework--it's use for semantic addressing is dependent on the definition of just the sort semantic addressing schemes represented by URNs, full-text retrieval systems, and other ways that computers can represent or understand the semantic nature of data. -- \<Address HyTime=bibloc> W. Eliot Kimber (kimber@passage.com) Systems Analyst and HyTime Consultant Passage Systems, Inc., 2608 Pinewood Terr., Austin TX 78757 (512)339-1400 10596 N. Tantau Ave, Cupertino CA, 95014, (408) 366-0300 "SGML or die" \</Address> </message> <message id="<DI973L.9CE@mv.mv.com>" date="3025711281" seqno="11485"> Newsgroups: comp.text.sgml From: kaikow@standards.com Subject: Re: SGML book? Originator: kaikow@mv.mv.com Message-ID: \<DI973L.9CE@mv.mv.com> Sender: kaikow@standards.com Reply-To: kaikow@standards.com Organization: MV Communications, Inc. Date: Sat, 18 Nov 1995 19:21:21 GMT References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI625t.F6M@curia.ucc.ie> Followup-To: kaikow@standards.com Is the Alschuler book better than the Practical SGML book by Van Herwijejnen? If so, how? In article \<DI625t.F6M@curia.ucc.ie>, Peter Flynn \<pflynn@www.ucc.ie> wrote: >Dixit pyachnes@ucs.indiana.edu (paul alan yachnes): > >> Can someone recommend the best book for learning SGML? > >The best one I can recommend is "ABCD... SGML: A User's Guide >to Structured Information" by Liora Alschuler (ISBN: 1-850-32197-3) >just published by ITCP. Details at >http://www.thomson.com/itcp/liora/LA_HOME.HTM > >///Peter > </message> <message id="<CONNOLLY.95Nov18152049@w3.org>" date="3025714849" seqno="11486"> From: connolly@w3.org (Dan Connolly) Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: 18 Nov 1995 20:20:49 GMT Organization: Massachvsetts Institvte of Technology Message-ID: \<CONNOLLY.95Nov18152049@w3.org> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>,<488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> <48gns4$frm@bell.maths.tcd.ie> \<CONNOLLY.95Nov17175945@w3.org> <19951118T151733Z@naggum.no> In-reply-to: Erik Naggum's message of 18 Nov 1995 15:17:33 +0000 In article <19951118T151733Z@naggum.no> Erik Naggum \<erik@naggum.no> writes: > > [Dan Connolly] > > | By those definitions, the grammar given in the SGML spec is > | context-free. > > wrong. the productions alone do not define the grammar of SGML. Huh? I thought I was being painfully clear: by "the grammar given in the SGML spec," I meant the productions. Clearly that grammar is context-free. If by "the grammar of SGML," you mean some grammar which generates the SGML language, then I agree that it wouldn't be context free. I said as much in my post. > | Finally, SGML is not exactly the language generated by the grammar > | consisting of the productions in the standard. [semantic example] > > the requirements of the standard are the real definition of SGML. the > syntactic productions are simplifications, sometimes gross, of the real > grammar. thus, any reasoning from the productions that fails to recognize > the informal requirements of the standard is irrelevant. Umm... aren't we saying exactly the same thing? > there is a fundamental difference between syntactic and semantic context. > when the syntax of the language can be reduced without regard to the > context of the lexical tokens, the syntax is context-free. C's syntax is > context-free. SGML's is not. SGML's syntax is driven by the semantics of > the language, as in CDATA/RCDATA elements, as in CONREF attributes, as in > short reference delimiters. SGML cannot be context-free syntactically. Ooops. Yes, I suppose \<x> and \<y> are syntactically the same (i.e. they are represented as the same sequence of terminals in the grammar), and so my proof that SGML isn't context-free is pretty bogus. And I suppose C _is_ context-free at that same level. > if you want to look into _complete_ syntax productions for SGML, you will > find that the left-hand-side (it appears that you confuse rhs with lhs) > needs to specify the context if the productions are to stand on their own. Yup. Dan -- Daniel W. Connolly "We believe in the interconnectedness of all things" Research Scientist, MIT/W3C PGP: EDF8 A8E4 F3BB 0F3C FD1B 7BE0 716C FF21 \<connolly@w3.org> http://www.w3.org/hypertext/WWW/People/Connolly/ </message> <message id="<SatNov18222818EST1995jbw@cs.bu.edu>" date="3025740809" seqno="11487"> From: jbw@bigbird.bu.edu (Joe Wells) Newsgroups: comp.text.sgml Subject: assorted HTML and SGML questions Date: 19 Nov 95 03:33:29 GMT Organization: Boston University Computer Science Department Message-ID: \<SatNov18222818EST1995jbw@cs.bu.edu> To: www-html@w3.org Hi, HTML and SGML gurus, I've got some more questions the answers to which I haven't been able to find in my WWW browsing. Some of these questions are about HTML, some are about SGML, and some are about HTML as an SGML document type. Q: (("text/html" Internet Media Type)) Does text/html forbid including the SGML declaration (\<!SGML ...>)? I know it forbids including a document type declaration subset, but the standard is unclear on whether the SGML declaration is allowed. Q: ((Internet Media Types for SGML)) Since the text/html Internet media type forbids including a DTD subset, what media type should one use if one wishes to transmit an HTML document with a DTD subset via HTTP? Is there something like a text/sgml media type defined anywhere? Q: ((HTML and Empty P Elements)) What are the semantics of an empty P element in HTML? The standard doesn't really seem to deal with this. There are *lots* of documents on the net with *lots* of empty P elements. Is it reasonable for a user agent to issue a warning that this is bad HTML? Q: ((SGML Mixed Content)) I'm not sure if I understand the mixed content rules properly. Let me state what I guess the rules are so that you can tell me if I got it right or wrong. Here is what I think the rules are: * If a content model contains #PCDATA anywhere, the the entire element has "mixed content". * If an element does _not_ have mixed content, then a sequence of characters between two tags that is solely composed of whitespace (SPACE, TAB, RS, RE) is ignored, otherwise the whitespace is treated as ordinary data characters and must correspond to an occurrence of #PCDATA in the content model. Is this right? Q: ((SGML LITLEN)) Is the SGML limit on attribute value lengths (LITLEN) applied to the attribute value after parsing and entity replacement or before? Q: ((HTML PRE Containing FORM)) RFC 1866 says this: For example, a \<PRE> element may contain a \<FORM> element, ... This doesn't make any sense because it contradicts the DTD given in the same document. What's the story here? Q: ((HTML INPUT and SELECT Attributes)) Why is the SIZE attribute of the INPUT element specified to have type CDATA while the SIZE attribute of element SELECT is specified to have type NUMBER? Is this to allow dimension units to be specified? It doesn't say in the standard. Thanks for any help you can give me. -- Joe Wells \<jbw@cs.bu.edu> </message> <message id="<9511191347.AA29508@jytko.jyu.fi>" date="3025777642" seqno="11488"> Newsgroups: comp.text.sgml Date: 19 Nov 1995 15:47:22 +0200 From: Katri Kauppinen \<jytrio@cs.jyu.fi> Message-ID: <9511191347.AA29508@jytko.jyu.fi> Subject: HyTime links Hello, I have some SGML documents which I would like to connect with HyTime links. Is it recommendable to write the links into separate file, and detach them from the plain data documents by ilink? Can the existing SGML documents be unmodified, or should I incorporate any new information in them? How can I get the links to be contained in documents when I want to browse them? Could you please give me an example of the resolution of DTD and link file. Thank you! Katri Kauppinen jytrio-projekti </message> <message id="<19951119T145156Z@naggum.no>" date="3025781516" seqno="11489"> From: Erik Naggum \<erik@naggum.no> Newsgroups: comp.text.sgml Subject: Re: parsing SGML with Petri nets? Date: 19 Nov 1995 14:51:56 +0000 Organization: Naggum Software; +47 2295 0313 Message-ID: <19951119T145156Z@naggum.no> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>, <488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> <199511171131.LAA02593@sun1.cup.cam.ac.uk> <19951118T103550Z@naggum.no> \<JBW.95Nov18152744@bigbird.bu.edu> [Erik Naggum] | at SGML '95, Ron Turner will present a paper on parsing with Petri | nets. [Joe Wells] | Pointer to on-line version of this paper? as far as I know, none exists. however, Ron Turner \<swa@comtch.iea.com> probably won't mind a friendly request for information in his mailbox. #\<Erik 3025781516> -- "if you know what you want and you don't go out and do it yourself, you're basically a loser, you know." -- Bj�rk. </message> <message id="<48nd2s$25j@news2.delphi.com>" date="3025778204" seqno="11490"> From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 19 Nov 1995 13:56:44 GMT Organization: Galahad Message-ID: <48nd2s$25j@news2.delphi.com> References: <48ihmj$bge@newsbf02.news.aol.com> ISI Pres (America Online, Inc. (1-800-827-6364)) wrote: > may be missing something but, given the fact that SGML content is >supposed to be structured according to a DTD, how would one come into the >possession of "SGML Source" for which a DTD must be created? In the rare >instance in which data is available but the DTD is no longer still around, >one might try to reverse engineer one, but the caveat would be that such >inferrence would, at best, create a very tenuous picture of the intent of >the original creators. This brings up an interresting idea. If you were going to convert a large number of legacy documents into SGML one approach would be to start tagging everything. That is, if you have an ascii version of the document, just start adding tags. The idea would be to functionally tag every item of the documents. From this, you could easilly build a list of what type of data the documents really contain. Then you could use that as the starting point for the development of the DTD. This really does not differ much from what *should* be done during document analysis. When you sit down to develop a DTD for a set of legacy documents you have to do the analysis of what is actually in the documents. You could tag the documents at the same time you do the analysis. You need to keep track of what tags are used where, putting the tags into the sample documents at the same time keeps track of that for you. Once you have tagged the documents, you can then do some statistical analysis of where the tags occur. From that you can put together a DTD. The DTD would still need some polishing. But it would get you a quick start. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. </message> <message id="<48o45t$4s@Starbase.NeoSoft.COM>" date="3025801853" seqno="11491"> From: claird@Starbase.NeoSoft.COM (Cameron Laird) Newsgroups: comp.databases,comp.databases.object,comp.software-eng,comp.std.misc,sci.engr.control,comp.object,comp.soft-sys.matlab,comp.text.sgml Subject: Re: CDIF has new internet presence Followup-To: comp.text.sgml,comp.software-eng,comp.object Date: 19 Nov 1995 14:30:53 -0600 Organization: NeoSoft Internet Services +1 713 968 5800 Message-ID: <48o45t$4s@Starbase.NeoSoft.COM> References: <480418$c2f@hopper.isi.com> <816697080snz@tommy.demon.co.uk> In article <816697080snz@tommy.demon.co.uk>, John Nissen \<jn@tommy.demon.co.uk> dreamed: . . . >P.S. Likewise, I'd like to see SGML fit the OLE2 model, for same reason. . . . I don't get it. SGML is a (meta-)language, and simply is not one of the recognized OLE2 classes. Are you saying you'd like someone to build an application that uses OLE2 to share SGML data? What is there to "fit" in that? -- Cameron Laird http://starbase.neosoft.com/~claird/home.html claird@NeoSoft.com +1 713 623 8000 #227 +1 713 996 8546 FAX </message> <message id="<48m8mj$hi1@www1.hlc.net>" date="3025740947" seqno="11557"> From: harvey@eccnet.eccnet.com (Betty Harvey) Newsgroups: comp.text.sgml Subject: Re: Decent Source of "Commented" IETM DTD Date: 19 Nov 1995 03:35:47 GMT Organization: None Message-ID: <48m8mj$hi1@www1.hlc.net> References: <48jgvc$fb5@ixnews2.ix.netcom.com> Mark Klamerus (klamerus@ix.netcom.com) wrote: : I am trying to locate a "well-commented" copy of the CALS DTD for : IETMs. Does anyone know of a good source for this? The only DTDs : I've seen on-line are uncommented (and even hypertexted) and I can't : follow my usual source of ANSI and ISO standards (I don't think). I don't know of a 'commented' IETM DTD. However, there are several sources you may want to checkout. The first is http://navysgml.dt.navy.mil/ietm.html. This contains a minimal amount of information concerning IETMS. The IETM standards, MIL-M-87268, 87269, and 87270 can be downloaded from ftp://navysgml.dt.navy.mil/pub/ietms. The Navy also has an MIL-M-87269 Authoring Guide. You should be able to get a hard copy of this style guide from either Eric Jorgensen, (jorgense@oasys.dt.navy.mil) or John Junod, (junod@oasys.dt.navy.mil). Good luck. Betty </message> <message id="<1995Nov20.045420.29244@sq.com>" date="3025832060" seqno="11492"> Newsgroups: comp.text.sgml From: lee@sq.com (Liam R. E. Quin) Subject: Re: SGML book? Message-ID: <1995Nov20.045420.29244@sq.com> Organization: SoftQuad Inc., Toronto, Canada References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI625t.F6M@curia.ucc.ie> \<DI973L.9CE@mv.mv.com> Date: Mon, 20 Nov 1995 04:54:20 GMT Peter Flynn \<pflynn@www.ucc.ie> wrote: > The best one [book for learning SGML] I can recommend is "ABCD... > SGML: A User's Guide to Structured Information" by Liora Alschuler > (ISBN: 1-850-32197-3) just published by ITCP. Details at > http://www.thomson.com/itcp/liora/LA_HOME.HTM \<kaikow@standards.com> wrote: > Is the Alschuler book better than the Practical SGML book by Van > Herwijejnen? If so, how? I haven't read the ABCD...SGML book in detail yet, but have looked at a couple of chapters carefully, and skimmed the rest. It is better presented than most other books -- the typography in Readme 1st was so bad I couldn't bear to look at it -- and it does have a number of case studies that I think are quite helpful. Erik V.H.'s book is good if you already know that you are going to be using SGML, and have to learn it. It isn't inviting, but it does introduce the concepts, and the 2nd edition (orange cover, not pink) is less obviously IBM-Mainframe-o-centric. The quote about how ``you shouldn't choose long cryptic element names when clear short ones will do'' (paraphrased from memory) is gone, but the approach is still lower level than I'd like. Probably a combination of ABCD...SGML, Practical SGML, and excerpts from The HandBook carefully chosen and heavily wrapped in Lead would make the best combination for teaching. There is also the Brian Travis and ?? book on SGML Implementations, which is too heavy going for casual or beginner use, unfortunately. Both Readme 1st and the Impl. Guide have Strong Opinions on the World Wide Web, and are (I think) clearly hoping that it will go away and leave the SGML world back where it was 2 years ago. It won't, and it's best just to ignore the material in these books about the Web, and to use something like Ian Graham's soon-to-be-updated book for that instead. If you are only going to get 1 book, get Practical SGML. If you are going to get 2 books, get ABCD...SGML and the OUP SGML Handbook. If you are going to get three books, consider adding either the Travis and Walters (?) book, or Practical SGML, depending on whether the heavy text-book format of the former is offputting to you or not. If you are going to get four books, you might as well buy them all :-). Avoid Joan Smith's and Martin Bryant's books for a beginner -- they are rather old now. Joan's book is indispensible for the pictures of the character entities, but you won't need that at first. I'm afraid that I think Martin's book is a little long in the tooth, and spends a long time near the front dealing with dry issues such as character sets that are no longer so important for users to deal with. There is no reason I'm aware of to buy ISO 8879 insterad of the SGML Handbook. Finally, there will soon be new books in the same series as Readme 1st. People who can get past the *appalling* lack of typography (rather like listening to a symphony on a transistor radio with road works jackhammers in the background, but the lead violin is out of tune) have told me that the information in Readme 1st is useful, especially if you have someone who can help point out the errors. I hope later books in the series are worth while. Lee -- Liam Quin, SoftQuad Inc +1 416 239 4801 lee@sq.com \<URL:http://www.sq.com/> HexSweeper NeWS game;OPEN LOOK+XView+mf-fonts FAQs;lq-text unix text retrieval </message> <message id="<1995Nov20.051628.29522@sq.com>" date="3025833388" seqno="11493"> Newsgroups: comp.text.sgml From: lee@sq.com (Liam R. E. Quin) Subject: Re: Program to take SGML source and produce a DTD Message-ID: <1995Nov20.051628.29522@sq.com> Organization: SoftQuad Inc., Toronto, Canada References: <48ihmj$bge@newsbf02.news.aol.com> <48nd2s$25j@news2.delphi.com> Date: Mon, 20 Nov 1995 05:16:28 GMT Jeffrey McArthur \<j_mcarthur@BIX.com> wrote: > This brings up an interresting idea. If you were going to convert a large > number of legacy documents into SGML one approach would be to start tagging > everything. [...] The idea would be to functionally tag every item of the > documents. From this, you could easily build a list of what type of data > the documents really contain. Then you could use that as the starting > point for the development of the DTD. > > [...] Once you have tagged the documents, you can then do some statistical > analysis of where the tags occur. From that you can put together a DTD. > The DTD would still need some polishing. But it would get you a quick > start. Yes, this is how I usually write a new DTD. One other aspect of this is the case where you do not have any editorial authority over the data, and cannot consult the authors. For example, if you're marking up an Eighteenth Century text, you might not know any more thatn that a certain word is in italic. Is it for rhetorical emphasis? Or perhaps because it's the name of a Pope, or a foreign word? There is no better tag than \<I> or \<Italic> or \<Face.I> for this -- to use \<emphasis> or \<pope> or \<ProperName why="Pope"> is to impose an editorial interpretation on the text. SGML DTDs can serve two disparate functions. The first, and most common, is a set of prescriptive rules, saying what is allowed where. The second is to define a set of descriptive markup in which any restrictions on content models are merely for convenience in later processing, since the authoring is already done. I do not mean to imply that there is no overlap between these two uses, nor that the second only applies when the author is unavailable. I have an 18th C. book (John Leland's Itinerary, in Thos. Hearne's 2ns edn., the Sheldonian Theatre, Oxon, 1744) which has an entire chapter inside a footnote, complete with full page illustration. No amount of pointing at any DTD you like will change that. The book is published, and it is not `wrong' -- it was Hearne's decision to put a chapter inside a footnote, so any DTD for the book must of course allow that. In this sort of case, marking up the documents, then creating the DTD, then working out the errors and special cases, is probably the most pragmatic approach you can take. If you *do* have editorial control, though, it's not so good, as it turns out that any errors you make in your tagging will cause the DTD (that is later generated automatically) to allow those errors... In this sort of case, another approach is incremental. I sometimes define a simple DTD, use Author/Editor to enter (or edit) some text, or perhaps even vi.... and then type markup into Author/Editor. I then change the DTD to allow the new markup, save and re-import the file (`open' uses the old rules file from the in-memory cache) and proceed in a loop until tired. Be neat to be able to add elements to the DTD in A/E, but I know many customers would rather *not* have that feature, as they use SGML to get control over their technical authors, financial analysts, or whatever. Still, two approaches to DTD-writing, and two kinds of DTD, points at the ends of spectra, perhaps defining a rectangular co-ordinate space... Lee -- Liam Quin, SoftQuad Inc +1 416 239 4801 lee@sq.com \<URL:http://www.sq.com/> HexSweeper NeWS game;OPEN LOOK+XView+mf-fonts FAQs;lq-text unix text retrieval </message> <message id="<48pbff$pqd@news.kreonet.re.kr>" date="3025842489" seqno="11494"> From: sksong@rcunix.kotel.co.kr (Seongkyu Song) Newsgroups: comp.text.sgml Subject: [Q] Is there any SGML viewer for WWW? Date: Mon, 20 Nov 1995 07:48:09 GMT Organization: Korea Research Environment Open Network (KREONet) Message-ID: <48pbff$pqd@news.kreonet.re.kr> Thank you. </message> <message id="<48psgu$ep5@naurouze.cert.fr>" date="3025859550" seqno="11495"> From: Martin Hollender \<Martin.Hollender@onecert.fr> Newsgroups: comp.text.sgml Subject: Re: [Q] Is there any SGML viewer for WWW? Date: 20 Nov 1995 12:32:30 GMT Organization: EURISCO Message-ID: <48psgu$ep5@naurouze.cert.fr> References: <48pbff$pqd@news.kreonet.re.kr> To: sksong@rcunix.kotel.co.kr SoftQuad Panorama http://www.sq.com </message> <message id="<48q18e$ep5@naurouze.cert.fr>" date="3025864398" seqno="11496"> From: Martin Hollender \<Martin.Hollender@onecert.fr> Newsgroups: comp.text.sgml,comp.infosystems.harvest Subject: Re: Hypermedia Addressing [was: Solving the BIG problem] Date: 20 Nov 1995 13:53:18 GMT Organization: EURISCO Message-ID: <48q18e$ep5@naurouze.cert.fr> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> <9511100618.AA15754@fly.HiWAAY.net> \<tbray.10.30A3B6E9@opentext.com> \<CONNOLLY.95Nov12185209@w3.org> Maybe we should distinguish technical addressing (the computer science side) from cognitive addressing (the human factors side) In the following I want to make some remarks on the cognitive aspects a) Link addresses should be readable and interpretable by humans (Doug Englebart) A good address should indicate what information we can expect. Thus "http://www.w3.org/hypertext/WWW/People/Connoly" is much better than e.g. "ID45678x56" Such addresses are easier to be remembered and can be printed (many newspapers already cite electronic documents) a1)Because Hard-copys are very important, we should not forget them. It should be relatively easy to generate Hard-Copies with all the links and addresses being operational for the human reader b) I'm not shure if the retrieval of an Information Object has to be done directly by the addressing mechanism itself: If the documents have a good structure and good search engines are at hand, they can be used to identify the Information Object (no addresses visible for the human yet) Then it is veryfied if this Information Object is really what was looked for. Only in the last step the address of the Information Object is looked up. (it could be even totally invisible using the copy/paste mechanism) That means Library Catalogues, Table of contents, Indices etc. are all very important for the retrieval of Information Objects but do they need to be part of the address? c) If we have to make references to Information Objects we can or want not to retrieve before referencing them, we have to remember the address of the Information Objects by heart. This will be easier if the address contains as much semantic knowledge as possible. -- ----------------------------------------------------------------- Martin Hollender Research Associate eurisco - European Institute of Cognitive Sciences and Engineering 4, av Edouard Belin 31400 TOULOUSE, France Tel: (+33) 62.17.38.31 Fax: (+33) 62.17.38.39 ----------------------------------------------------------------- </message> <message id="<48q57o$8jj$1@mhafm.production.compuserve.com>" date="3025868472" seqno="11497"> From: Raymond H. Stachowiak <74453.1747@CompuServe.COM> Newsgroups: comp.text.sgml Subject: Re: Transporting SGML fragments Date: 20 Nov 1995 15:01:12 GMT Organization: System Consulting Message-ID: <48q57o$8jj$1@mhafm.production.compuserve.com> References: <48inf9$cu1@newsbf02.news.aol.com> In response to Barry Schaeffer's remarks about SGML fragments, it comes to mind that there may in fact be two references to SGML framents. The Document fragment, which Barry describes to detail, and the DTD fragment, which in fact may be a logical construct that is used much as a module or subroutine in larger DTDs. This is often referred to as a DTD frament although there might be a better term. Essentially we are referring to somethin like a Table fragment that might be included as an entity reference in dtd, or other type of construct that might be used over and over. It could parse in its won right and also then parse within the document. As for Document fragemnts, often these are simply subsections of a complete document, such as a task from a maintenance manual. Most SGML editor provide some level of parsing and support for these fragments. Ray Stachowiak </message> <message id="<9511201830.AA12588@fly.HiWAAY.net>" date="3025881004" seqno="11498"> Newsgroups: comp.text.sgml Date: Mon, 20 Nov 1995 12:30:04 -0600 Message-ID: <9511201830.AA12588@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> Subject: RE: VRML Behaviors \<Crosspost from="vrml-behaviors@sdsc.edu"> \<intro> For comp-text-sgml readers, the original subject is the use of plug-ins to support multi-protocol systems. The advantage is to isolate the browser code from the protocol code. The disadvantage is work and expense. \</intro> \<bod> [John Morrison] >Hello Len, > >Thank you for your reply. You are welcome. Here is my opinion in non technical language as you have expressed the technical problem quite well. I agree with you that multi-protocol support is hard and expensive. Tower of Babel analogies are easy to find in systems implementations, myth, and the study of natural and artificial ecologies. However, the natural competition in the environment leads to separate communities. These communities may overlap and these are called ecotonal boundaries. At these boundaries, competition is at its highest, and so is the noise and resulting system instability and dynamism. Nature never worked out a way to prevent this. In fact, it is the engine of evolution, and it evolves superb cooperators, like ants. The Web is not a new technology. It is a marriage of existing technologies whose offspring is an industry. We are ants in that industry (No offense. They work like hell and give up the best crumbs to their queen). My caveat is that no amount of force seems capable of keeping all of the ants marching in a straight line across uneven terrain even though they are models of cooperation. The best that can be done is provide signals and trained behaviors by which they reorganize at suitable intervals and events. 1. Each protocol has been optimized for some dimension of the communications problem (DIS for heavy lifting of dynamic object movement in a simulation, HTTP for light lifting of documents, Z39.50 for heavy lifting (i am told) of documents). 2. File formats are optimized for the representation of content and rendering of the data type (SGML for complex compound documents, HTML for non-complex compound documents, VRML for explicit but light representation of 3D graphics, CGM for raster/vector combinations, IGES for design data, PDES for everything under the sun, etc). 3. We can't amass enough guns to stop it. CALS tried this and has fallen on its face and that is with the power of the Federal government behind it. Apparently, such organization is simply unnatural and supernatural means aren't available. My approach because I am an SGML person and not a graphics person is to use the SGML document type and notation facilities to organize the interfaces and resources that I want to use. That way, at least within the compound document I can exactly and testably specify what is needed to support a compound document class. If I know, for example, for an enterprise which classes of document types I must support, what their order of exchange is (e.g., you send Request for Quote, I send Quote or Request for Information) then at least at the level that the average user sees things, this is understandable tractable, and appears to be within the bounds of Internet energy budgets (packets, bandwidth, etc). The DTD writer in conjunction with the system support guys must figure out what the senders and receivers can support, and therefore, which plug-ins have to be on both sides of the fence to interoperate and communicate. Can that lead to protocol and format explosion? It can but it may not for very long for exactly the reasons you cite, in essence, too much work so local system and employees fall down dead trying. "Her Majesty's a pretty nice girl, but she changes from day to day." The link and location models are the vital component that should be standard. The script processing will not be interoperable and that is very bad news but solvable only by evolution, and then, not completely. Competitors will exist. VRML and SGML systems will thrive by cooperation in this. Define the boundaries and the overlap. Optimize these based on common needs among the communities and try not to shut down the government in the process. ;-) From the Tales of Nasrudin (an old caravan trader): "Be on good terms with thy ass for it bears thee." I think that as the ass brays loudly, the riders will begin to see the sense of keeping it happy. So, our best bet here is to provide a saddle which both the rider and the ridden are comfortable with. len bullard\</bod> \</crosspost> Len Bullard </message> <message id="<48q6u9$5d3@dove.nist.gov>" date="3025870217" seqno="11499"> From: bagwill@kangaroo.ncsl.nist.gov (Robert Bagwill) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 20 Nov 1995 15:30:17 GMT Organization: NIST Message-ID: <48q6u9$5d3@dove.nist.gov> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> In article \<DI79G4.B42@udcf.gla.ac.uk>, flavell@v2.ph.gla.ac.uk (Alan J. Flavell) writes: >>In <487tql$mus@dove.nist.gov> bagwill@kangaroo.ncsl.nist.gov writes: >>> Nostradamus predicts: In a few more months, almost no HTML will be >>> hand-generated. > >I put it to you that you are peering into the wrong end of the >telescope. It does not in the least matter what the document looks >like to its author. Baloney [replace with British equivalent]. All publishers want control. All authors raised on WYSIWYG wordprocessors and desktop publishing want control. Resistance is futile. HTML will be assimilated [stupid but appropriate STTNG Borg reference]. >Unless, of course, you are writing for a captive audience The best audience is a captive audience! Preferably bound and gagged, so you can get at their wallets more easily. :-) >Well, either that or rediscover the original intentions of HTML, i.e as >marking up the structure of information. My impression is that Tim B.L.'s original intention was to add quick-and-dirty hypertext links to CERN documents. The NCSA developers who now work for Netscape added quick-and-dirty WYSIWYG markup. SGMLophiles are trying to rationalize HTML ex post facto. I am 210% in favor of semantic markup. But if the postings to comp.infosystems.www.authoring.html are any indication of what authors want, fonts and frames and pretty pictures are what's important. The customer's always right. As they say in China: Woks populi, woks day. -- Bob Bagwill \<rbagwill@nist.gov> </message> <message id="<psystemsDID2n7.7p3@netcom.com>" date="3025892131" seqno="11501"> Newsgroups: comp.text.sgml From: psystems@netcom.com (Passage Systems) Subject: SGML Applications Engineer, Cupertino, CA Message-ID: \<psystemsDID2n7.7p3@netcom.com> Organization: NETCOM On-line Communication Services (408 261-4700 guest) Date: Mon, 20 Nov 1995 21:35:31 GMT Sender: psystems@netcom3.netcom.com ================================================== SGML Applications Engineer, Cupertino, CA ================================================== Passage Systems, specialists in software development for technical publications customers, is seeking highly motivated software engineers to join one of the industry's fastest growing companies. We are in search of people with a computer science back- ground and strong analytical skills. Experience in the following areas required: - B.S. in Computer Science, MS a plus - C or C++ programming - Internet publishing - Standard Generalized Markup Language (SGML) or Hypertext Markup Language (HTML) Experience with the following is highly desired: - OmniMark programming language - Perl - SGML tools (parsers, editors, viewers) - Word Processors (MS Word, Framemaker, WP) - Free-text retreival Job Description: Join our team of the industry's top engineers to develop hypertext and multimedia publishing applications for CDROM and Internet delivery. About Passage Systems: Passage, headquartered in Cupertino, is one of the fastest growing SGML software and services companies. Join our energetic start-up company, Passage offers an exciting work environment tailored to goal oriented, highly productive technical experts. Our company offers competitive salary and benefits, plus a sabbatical and equity in the company commensurate with experience. If this opportunity sounds interesting to you, then please email or fax your resume to the address listed below. Principals only. Bryan K. Caporlette Manager, Applications Engineering Passage Systems, Inc. 10596 N. Tantau Avenue Cupertino, CA 95014 Fax: 408-366-0320 Email: caps@passage.com </message> <message id="<N.112095.232631.48@dialup8.isicom.fr>" date="3025895191" seqno="11502"> From: aperier@planetepc.fr Newsgroups: comp.text.sgml Subject: QWERTZ DTD and SofQuad/Rules Builder Date: Mon, 20 Nov 95 22:26:31 GMT Organization: Internet Way Message-ID: \<N.112095.232631.48@dialup8.isicom.fr> I can't compile the qwertz DTD using rules builder 3.0 release 3.28 for Windows 3.11. The error is: ERROR in the document type definition: general error SHORTREF, and USEMAP declarations are not currently supported </message> <message id="<1995Nov20.214942.10037@sq.com>" date="3025892982" seqno="11503"> Newsgroups: comp.text.sgml From: lee@sq.com (Liam R. E. Quin) Subject: Re: SGML book? Message-ID: <1995Nov20.214942.10037@sq.com> Organization: SoftQuad Inc., Toronto, Canada References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI625t.F6M@curia.ucc.ie> \<DI973L.9CE@mv.mv.com> <1995Nov20.045420.29244@sq.com> Date: Mon, 20 Nov 1995 21:49:42 GMT A couple of people have asked, so I just wanted to point out that the comments I made about various SGML books are my initial opinions based on a quick look (as I said in the article, but people asked anyway!). Please, if you're going to buy books, look at them, and read reviews! We're very fortunate that there are quite a few books on SGML, and more seem to be on the way, so you can choose ones that will suit you. Lee -- Liam Quin, SoftQuad Inc +1 416 239 4801 lee@sq.com \<URL:http://www.sq.com/> HexSweeper NeWS game;OPEN LOOK+XView+mf-fonts FAQs;lq-text unix text retrieval </message> <message id="<DICyvy.2yn@news.zippo.com>" date="3025886375" seqno="11504"> Newsgroups: comp.text.sgml From: zsoos@elan-gmk.com (Zoltan Soos) Subject: Expert help needed Reply-To: zsoos@elan-gmk.com Sender: usenet@news.zippo.com Organization: ELAN GMK Message-ID: \<DICyvy.2yn@news.zippo.com> Date: Mon, 20 Nov 1995 19:59:35 GMT Hello, As a service bureau, we are involved in several SGML conversion projects. Although we do have in-house SGML related talent, we would like to get a second opinion on specific problems. Other than assisting us with an immediate problem, this relationship can expand into a fairly regular consultation arrangement. Tasks might include DTD validation, interpretation and design, database integration, possible educational assignment. Please reply via mail to zsoos@elan-gmk.com Zoltan </message> <message id="<48qmfi$14r@bell.maths.tcd.ie>" date="3025886130" seqno="11505"> From: tim@maths.tcd.ie (Timothy Murphy) Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 20 Nov 1995 19:55:30 -0000 Organization: Dept. of Maths, Trinity College, Dublin, Ireland. Message-ID: <48qmfi$14r@bell.maths.tcd.ie> References: <9511171331.AA12385@fly.HiWAAY.net> Len Bullard \<cbullard@HiWAAY.net> writes: >The numbers of people in the US signed up >to the Internet still remains a low percentage although growing daily. Is this true? I read the following statistics last week. ============================================= The Electronic Bulletin of the GLObal Systems Analysis and Simulation Association in the USA Email distribution: GLOSAS@vm1.mcgill.ca According to a recent Internet Society survey, there are about 3.4 million "host" computers hooked up to the Internet in the United States and just over 500,000 in Western Europe. WHO'S ONLINE? * 37 million, or 17% of the U.S. and Canadian population 16 and older, have access to the Internet, * 24 million used the Internet in the past three months, * They spend an average of 5 hr. 28 min. on the Internet per week, * 34% are women, * 66% reached the Internet from work, * 25% of World Wide Web users have incomes of more than $80,000. (TIME, Nov. 13, 1995, Page 121) ============================================= -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland </message> <message id="<199511201748.RAA01462@sun1.cup.cam.ac.uk>" date="3025878035" seqno="11506"> From: Paul Treadaway \<pt100@cup.cam.ac.uk> Newsgroups: comp.text.sgml Subject: Re: Slight flame on SGML Date: Mon, 20 Nov 95 17:40:35 GMT Organization: Cambridge University Press Message-ID: <199511201748.RAA01462@sun1.cup.cam.ac.uk> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <47t53d$rqb@News1.mcs.net> \<DHssL8.8sD@world.std.com>, <488r3b$ruf@bell.maths.tcd.ie> <48advl$bvi@hopper.acm.org> <199511171131.LAA02593@sun1.cup.cam.ac.uk> <19951118T103550Z@naggum.no> In-Reply-To: <19951118T103550Z@naggum.no> In article <19951118T103550Z@naggum.no>, Erik Naggum says... >[Dave Peterson] > >| SGML is certainly more than just a grammar, though certainly a grammar >| is used in its definition. That grammar definitely is not >| context-free. Simple example: how you parse a net ("/") depends on >| whether you are currently within an element started by a net-enabling >| start-tag. > >[Paul Treadaway] > >| Are you sure? Surely if SGML couldn't be represented as a CFG then it >| couldn't be deterministically parsed in polynomial time? > >I don't see how that follows. Well, I was under the impression that the Context Free/Context Sensitive boundary was also the P/NP boundary for deterministic parsing. > but let me first address your first >question. a context-free grammar is one where the left-hand side of all >productions are single meta-syntactic tokens, such that anywhere the >left-hand side occurs, the right-hand side may replace it. SGML cannot be >expressed this way, for several reasons. > >e.g., a end-tag can be expressed thus: > > end-tag = ETAGO, gi, TAGC > >but with the null end-tag, we don't have this simple rule, we have > > start-tag contents end-tag = net-enabling-start-tag contents NET > >it is the multi-token left-hand-side that defines the context in the >production. The question is whether that could be represented as a CFG. In this case, for example, you have two distinct alternatives, don't you, so you should be able to rewrite that as: e -> start-tag contents end-tag e -> net-enabling-start-tag contents NET net-enabling-start-tag -> STAGO gi attlist s* NET start-tag -> STAGO dts gi attlist s* TAGC etc. where e is some non-terminal. Which is context free. Now, I don't know that this is true for all SGML, but unless I'm missing something significant in the above, it is the case here. Basically, my question is whether SGML is in general expressable as a CFG. I remember reading somewhere that it is, but that converting to some standard representation of a CFG might be computationally hard. >note that no language is without context at some level of interpretation or >another. in Lisp, for instance, the semantics of a binding is defined by >whether the symbol name is declared special or not. however, this is only >a semantic constraint, and has no influence on any syntactic processing. >similarly, the type of a variable provides a semantic context for the >compiler that may return violations of the semantics. for a syntactic >modification, we may consider another context-sensitive grammar, namely >that of C++, where a class definition introduces a new keyword into the >language. this means that a given statement may be a syntax error if the >proper class declaration is not in scope. Note that the word 'context' in Context Free and Context Sensitive grammars is not actually very meaningful. The terms Type 0, 1, 2, 3 for Regular grammars, Context Free grammars, Context Sensitive grammars and Unrestricted grammars are probably preferable. Actually, I think your C++ example is also one of semantics, when viewed from a type-theoretical point of view. It's simply presented as something different to that within the designer's model of the language. >in SGML, we also have the deprecated (by me) CDATA and RCDATA "declared >contents": > > start-tag contents end-tag = cdata-start-tag character-data end-tag > start-tag contents end-tag = > rcdata-start-tag replaceable-character-data end-tag > >now, one might argue that the NET is of the same category as the string >syntax in other languages, since it has a special syntax to identify both >start and end. however, it also affects elements inside it, and that's not >so good. Unless I'm missing something, these are also semantic considerations. Surely these could also be represented with Context Free productions? >to your point about polynomial time, there are several interesting parsing >technologies around. at SGML '95, Ron Turner will present a paper on >parsing with Petri nets. the most common parsing technology in the Unix >world, partly because lex and yacc are so easy to use, is lexically context >sensitive tokenization with regular expressions (lex) and parsing by finite >state automatons built from LALR grammars with yacc. if the grammar is >LL(k), one may instead use recursive descent parsing, wherein which tokens >are looked for by functions that call other functions to handle whatever it >finds. LL-parsing is polynomial in the number of tokens and productions, >whereas LR-parsing is polynomial in the number of state transitions, which >is not necessarily polynomial in the number of tokens and productions. (at >least that's how I have understood it, I haven't proven this myself.) I'd be interested to read the paper on parsing with Petri nets, but is it deterministic or non-deterministic? Certainly LALR, LL(k) and LR(k) are subsets of context free grammars, aren't they? Finite state automatons are isomorphic with regular grammars (Chomsky type 0, a proper subset of context free). Certainly LL(k) and LR(k) can be parsed deterministically (using recursive descent and shift reduce parsing respectively) in poynomial time. </message> <message id="<SMISHRA.95Nov20154436@gaia.cc.gatech.edu>" date="3025889076" seqno="11507"> From: smishra@gaia.cc.gatech.edu (Sunil Mishra) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 20 Nov 1995 20:44:36 GMT Organization: Bizarre Inc. Message-ID: \<SMISHRA.95Nov20154436@gaia.cc.gatech.edu> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> <48q6u9$5d3@dove.nist.gov> Reply-To: smishra@cc.gatech.edu In-reply-to: bagwill@kangaroo.ncsl.nist.gov's message of 20 Nov 1995 15:30:17 GMT In article <48q6u9$5d3@dove.nist.gov> bagwill@kangaroo.ncsl.nist.gov (Robert Bagwill) writes: \\\\ favor of semantic markup. But if the postings to \\\\ comp.infosystems.www.authoring.html are any indication of what authors \\\\ want, fonts and frames and pretty pictures are what's important. The \\\\ customer's always right. As they say in China: Woks populi, woks day. \\\\ \\\\ -- \\\\ Bob Bagwill \<rbagwill@nist.gov> \\\\ That's why they have proposed stylesheets, \<ul src=...> and a host of other *better* ways of giving presentation. Unfortunately, the customer is the one being ignored by netscape and others. Sunil </message> <message id="<30B106FE.5E85@novell.com>" date="3025896830" seqno="11508"> From: Ben Brimhall \<bbrimhall@novell.com> Newsgroups: comp.text.sgml Subject: Re: Where can I find the ISO char sets Date: Mon, 20 Nov 1995 16:53:50 -0600 Organization: Novell, Inc. Message-ID: <30B106FE.5E85@novell.com> References: \<DHF9Lv.M4z@oasis.icl.co.uk> I work with SGML on a regular basis, and was roaming around on the internet searching for entity lists. I found a page that had what I wanted it is at "http://www.alis.com:8085/ietf/html/index.html" Good Luck Ben </message> <message id="<48q8ha$7gc@barnacle.iol.ie>" date="3025904114" seqno="11500"> From: digitome@iol.ie (Sean Mc Grath) Newsgroups: comp.text.sgml Subject: Re: SGML + Lotus Notes Date: Tue, 21 Nov 1995 00:55:14 GMT Organization: Digitome Ltd. Message-ID: <48q8ha$7gc@barnacle.iol.ie> References: <489p7e$gdi@gidora.kralizec.net.au> jon@zeta.org.au wrote: >Has anyone seen any good examples of a hybrid Lotus Notes/SGML system or >any toolkits that can be used to help build such systems? >jon. I do not know of any hybrid systems. However, we have an SGML to Lotus Notes formatting engine as part of our IDM SGML Transformation Toolkit. It has a Lotus Notes aware stylesheet language that allows the production of richly formatted, richly doclinked Notes databases direct from SGML source documents. Hope this helps, Sean Mc Grath digitome@iol.ie > </message> <message id="<48sokk$cnm@news.Belgium.EU.net>" date="3025953876" seqno="11509"> From: Jacques Deseyne \<jad@sema.be> Newsgroups: comp.text.sgml Subject: Re: [Q] system declaration for MS SGML Author for Word? Date: 21 Nov 1995 14:44:36 GMT Organization: SEMA Group Belgium Message-ID: <48sokk$cnm@news.Belgium.EU.net> References: <48jb3u$k1i@mailgate.bender.com> "Jay.Wood@bender.com" \<Jay.Wood@bender.com> wrote: > >Does anybody have or know where I might get a system declaration for >Microsoft's SGML Author for Word? I have tried repeatedly to get this >from Microsoft's sales and support lines and web site, with no success. > >Thanks in advance... > You are right in requesting this "fact sheet" from vendors, although most are probably not interested to supply the information. I am afraid you will have to build a system declaration yourself, and still then not all information is available, except maybe by trial and error. Concerning the sections of a system declaration for SGML Author : * CAPACITY numbers are not specified in the documentation * FEATURES supported are described somewhat on p. 121 of the System Administrator's Guide. Remarkably enough, the text says "All types of markup minimization are supported from SGML to Word.", and two lines further : "Short references, start or end tag minimization, CDATA character entity references and usemaps are not supported." SUBDOC, DATATAG, CONCUR, RANK and LINK are explicitly mentioned as not being supported. * SCOPE is in a certain way both DOCUMENT and INSTANCE, as for SYNTAX only the reference concrete syntax is supported. * VALIDATION services are not specified. I have the feeling that most entries should be YES, but I'm not sure if the Author application loading a DTD for defining a mapping file ever takes the information from an SGML declaration into account. Best regards, ------------------------------------------------------ Jacques Deseyne \<jad@sema.be> Sema Group Belgium - Stallestraat 96 - B-1180 Brussels ------------------------------------------------------ </message> <message id="<9511211904.AA13420@fly.HiWAAY.net>" date="3025969445" seqno="11510"> Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Date: Tue, 21 Nov 1995 13:04:05 -0600 Message-ID: <9511211904.AA13420@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> References: <9511171331.AA12385@fly.HiWAAY.net> <48qmfi$14r@bell.maths.tcd.ie> Subject: Re: Slight flame on SGML [Len Bullard] >The numbers of people in the US signed up to the Internet still remains a >low percentage although growing daily. [Timothy Murphy] >Is this true? I read the following statistics last week. Yes, I made a sweeping statement, so I can't say if it true or untrue. Neither can you. But I have a problem with the statistics I'm seeing on Internet usage and their use to justify just about any claim about the given technologies. I get different numbers and all of them indicate that the rage to get on the net is in full swing, but exaggerated. It will be interesting to see if there is also an increase in dropped accounts if a global recession sets in and folks have to choose between the Net and CableTV. To me, the most significant indicator of the rage is the self-evident fact that the time to get a dial tone from the service provider is rapidly increasing, and the network brownouts and router congestion is growing. Part of this is bigger numbers, part of it is longer times spent online and bigger transactions (hey, HTML and Jpegs! Cool!) 1. It's too easy to bake the numbers and without the method for gathering them, hard to determine if they are significant. 2. How do you or any of the rest of us know who is hooked up, how, and how many use that connection? To whom is it significant if you do know? For example, "25% of WWW users have incomes of more than $80,000 dollars" just doesn't seem right. Could be but could be joint incomes. If true, that means the AOL marketing adds sure are targeting the wrong market. 3. Which Web technologies are they really using, or is this just a composite of all users of all Internet technologies? FTP can be had in an HTML browser but it's not a WebThang. The facts are hard to come by and they too often reflect the wishes of people who have discovered they can control perceptions with such and that perceptions sell systems and vendors. Nothing new in that. None of these are convincing arguments that extant Web technologies are the right technologies. I find it a bit distressing for example that we see "SGMLophiles are ex post facto justifying the design of HTML" and in the same day "SGML authors are hoping the Web will quietly go away". Just ain't true. Throw in some stats like the ones cited above, and it seems to me some are trying to make the case that we should just accept the current players and technologies as the winners and go home. THAT is what is not going to happen. It's early in the first quarter here. The other guy has made a touchdown based on a fumble, but little else. HTML is an SGML technology. Fabulous! We can use the Internet plumbing to springboard more powerful SGML apps and are doing so. So what. HTML won't be damaged and users with different requirements will get what they want. You don't like SGML? So what. You can invent your own language and have at in the free market with the rest of us. This is free enterprise, not socialist driveling from college students and ambitious corporate moguls who discovered the tactics of Goebbels and use it to drive the young and gullible. C'mon Tim, upfront dude: what do you want from us here, or from the Net in general? I've read a lot of negative stuff and cries for more "control" from you but who is going to bell the cat? You? I don't think so unless you are armed with a pretty vast fortune. Bill Gates? Nah, even with all of his money and resources, he has to cooperate with vendors, legacy code, a huge customer base, and all of us nipping at this heels. Charles Goldfarb? Nah, he just wants to see SGML used well and he turned his back on the *BIG MONEY* years ago. ME? Not without a golden parachute? We can tediously and pedantically debate the design of SGML on CTS, and have for quite a long time. Most of us already know these things and they are most certainly subjects for the SGML review. Contribute to that and do some good for the world. But the HTML vs SGML brouhaha is over. It's SGML, but the system evolution just doesn't stop with SGML or HTML. SGML is fine hub for a compound document, but that NOTATION declaration is there for a reason. Let's have a show of hands: o Anyone out there who doesn't accept that HTML is SGML and that the Internet is a viable communications medium raise your hand. o Anyone out there who thinks the WWW will all go away, raise your hand. o Anyone out there who thinks one solution and one design can dominate the march of Ants across the Alps, raise your hand. Now, all of you who raised your hands, please submit resumes to Microsoft or Apple. You're their kinda folks. ;-) So, again, Tim, up front and publically posted: what do you want? cheers, Len Bullard </message> <message id="<PRIESTDO.95Nov21110233@cssun1.vassar.edu>" date="3025958553" seqno="11511"> From: priestdo@cs.vassar.edu (Greg Priest-Dorman) Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 21 Nov 1995 16:02:33 GMT Organization: Vassar College Message-ID: \<PRIESTDO.95Nov21110233@cssun1.vassar.edu> References: <48ihmj$bge@newsbf02.news.aol.com> <48nd2s$25j@news2.delphi.com> In-reply-to: j_mcarthur@BIX.com's message of 19 Nov 1995 13:56:44 GMT I believe "Fred: The SGML Grammar Builder" will do the tast you ask. I have never used it but it's home page is: http://www.oclc.org/fred/ The "About Fred" link states "Fred can automatically build DTDs from tagged text." So I think this is what you are looking for. I hope this helps, Greg Priest-Dorman </message> <message id="<816971953.18186@hchworth.demon.co.uk>" date="3025960753" seqno="11512"> From: Paul Chan \<paul> Newsgroups: comp.text.frame,comp.text.sgml,comp.text.desktop,comp.text.pdf,comp.text.tex,alt.aldus.pagemaker,uk.jobs.offered Subject: DTP Experienced Person Required for UK-based operation. Date: Tue, 21 Nov 1995 16:39:13 GMT Message-ID: <816971953.18186@hchworth.demon.co.uk> Apologies for the crosspost. Hopefully, this message is relevant to at least some of you out there. We require a DTP skilled person to set up and manage a Framemaker operation with our company at a UK based site. Please apply to mailto:Geoff.Aitken@Charlesworth.com Thanks for your time. </message> <message id="<48t627$fr6@voyager.Internex.NET>" date="3025967619" seqno="11513"> From: davido@apocalypse.org (David Ornstein) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: Tue, 21 Nov 1995 18:33:39 GMT Organization: InterNex Information Services 1-800-595-3333 Message-ID: <48t627$fr6@voyager.Internex.NET> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> <48q6u9$5d3@dove.nist.gov> \<SMISHRA.95Nov20154436@gaia.cc.gatech.edu> smishra@gaia.cc.gatech.edu (Sunil Mishra) wrote: >In article <48q6u9$5d3@dove.nist.gov> bagwill@kangaroo.ncsl.nist.gov (Robert Bagwill) writes: >\\\\ favor of semantic markup. But if the postings to >\\\\ comp.infosystems.www.authoring.html are any indication of what authors >\\\\ want, fonts and frames and pretty pictures are what's important. The >\\\\ customer's always right. As they say in China: Woks populi, woks day. >\\\\ >\\\\ -- >\\\\ Bob Bagwill \<rbagwill@nist.gov> >\\\\ >That's why they have proposed stylesheets, \<ul src=...> and a host of other >*better* ways of giving presentation. Unfortunately, the customer is the >one being ignored by netscape and others. At the risk of stepping out having not read all the important documents in detail, let me ask about style sheets and markup. It has seemed to me that the CLASSes being added to the various tags (e.g., \<EM class=Person>) are really a poor-man's way to add new tags (e.g., \<Person>). Is this basically right? Ideally (of course?) browsers would support displaying custom tags defined in DTDs that are specific to the documents being viewed. I know they don't do this (I assume because it's a lot of work and they have too many other things on thier minds), but why didn't the style sheet folks design something that was a bit more in line (like allowing the style sheet information to define the new tags for a browser using stuff in the header -- thus avoiding the full complexity of DTDs)? [Sorry for the rough phrasing...] --------------------------------------------- David Ornstein BrowserCaps: http://objarts.com/bc Outbreak: http://ichiban.objarts.com/outbreak-unreg Personal Info: http://objarts.com/davido </message> <message id="<48t8d7$fr6@voyager.Internex.NET>" date="3025970018" seqno="11514"> From: davido@apocalypse.org (David Ornstein) Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: Tue, 21 Nov 1995 19:13:38 GMT Organization: InterNex Information Services 1-800-595-3333 Message-ID: <48t8d7$fr6@voyager.Internex.NET> References: <9511171331.AA12385@fly.HiWAAY.net> <48qmfi$14r@bell.maths.tcd.ie> tim@maths.tcd.ie (Timothy Murphy) wrote: >Len Bullard \<cbullard@HiWAAY.net> writes: >>The numbers of people in the US signed up >>to the Internet still remains a low percentage although growing daily. >Is this true? >I read the following statistics last week. [clip] For a *really* thoughtful analysis of all these recent surveys about how many folks are on the Internet, I recommend Jack's fantastic editorial this month in Boardwatch magazine. \<URL:http://www.boardwatch.com/mag/95/dec/bwm1.htm> --------------------------------------------- David Ornstein BrowserCaps: http://objarts.com/bc Outbreak: http://ichiban.objarts.com/outbreak-unreg Personal Info: http://objarts.com/davido </message> <message id="<power.58.30B23669@inch.com>" date="3025974506" seqno="11515"> From: power@inch.com (Power Media Communications) Newsgroups: comp.sys.ti,comp.sys.transputer,comp.terminals,comp.text,comp.text.desktop,comp.text.frame,comp.text.interleaf,comp.text.sgml,comp.text.tex,comp.theory Subject: * Experienced Video Telecommunications Programmer Wanted * Date: Tue, 21 Nov 1995 20:28:26 GMT Organization: The Internet Channel Message-ID: \<power.58.30B23669@inch.com> New York City based company seeks experienced programmer to work on new innovative video teleconferencing application. Should have knowledge of the internet, web servers, video teleconferencing, dialogic boards, and general programming skills. Competitive compensation. [ Power Media Communications ] Please email resume and salary requirements to: power@inch.com Or fax to: (212) 223-8765 </message> <message id="<11-21-1995.21451@mbfw451>" date="3025974726" seqno="11516"> Organization: MBF Systems Corporation Subject: Re: Slight flame on SGML References: <9511171331.AA12385@fly.hiwaay.net> <48qmfi$14r@bell.maths.tcd.ie> Date: Tue, 21 Nov 95 12:32:06 -0800 From: ricday@mbfw451.mlnet.com (Richard S. Day) Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla Message-ID: <11-21-1995.21451@mbfw451> Followup-To: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla In article <48qmfi$14r@bell.maths.tcd.ie> tim@maths.tcd.ie (Timothy Murphy) writes: >Len Bullard \<cbullard@HiWAAY.net> writes: > >>The numbers of people in the US signed up >>to the Internet still remains a low percentage although growing daily. > >Is this true? >I read the following statistics last week. > >============================================= > > The Electronic Bulletin of the > GLObal Systems Analysis and Simulation Association > in the USA > > Email distribution: GLOSAS@vm1.mcgill.ca > > According to a recent Internet Society survey, there are about 3.4 >million "host" computers hooked up to the Internet in the United States and >just over 500,000 in Western Europe. > > WHO'S ONLINE? > > * 37 million, or 17% of the U.S. and Canadian population 16 and older, > have access to the Internet, > * 24 million used the Internet in the past three months, > * They spend an average of 5 hr. 28 min. on the Internet per week, > * 34% are women, > * 66% reached the Internet from work, > * 25% of World Wide Web users have incomes of more than $80,000. > > (TIME, Nov. 13, 1995, Page 121) > >============================================= Timothy The problem with such surveys is the fundamental question of accuracy. If someone taking the survey you describe had called me, I would have said * I am older than 16 * I have used the net within the past three months (hours...) * I (measured by the activity on my accounts) spend more than 5.5 hrs a month online * I am not female * I reach the net from work (and home) * my income is above $80,000 So, I would "confirm" the survey results. But a closer look would show: I work from home, so "home" and "business" are the same. I have four connections to the net; each is probably used 5.5 hours or more per month. So far, so good. But -- 60 to 70 percent of "my" online time is actually time spent by my teenage kids, who use it for homework research, emailing friends and in my son's case, reinforcing his math and programming skills. And my wife is online, corresponding with friends overseas. Time magazine and others put together these surveys to promote the notion of the net as a good place to do business. So they look for the sort of profile suggested by the above survey. Most of my friends' usage patterns correspond with mine: personal and business access, but most of the "volume" actually generated by their kids. I suggest while there are a large number of people with internet access in North America today, the majority are (a) casual users through services such as AOL; and (b) far younger and less affluent than the net-as-commerce proponents would have one believe. If Len Bullard is suggesting the numbers of serious users signed up is still low, I would have to agree with him. Most of "the numbers" being bandied about are young, occasional users; I suspect many are mindless "explorers" and/or "creators" of equally mindless "home pages". Thankfully, most have managed to steer clear of comp.text.sgml and the few intelligent discussions left on the net. Ric (a refugee from Cork and most things Irish) </message> <message id="<48tc79$clo@rap.SanDiegoCA.ATTGIS.COM>" date="3025973929" seqno="11517"> From: swf@newselsegundoca.attgis.com (Stan Friesen) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 21 Nov 1995 20:18:49 GMT Organization: NCR Teradata Database Business Unit Message-ID: <48tc79$clo@rap.SanDiegoCA.ATTGIS.COM> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> <48q6u9$5d3@dove.nist.gov> Reply-To: swf@elsegundoca.ncr.com In article <48q6u9$5d3@dove.nist.gov>, bagwill@kangaroo.ncsl.nist.gov (Robert Bagwill) writes: |> |> Baloney [replace with British equivalent]. |> |> All publishers want control. All authors raised on WYSIWYG wordprocessors |> and desktop publishing want control. Resistance is futile. HTML will |> be assimilated [stupid but appropriate STTNG Borg reference]. Unfortunately, they will not get what they want - it is *physically* *impossible. Disparate platforms *cannot* produce the same layout. Merely changing the monitor, while keeping the colormaps, can make a noticable change in appearance. Colors that look good on one monitor may be ugly on another. -- swf@elsegundoca.attgis.com sarima@netcom.com The peace of God be with you. </message> <message id="<48tn6m$5lu@news.INbe.net>" date="3025985174" seqno="11518"> From: "M.T. Carrasco Benitez" \<carrasco@innet.lu> Newsgroups: comp.infosystems.www.misc,comp.std.internat,comp.software.international,comp.text,comp.text.sgml,sci.lang,sci.lang.translation Subject: WInter (Web Internationalization & Multilinguism) Date: 21 Nov 1995 23:26:14 GMT Organization: INnet NV (post doesn't reflect views of INnet NV) Message-ID: <48tn6m$5lu@news.INbe.net> WInter (Web Internationalization & Multilinguism) The mailing-list for WInter is now in place. Administration first: 1) Unsubscribe majordomo@dorado.crpht.lu unsubscribe winter 2) Subscribe majordomo@dorado.crpht.lu subscribe winter 3) Help (It is run majordomo) majordomo@dorado.crpht.lu help 4) Post to the list winter@dorado.crpht.lu 5) Address of the list owner carrasco@coctel.lu 6) WInter Web page http://www.echo.lu/other/winter I propose the following preliminary objectives: 1) The WInter Manifesto A short introductory document pointing in the general direction. It can be updated as we go along. 2) The agenda for Boston 3) The production of a working document I thank the Centre de Recherche Public Henri Tudor for hosting the list. http://www.crpht.lu I apologize for the delay. Tomas </message> <message id="<61rh209z@qnx.com>" date="3025968886" seqno="11533"> Newsgroups: comp.text.sgml From: chrish@qnx.com (Chris Herborth) Subject: Re: SGML -> HTML Message-ID: <61rh209z@qnx.com> Date: Tue, 21 Nov 95 18:54:46 GMT Organization: QNX Software Systems References: <48g54e$8d3@news.INbe.net> In article <48g54e$8d3@news.INbe.net>, M.T. Carrasco Benitez \<carrasco@innet.lu> wrote: > I am looking for SGML->HTML filter, or a pointer to the FAQ where this > kind of answer could be. Mail prefered. This depends totally on the SGML DTD being used to write the documents; it would probably be impossible (or at least impractical) to write a general-purpose SGML -> HTML converter. We're currently using OmniMark to convert our SGML documents into HTML for online publishing; we used to use perl, which is probably suitable if your documents aren't very complex (or maybe even if they are, depending on how much of a perl fan you are). -- ----------========================================================---------- Chris Herborth, R\&D Technical Writer Arcane Dragon chrish@qnx.com QNX Software Systems, Ltd. -==(UDIC)==- ||| JAGUAR http://www.qnx.com/~chrish/ DNRC / | \\ 64-bit </message> <message id="<tim-2111952119120001@slug-ppp17.wustl.edu>" date="3025999152" seqno="11519"> From: tim@michelob.wustl.edu (Tim Rand) Newsgroups: comp.text.sgml Subject: Numeric Tags and Parsing Errors Date: Tue, 21 Nov 1995 21:19:12 -0600 Organization: St. Louis Unix Users Group Message-ID: \<tim-2111952119120001@slug-ppp17.wustl.edu> We have recently run across a problem were codes that begin with numerics are passed through parsing, yet are invalid. These codes are not defined in the DTD, so we suspect that the problem is related to our declaration (included below), but we aren't sure. We have tried the validation using both commercial parsing tools as well as NSGMLS and all behave the same. Feedback as to were to look for the source of the problem would be appreciated. The DTD is not structured for codes starting with numerics. However, when validating instance files, there were some numeric codes that passed through. Where in the declaration can this be controlled and how? Example: <3a> passed through and it isn't defined in the DTD. A copy of the declaration is listed below: \<!SGML "ISO 8879-1986" -- Default SGML declaration using the Reference concrete syntax -- CHARSET BASESET "ISO 646-1983//CHARSET International Reference Version (IRV)//ESC 2/5 4/0" DESCSET 0 128 0 128 128 UNUSED CAPACITY PUBLIC "ISO 8879-1986//CAPACITY Reference//EN" SCOPE DOCUMENT SYNTAX SHUNCHAR CONTROLS 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 127 255 BASESET "ISO 646-1983//CHARSET International Reference Version (IRV)//ESC 2/5 4/0" DESCSET 0 128 0 FUNCTION RE 13 RS 10 SPACE 32 TAB SEPCHAR 9 NAMING LCNMSTRT "" UCNMSTRT "" LCNMCHAR "-." -- Lower-case hyphen, period are -- UCNMCHAR "-." -- same as upper-case (45 46). -- NAMECASE GENERAL YES ENTITY NO DELIM GENERAL SGMLREF SHORTREF SGMLREF NAMES SGMLREF QUANTITY SGMLREF NAMELEN 15 FEATURES MINIMIZE DATATAG YES OMITTAG YES RANK YES SHORTTAG YES LINK SIMPLE NO IMPLICIT YES EXPLICIT NO OTHER CONCUR NO SUBDOC NO FORMAL YES APPINFO NONE> Your assistance appreciated. -- Tim Rand \<tim@michelob.wustl.edu> Checked regularly All messages will get a response. No black holes here! </message> <message id="<48ts0d$669@nnews.ims.com>" date="3025990093" seqno="11520"> From: Steve Homer \<shomer> Newsgroups: comp.text.sgml Subject: Pinnacles info needed Date: 22 Nov 1995 00:48:13 GMT Organization: Integrated Measurement Systems, Inc. Message-ID: <48ts0d$669@nnews.ims.com> I'm looking for information about the Pinnacles Component Information Standard. Ideally I'd like to get in touch with engineers who have worked with Pinnacles documents and have used IMS or similar test equipment. Also, any pointers to information about Pinnacles or to sample data sheets would be appreciated. ________________________________________________________ Steve Homer shomer@ims.com (503)626-5453 Senior Technical Writer Integrated Measurement Systems 9525 SW Gemini Drive, Beaverton, Oregon 97008 </message> <message id="<48v78e$j07@cliffy.lfwc.lockheed.com>" date="3026034382" seqno="11521"> From: mcclellantj@harrier (Tad McClellan) Newsgroups: comp.text.sgml Subject: Re: VRML Behaviors Date: 22 Nov 1995 13:06:22 GMT Organization: Lockheed Martin Tactical Aircraft Systems Message-ID: <48v78e$j07@cliffy.lfwc.lockheed.com> References: <9511201830.AA12588@fly.HiWAAY.net> Reply-To: mcclellantj@lfwc.lockheed.com Len Bullard (cbullard@HiWAAY.net) wrote: : [snip] Nature never worked out a way : to prevent this. In fact, it is the engine of evolution, and it evolves : superb cooperators, like ants. And miscre-ants ;-) -- Tad McClellan, Logistics Specialist (IETMs and SGML guy) email: mcclellantj@lfwc.lockheed.com Interesting trivia: If you took all the sand in North Africa and spread it out... it would cover the Sahara desert. </message> <message id="<48vf7f$nuc@dove.nist.gov>" date="3026042543" seqno="11522"> From: bagwill@kangaroo.ncsl.nist.gov (Robert Bagwill) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 22 Nov 1995 15:22:23 GMT Organization: NIST Message-ID: <48vf7f$nuc@dove.nist.gov> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> <48q6u9$5d3@dove.nist.gov> <48tc79$clo@rap.sandiegoca.attgis.com> In article <48tc79$clo@rap.sandiegoca.attgis.com>, swf@newselsegundoca.attgis.com (Stan Friesen) writes: >Unfortunately, they will not get what they want - it is *physically* >*impossible. Disparate platforms *cannot* produce the same layout. >Merely changing the monitor, while keeping the colormaps, can make >a noticable change in appearance. Colors that look good on one monitor >may be ugly on another. There are fixes for that, but I don't think most people care too much. -- Bob Bagwill \<rbagwill@nist.gov> </message> <message id="<SMISHRA.95Nov22112944@gaia.cc.gatech.edu>" date="3026046584" seqno="11523"> From: smishra@gaia.cc.gatech.edu (Sunil Mishra) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 22 Nov 1995 16:29:44 GMT Organization: Bizarre Inc. Message-ID: \<SMISHRA.95Nov22112944@gaia.cc.gatech.edu> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> <48q6u9$5d3@dove.nist.gov> \<SMISHRA.95Nov20154436@gaia.cc.gatech.edu> <48t627$fr6@voyager.Internex.NET> Reply-To: smishra@cc.gatech.edu In-reply-to: davido@apocalypse.org's message of Tue, 21 Nov 1995 18:33:39 GMT In article <48t627$fr6@voyager.Internex.NET> davido@apocalypse.org (David Ornstein) writes: \\\\ \\\\ At the risk of stepping out having not read all the important \\\\ documents in detail, let me ask about style sheets and markup. It has \\\\ seemed to me that the CLASSes being added to the various tags (e.g., \\\\ \<EM class=Person>) are really a poor-man's way to add new tags (e.g., \\\\ \<Person>). Is this basically right? No, not really. A style sheet is a way to specify how you want your documents to be presented to the reader. The reader should have the option to override your choices. If you construct a new tag \<person>, then the reader would have no way of specifying her defaults for the tag. If, on the other hand, you said \<em class=person>, the defaults specified by the reader for \<em> could in theory override what you specify. Also, a reader with a browser that does not possess a style sheet would stil be able to produce a reasonable representation of \<em>, just not precisely how you, the author, wanted it. \\\\ Ideally (of course?) browsers would support displaying custom tags \\\\ defined in DTDs that are specific to the documents being viewed. I \\\\ know they don't do this (I assume because it's a lot of work and they \\\\ have too many other things on thier minds), but why didn't the style \\\\ sheet folks design something that was a bit more in line (like \\\\ allowing the style sheet information to define the new tags for a \\\\ browser using stuff in the header -- thus avoiding the full complexity \\\\ of DTDs)? This is certainly a possibility that has been brought up before, and I don't know enough to say much about it. Sunil </message> <message id="<DIG9zL.uy@sunsrvr6.cci.com>" date="3026041617" seqno="11524"> Newsgroups: comp.text.sgml From: rs@cci.com (Rick Silterra) Subject: HyTime and OpenDoc and OLE Message-ID: \<DIG9zL.uy@sunsrvr6.cci.com> Keywords: HyTime,OpenDoc,OLE Sender: root@sunsrvr6.cci.com (Operator) Organization: Northern Telecom, Network Application Systems Date: Wed, 22 Nov 1995 15:06:57 GMT I have been asked the question: How do HyTime, and OpenDoc relate? I was somewhat flummoxed. My first reaction is that HyTime is 1) HyTime is an interchange format, which describes links and time and space. 2) OpenDoc/OLE is a presentation format which describes the display of information - it is the multi-media equivalent of a wysiwyg editor. 3) A HyTime document might serve as an interchange format for an OpenDoc document. Since I do not understand HyTime deeply, and I understand OpenDoc only from marketing write-up, I am not at all sure about whether these comments make any sense. Thanks in advance for any comments on this. Rick Silterra -- Rick Silterra Usenet: rs@cci.com Mail: Rick Silterra Northern Telecom Network Applications Systems Group 97 Humboldt St. Rochester NY 14609-7493 Any opinions expressed above are my own, and not those of any group with which I am associated. </message> <message id="<9511221357.AA18002@fly.HiWAAY.net>" date="3026037476" seqno="11525"> Newsgroups: comp.text.sgml Date: 22 Nov 1995 07:57:56 -0600 Message-ID: <9511221357.AA18002@fly.HiWAAY.net> From: Len Bullard \<cbullard@HiWAAY.net> Subject: Whose Online? [Was: Slight Flame on SGML] At the risk of keeping this issue alive, here are stats from an Associated Press article in the local paper of last night. The source is Rand Corp: o It will cost $1 billion to provide electronic mail to 10 percent of all Americans (sorry about the centrism) o There are roughly 6.7 million consumer electronic mailboxes vs business mailboxes o America Online says that only 8 to 9 % of American homes "surf the net" or subscribe to an online service o 37% of American homes have a computer. The report states that all american homes should have online service as the economic benefit is there and the social disadvantages of not having one will accrue. Sounds like the business is solid. ;-) Now we have to get back to the features war that provides the discriminators that as they say, "gets rears into the seats." Ah well, Montreal next week and with this bloody cold.... sniff.... sniff.... sniff... </message> <message id="<DIGuF7.L7w@mv.mv.com>" date="3026068098" seqno="11526"> Newsgroups: comp.text.sgml From: kaikow@standards.com Subject: Is there a thing such as http://www.sgmlopen.home? Originator: kaikow@mv.mv.com Message-ID: \<DIGuF7.L7w@mv.mv.com> Sender: kaikow@standards.com Reply-To: kaikow@standards.com Organization: MV Communications, Inc. Date: Wed, 22 Nov 1995 22:28:18 GMT Followup-To: kaikow@standards.com Or something like that? </message> <message id="<DIGqy3.4IC@mail.appliedmicro.ns.ca>" date="3026063594" seqno="11527"> Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml From: Richard Akerman \<akerman@appliedmicro.ns.ca> Subject: !DOCTYPE (for HTML 3.0 Strict) Message-ID: \<DIGqy3.4IC@mail.appliedmicro.ns.ca> Sender: nobody@mail.appliedmicro.ns.ca Organization: Applied Microelectronics Date: Wed, 22 Nov 1995 21:13:14 GMT I want to put the correct DOCTYPE at the top of my HTML documents. I am trying to write strict HTML 3.0 The main HTML 3.0 things I am using are \<P ALIGN=CENTER> \<BIG> i.e. just a few of the minor extensions supported by current browsers What is the syntax for the quoted part of the SGML DOCTYPE declaration? I understand \<!DOCTYPE HTML PUBLIC but what about "-//IETF//DTD HTML 2.0//EN"> The best I can parse this is: "official standard #//issuing authority//what it is//language//?unknown?" RFC 1866 also suggests a very wide variety of possibilities for the "what it is" section, including DTD HTML DTD HTML Level 1 DTD HTML Level 2 DTD HTML Strict DTD HTML Strict Level 1 DTD HTML Strict Level 2 DTD HTML 2.0 Level 1 DTD HTML 2.0 Level 2 DTD HTML 2.0 Strict DTD HTML 2.0 Strict Level 1 I would think then that if I am writing general strict HTML I can just use \<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML Strict//EN"> but then my page will fail on HALsoft validation, but I guess this is because HALsoft won't do "this is valid HTML of some version" validation? HALsoft will validate if I use a DOCTYPE of ... //DTD HTML 3.0// but what is the status of using such a DOCTYPE, since I understand HTML 3.0 has been withdrawn and is being re-proposed as a set of smaller standards (i18n, tables, etc.). Is it valid to use \<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 3.0 Strict//EN"> instead of this, which HALsoft likes: \<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 3.0//EN//" [ \<!ENTITY % HTML.Recommended "INCLUDE"> ] > which incidentally Netscape parses messily, leaving a ] > on the screen, which is I guess understandable since Netscape isn't a SGML-based parser. -- Richard Akerman | Akerman@AppliedMicro.NS.Ca LAN Systems Manager | Applied Microelectronics Inc. | Halifax, N.S., Canada | </message> <message id="<4908m2$1qq@franklin.cc.utas.edu.au>" date="3026068610" seqno="11528"> From: John Lamp \<John.Lamp@cs.utas.edu.au> Newsgroups: comp.text.sgml Subject: Re: Where can I find the ISO char sets Date: 22 Nov 1995 22:36:50 GMT Organization: University of Tasmania, Department of Computer Science Message-ID: <4908m2$1qq@franklin.cc.utas.edu.au> References: \<DHF9Lv.M4z@oasis.icl.co.uk> <30B106FE.5E85@novell.com> Ben Brimhall \<bbrimhall@novell.com> wrote: >I work with SGML on a regular basis, and was roaming around on the >internet searching for entity lists. I found a page that had what I >wanted it is at "http://www.alis.com:8085/ietf/html/index.html" There's a link to an entity store at http://lamp.cs.utas.edu.au/net.html Cheers John -- _--_|\\ John Lamp, originating in Hobart, Tasmania / \\ Phone: 002 20 2375 - Fax: 002 20 2913 \\_.--._/ email: John.Lamp@cs.utas.edu.au v <--<< http://lamp.cs.utas.edu.au/jw_lamp.html Information Systems Group, Dept of Comp Science, Uni of Tasmania </message> <message id="<sjd-2211951806500001@sulia.ebt.com>" date="3026070410" seqno="11529"> From: sjd@ebt.com (Steven J. DeRose) Newsgroups: comp.text.sgml Subject: Re: Numeric Tags and Parsing Errors Date: Wed, 22 Nov 1995 18:06:50 -0500 Organization: EBT, Inc. Message-ID: \<sjd-2211951806500001@sulia.ebt.com> References: \<tim-2111952119120001@slug-ppp17.wustl.edu> In article \<tim-2111952119120001@slug-ppp17.wustl.edu>, tim@michelob.wustl.edu (Tim Rand) wrote: > We have recently run across a problem were codes that begin with numerics > are passed through parsing, yet are invalid. These codes are not defined > in the DTD, so we suspect that the problem is related to our declaration > (included below), but we aren't sure. We have tried the validation using > both commercial parsing tools as well as NSGMLS and all behave the same. >... > Example: <3a> passed through and it isn't defined in the DTD. This case might be a bit puzzling the first time you run into it, but if I'm understanding your description right, this is SGML's correct defined behavior for that case. "<3a>" is valid given the reference concrete syntax, but is not a tag. Figure 3 specifies that STAGO (the start-tag open delimiter, commonlly known as "<") is recognized only when the contextual constraint "GI" applies. Clause 9.6.2 defines the "GI" contextual constraint, the relevant part being that the delimiter must be followed by a name start character. In the SGML declaration you thoughtfully supplied the usual naming rules are in effect, so "3" is a name character but not a name *start* character (that is, names like GIs can't start with it). So in "<3a>" the "<" is not a STAGO in context, and is just data. Assuming #PCDATA is allowed at the points in question, that won't produce an error. If you view your data with a tool that shows you the result of parsing (via a tree or structure view, an ESIS output stream, or some such), I think you'll find that <3a> is not being taken as a start-tag, but just as literal data. If you want <3a> as a tag, you'd need to add digits to the list of name start characters, and then declare the element type as usual. Hope that helps. Steven J. DeRose, Ph.D. Sr. Systems Architect, EBT sjd@ebt.com </message> <message id="<48vu69$dt6@rap.SanDiegoCA.ATTGIS.COM>" date="3026057865" seqno="11531"> From: swf@newselsegundoca.attgis.com (Stan Friesen) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 22 Nov 1995 19:37:45 GMT Organization: NCR Teradata Database Business Unit Message-ID: <48vu69$dt6@rap.SanDiegoCA.ATTGIS.COM> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> <48q6u9$5d3@dove.nist.gov> \<SMISHRA.95Nov20154436@gaia.cc.gatech.edu> <48t627$fr6@voyager.Internex.NET> Reply-To: swf@elsegundoca.ncr.com In article <48t627$fr6@voyager.Internex.NET>, davido@apocalypse.org (David Ornstein) writes: |> |> At the risk of stepping out having not read all the important |> documents in detail, let me ask about style sheets and markup. It has |> seemed to me that the CLASSes being added to the various tags (e.g., |> \<EM class=Person>) are really a poor-man's way to add new tags (e.g., |> \<Person>). Is this basically right? Yes - with the caveat that they *also* provide a way to specify a reasonable default behavior. That is using, or instance, \<H2 CLASS="appendix"> says that this is an appendix title, *and* that it can be safely considered to be equivalent to a second level header without doing damage to the document content. Saying \<APPENDIX> says this is an appendix, and who the heck *knows* how to treat it if it cannot be treated as an appendix title. This difference - the mechanism for attaching default meaning - is the critical one that makes it a useful and *safe* means of extending HTML. The lack of this is why \<CENTER> was such a bad idea. What could this man to a browser that doesn't have this extension? On the other hand using \<P align=CENTER> at least says to treat the text as a paragraph even if centering is not supported. -- swf@elsegundoca.attgis.com sarima@netcom.com The peace of God be with you. </message> <message id="<vwj7n0sl3bh.fsf@osfb.aber.ac.uk>" date="3026066290" seqno="11536"> From: pcg@aber.ac.uk (Piercarlo Grandi) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 22 Nov 1995 21:58:10 +0000 Organization: Prifysgol Cymru, Aberystwyth Sender: pcg@osfb.aber.ac.uk Message-ID: \<vwj7n0sl3bh.fsf@osfb.aber.ac.uk> References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48vf7f$nuc@dove.nist.gov> Reply-To: pcg@aber.ac.uk (Piercarlo Grandi) In-reply-to: bagwill@kangaroo.ncsl.nist.gov's message of 22 Nov 1995 15:22:23 GMT Precedence: first-class >>> On 22 Nov 1995 15:22:23 GMT, bagwill@kangaroo.ncsl.nist.gov (Robert >>> Bagwill) said: Robert> In article <48tc79$clo@rap.sandiegoca.attgis.com>, Robert> swf@newselsegundoca.attgis.com (Stan Friesen) writes: Stan> Unfortunately, they will not get what they want - it is *physically* Stan> *impossible. Disparate platforms *cannot* produce the same layout. Stan> Merely changing the monitor, while keeping the colormaps, can make Stan> a noticable change in appearance. Colors that look good on one monitor Stan> may be ugly on another. Robert> There are fixes for that, but I don't think most people care too Robert> much. No, there are not fixes at all. a color management scheme will only make colors that can be represented in the palette look the same. But this only if they can fit in the palette of your display device -- if they cannot dithering or nearest color or whatever will have to be done, and this can awful consequences. Consider: dixy> From: dcreelma@gate.net (dixy flatline) dixy> Subject: Background Color Problem dixy> Date: Sat, 28 Oct 95 17:56:18 GMT dixy> [ ... ] I have been using background colors on the pages I am dixy> creating, as many do. On my computer using netscape with a SVGA dixy> monitor the background color for any web page I create looks dixy> fine to me. But I was over a guys office the other day, dixy> unfortunately the guy who is paying me to write his organization dixy> a web page, and we used his computer running netscape to access dixy> one of my pages. The background color appeared textured and not dixy> smooth. We could not even read the text on the page. Now I am dixy> 99.9% sure he has a SVGA monitor but he may have been using dixy> Netscape 1.1 and not the newer version I use. [ ... ] What do you do in this case? \<BODY BGCOLOR="#xxxxxx" SVGACARD="hdp://freebie.org/pc/svga/24bit.pci"> (hdp of course is the "hardware download protocol") seems rather hard to do. </message> <message id="<4903v3$2ne@alpha.aladdin.co.uk>" date="3026092512" seqno="11530"> From: nikb@aladdin.co.uk (Nick barron) Newsgroups: comp.text.sgml Subject: Example of IETMs Date: Thu, 23 Nov 1995 05:15:12 GMT Organization: SoNet - The first Internet provider on the south coast Message-ID: <4903v3$2ne@alpha.aladdin.co.uk> Can anyone point me at some examples or further information of IETMs to MIL-D-87269? I'm specifically interested in examples of the use of the precond/postcond elements. Any information or pointers to on-line sources would be gratefully received. </message> <message id="<491tij$2m9@news.Belgium.EU.net>" date="3026122771" seqno="11532"> From: Jacques Deseyne \<jad@sema.be> Newsgroups: comp.text.sgml Subject: Re: Is there a thing such as http://www.sgmlopen.home? Date: 23 Nov 1995 13:39:31 GMT Organization: SEMA Group Belgium Message-ID: <491tij$2m9@news.Belgium.EU.net> kaikow@standards.com wrote: Subject: Re: Is there a thing such as http://www.sgmlopen.home? > >Or something like that? The home page for SGML Open is reachable through http://www.sgmlopen.org I am not sure whether "home" would be an appropriate name for an Internet domain. Best regards, ------------------------------------------------------ Jacques Deseyne \<jad@sema.be> Sema Group Belgium - Stallestraat 96 - B-1180 Brussels ------------------------------------------------------ </message> <message id="<492g3n$p9d@dirac.isa.de>" date="3026141751" seqno="11534"> From: taube@isa.de (Christian Taube) Newsgroups: comp.text.sgml Subject: Mapping RTF hex codes to ISO Latin entities? Date: 23 Nov 1995 19:55:51 +0100 Organization: ISA GmbH, Azenbergstr. 35, D-70174 Stuttgart, FRG Message-ID: <492g3n$p9d@dirac.isa.de> Keywords: RTF, ISO Latin, mapping, character sets Sorry to bug you guys again with stupid RTF stuff, but I need to work this out ASAP ... :-( I'm looking for some sort of table that lists all the RTF hexadecimal codes -- the ones that look like \\'\<xx> and are replaced by special characters and "Umlaute" according to The Microsoft Way when you display the document -- and their corresponding ISO 8859 entities. Does anyone know if and where such a beast exists? (No, I don't regularly work with RTF ... :-) Regards. Christian -- Christian Taube -- \<A HREF="http://www.isa.de/~taube">taube@isa.de\</A> Any opinions expressed are mine, and mine only! Thought for today: These are the children of the race "Enterprise." Her 5-year mission - to logically go where no Romulan has gone before. </message> <message id="<1995Nov23.225010.13970@maroki.netzservice.de>" date="3026155810" seqno="11535"> Newsgroups: comp.text.sgml From: maro@maroki.netzservice.de (Martin Rost) Subject: linuxdoc Organization: Privat-Site Message-ID: <1995Nov23.225010.13970@maroki.netzservice.de> Date: Thu, 23 Nov 1995 22:50:10 GMT I am enthusiastic already long by the linuxdoc-package, even I am not really familiar with sgml. I have told some people of it, which don't want to work in themselves for that reason in Linux, but have comprehended the possibilities of sgml. Now my question: Does there is perhaps already a port of the whole linuxdoc-package to DOS? I am not looking only for an sgml-compiler for DOS. I personally dispose of no C-compiler under DOS, so that I could dare it to compile the whole sources by myself. Greeting, Martin -- Martin Rost / Germany, Kiel / maro@maroki.netzservice.de </message> <message id="<49318u$l9u@artemis.it.luc.edu>" date="3026159326" seqno="11537"> From: tkaczma@orion.it.luc.edu (Tomasz J. Kaczmarski) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Followup-To: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Date: 23 Nov 1995 23:48:46 GMT Organization: Loyola University Chicago Message-ID: <49318u$l9u@artemis.it.luc.edu> References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48q6u9$5d3@dove.nist.gov> Robert Bagwill (bagwill@kangaroo.ncsl.nist.gov) wrote: : Resistance is futile. HTML will : be assimilated [stupid but appropriate STTNG Borg reference]. Great quote ;) I'll add it to my list. -- ? ''~`` ( o o ) +----------------------------------.oooO--(_)--Oooo.------------------------+ mailto:tkaczma@luc.edu http://orion.it.luc.edu/~tkaczma .oooO (312) 512-7302 ( ) Oooo. +-------------------------------------\\ (----( )--------------------------+ \\_) ) / (_/ </message> <message id="<vwjag5ne2yy.fsf@osfb.aber.ac.uk>" date="3026131189" seqno="11540"> From: pcg@aber.ac.uk (Piercarlo Grandi) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 23 Nov 1995 15:59:49 +0000 Organization: Prifysgol Cymru, Aberystwyth Sender: pcg@osfb.aber.ac.uk Message-ID: \<vwjag5ne2yy.fsf@osfb.aber.ac.uk> References: <47d5ro$o3j@hamilton.maths.tcd.ie> \<vwjivkxguhs.fsf@osfb.aber.ac.uk> <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \<DI79G4.B42@udcf.gla.ac.uk> <48q6u9$5d3@dove.nist.gov> \<SMISHRA.95Nov20154436@gaia.cc.gatech.edu> <48t627$fr6@voyager.Internex.NET> <48vu69$dt6@rap.SanDiegoCA.ATTGIS.COM> Reply-To: pcg@aber.ac.uk (Piercarlo Grandi) Precedence: first-class >>> On 22 Nov 1995 19:37:45 GMT, swf@newselsegundoca.attgis.com (Stan >>> Friesen) said: Stan> In article <48t627$fr6@voyager.Internex.NET>, Stan> davido@apocalypse.org (David Ornstein) writes: David> At the risk of stepping out having not read all the important David> documents in detail, let me ask about style sheets and markup. It has David> seemed to me that the CLASSes being added to the various tags (e.g., David> \<EM class=Person>) are really a poor-man's way to add new tags (e.g., David> \<Person>). Is this basically right? In practice yes: in theory it is a monstrosity. CLASSes ought to be used exclusively to differentiate between different flavours of the same thing, not to extend stealthily the range of HTML elements. Even if some HTML3 papers have examples like '\<DIV CLASS=abstract>', it is an abomination. The main problem is that the values that the 'CLASS' attribute can take are not part of the HTML standard at all, therefore they are arbitrary, and therefore they should be only used to specify rendering possibilities, not structure (for structure _must_ be document-independent). Stan> Yes - with the caveat that they *also* provide a way to specify a Stan> reasonable default behavior. Let me add: a reasonable _HTML_ default behaviour. Because one of the important consideration in the HTML3 design has been backwards compatibility with HTML2 implementations. This has obvious benefits in terms of continuity, as compared to "OK, HTML was small and fun, good to start with, but now there is no problem in switching to full SGML-capable browsers". The obvious snag is that one is then encouraged to do things like: Stan> That is using, or instance, \<H2 CLASS="appendix"> says that this Stan> is an appendix title, *and* that it can be safely considered to be Stan> equivalent to a second level header without doing damage to the Stan> document content. Which, as argued above, strictly speaking, are horrors. Stan> Saying \<APPENDIX> says this is an appendix, and who the heck *knows* Stan> how to treat it if it cannot be treated as an appendix title. Which would be the right thing to do, _if_ \<APPENDIX> were part of HTML's content model. Stan> This difference - the mechanism for attaching default meaning - is Stan> the critical one that makes it a useful and *safe* means of Stan> extending HTML. Again, the point is that _really_ 'CLASS' should not imply any new "meaning", just customization of an existing thingie. Stan> The lack of this is why \<CENTER> was such a bad idea. What could Stan> this man to a browser that doesn't have this extension? On the Stan> other hand using \<P align=CENTER> at least says to treat the text Stan> as a paragraph even if centering is not supported. This works because 'ALIGN' has no semantic connotation. '\<P>' is a paragraph however it is aligned. But consider instead \<P CLASS=reference> \<P CLASS=quote> \<P CLASS=title> If this usage spreads (which in some measure is inevitable), then HTML could be replaced by HTML(-1), with a single \<THINGIE CLASS={href,h1,title, head,p,...}> construct. :-) </message> <message id="<DIJ8sB.1Lq@mv.mv.com>" date="3026180026" seqno="11538"> Newsgroups: comp.text.sgml From: kaikow@standards.com Subject: Re: SGML book? Originator: kaikow@mv.mv.com Message-ID: \<DIJ8sB.1Lq@mv.mv.com> Sender: kaikow@standards.com Reply-To: kaikow@standards.com Organization: MV Communications, Inc. Date: Fri, 24 Nov 1995 05:33:46 GMT References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI973L.9CE@mv.mv.com> <1995Nov20.045420.29244@sq.com> <1995Nov20.214942.10037@sq.com> Followup-To: kaikow@standards.com I agree, but a few bookstores stock SGML books. Others will get copies of most books for one to look at with no obligation to buy, however, most will only do that if the book is available thru their supplier. Alas, Practical SGML is not so available. I don't like buying any book without first seeing it, but SGML books are just not on store shelves everywhere. Now, if you want HTML, there are many to choose from! In article <1995Nov20.214942.10037@sq.com>, Liam R. E. Quin \<lee@sq.com> wrote: >A couple of people have asked, so I just wanted to point out that the >comments I made about various SGML books are my initial opinions based >on a quick look (as I said in the article, but people asked anyway!). > >Please, if you're going to buy books, look at them, and read reviews! >We're very fortunate that there are quite a few books on SGML, and more >seem to be on the way, so you can choose ones that will suit you. > >Lee > >-- >Liam Quin, SoftQuad Inc +1 416 239 4801 lee@sq.com \<URL:http://www.sq.com/> >HexSweeper NeWS game;OPEN LOOK+XView+mf-fonts FAQs;lq-text unix text retrieval </message> <message id="<4940fd$anb@usenet1.interramp.com>" date="3026191275" seqno="11539"> From: Goldfarb@InterRamp.com (Charles F. Goldfarb) Newsgroups: comp.text.sgml Subject: Re: SGML book? Date: Fri, 24 Nov 1995 08:41:15 GMT Organization: Information Management Consulting Message-ID: <4940fd$anb@usenet1.interramp.com> References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI625t.F6M@curia.ucc.ie> \<DI973L.9CE@mv.mv.com> <1995Nov20.045420.29244@sq.com> Reply-To: Goldfarb@InterRamp.com lee@sq.com (Liam R. E. Quin) wrote: I would appreciate hearing about any errors in README.1ST or other books in my series as I am not aware of any. Please write to Goldfarb@SGMLSource.com. Thanks, Charles Goldfarb </message> <message id="<199511241139.MAA22270@cygne.ais.berger-levrault.fr>" date="3026201962" seqno="11541"> Newsgroups: comp.text.sgml Date: 24 Nov 1995 12:39:22 +0100 From: St�phane Bornnerand \<sbor@Berger-Levrault.fr> Message-ID: <199511241139.MAA22270@cygne.ais.berger-levrault.fr> References: <47togkINNoc1@oasys.dt.navy.mil> Subject: Re: SGML Document Management System Reply to the following article based on the product ADOC(TM) developed both by AIS S.A. and the AVIONICS and SYSTEMS Division of AEROSPATIALE (France). I directly put the answers to your questions within the text and additional comments at the end of the message. [Lori Westbrook] | Does anyone know of any software that could be used to build an SGML | document management system for technical manuals that satisfies the | requirements listed below. If you know of any products that satisfy all | or a majority of these requirements please respond by e-mail to | "sgmldbms@oasys.dt.navy.mil". The desired system must: | 1. maintain structure of an SGML instance. ADOC is based on SGML/Store(TM) storage sub-system which has been designed to store and to manipulate SGML document or collection of documents under a database format. SGML/Store (and so ADOC) accepts arbitrary DTDs, without requiring any specific database schema definition or settings step. A SGML/Store (and so ADOC) database can manage heterogeneous document collections which are instance of multiple DTDs. | 2. maintain multiple versions of the same document without redundancy | of data storage. Yes. The manager of the database is open to define the granularity of the SGML storage management. The manager declares how an SGML document or an SGML fragment will be managed (access rights, versioning). Furthermore, ADOC makes distinction between "evolution" and version. A version is a document or a collection of documents with a number and it could no more be updated. The "evolution" keeps tracking of different steps. ADOC maintains the database consistency and the last version is always used through cross-references. | 3. maintain/manage binary objects, e.g, images, ASCII text. ADOC is able to manage both SGML instances database and illustration database (graphics). It is possible to store/retrieve/access also other binary objects (WORD, EXCEL,...) in the database using external application for editing. ADOC 2.0 will provide versioning on graphics. In ADOC 1.2, the versioning of a graphics is linked with the SGML fragment where the graphics is included. | 4. provide configuration management to include | revision management and workflow management. ADOC 1.2 provides configration management with revision management but no workflow management. ADOC 2.0 will provide user-defined properties and APIs in order to be able to easily integrate a workflow product within ADOC. | 5. support multi-platform development and shall provide a standard | Application Programming Interface. ADOC 1.2 provide an SGML automatic transformation engine in order to execute scripts to modify SGML instances. ADOC 2.0 will provide a set of APIs and a toolkit for developing specific gateway , customizing ADOC environment or adding new functions with plug-in modules. | 6. provide for client/server interfaces across platforms e.g., NT, | Unix/Sun, Windows-based, etc. ADOC 1.2 is client/server but it requires to mount NFS. It is running on SUN OS. ADOC 2.0 will be client/server with RPC on TCP. Alpha release is running on Solaris 2. A windows-based client is planned. | 7. be browser independent. ADOC is editor independent but it requires a gateway to plug-in the editor onto the ADOC workbench. In the release 1.2, ADOC provides a gateway only for Author/Editor from SoftQuad. | 8. export SGML files directly from the database | for use in independent software products. Yes. Export of one document or export of a collection of documents. Furthermore, ADOC is able to mark the delta between the version n-i and n-j of a document during the export process. | 9. support all plug-ins via API, an example would be an enhanced | graphics viewer. ADOC 1.2 provides APIs to plugs external tools for editing or viewing. | 10. efficiently manage 30,000 to 40,000 documents, each averaging 3MB | of text and 17MB of graphics. ADOC effeciently supports very large documents of several tens of Mbytes in size, made up of several million SGML elements. ADOC offers concurrent editing where 10 users can have concurrent access to the same document. ADOC 1.2 (Available) - Documentation under CALS standards: SGML, CGM & CCITT G4 - DTD structuration: arbitrary DTD, definition of SGML elements granularity level - Parsing SGML instance according to its DTD - SGML transformation engine for scipting - Author private working space. ADOC is native concurrent editing system - Integrated native SGML editor - Integrated Graphic CGM editor - Integrated Group 4 graphic editor - Formatted manuals database - OS CALS A & B composer with incremental formatting - Viewer/differential viewer - Printer/differential printer (Postscript) - Electronic publications maker ADOC 2.0 (2Q96) New features - customization of the client application - Client/server architecture (management, import/export, printing, viewing) - Document server based on a RDBMS - Configuration server is based on a RDBMS - SGML Entity management in SGML/Store (and so in ADOC) - Generic declaration at a Class level (DTD, FOSI, filter, access rights, granularity level,...) - Security refinement - APIs For further information, contact me or Isabelle Bornier: phone: [+33] 1 46 40 84 15 FAX: [+33] 1 46 40 84 10 email: ibor@ais.berger-levrault.fr Regards, Stephane Bornerand. -- _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/ Stephane Bornerand phone: [+33] 1 46 40 84 06 _/ _/ AIS S.A. FAX: [+33] 1 46 40 84 10 _/ _/ 35, Rue du Pont email: sbor@ais.berger-levrault.fr _/ _/ 92200, Neuilly, FRANCE WWW: http://ais.berger-levrault.fr _/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ </message> <message id="<494a66$qas@dirac.isa.de>" date="3026201222" seqno="11542"> From: taube@isa.de (Christian Taube) Newsgroups: comp.text.sgml Subject: Re: Mapping RTF hex codes to ISO Latin entities? Date: 24 Nov 1995 12:27:02 +0100 Organization: ISA GmbH, Azenbergstr. 35, D-70174 Stuttgart, FRG Message-ID: <494a66$qas@dirac.isa.de> References: <492g3n$p9d@dirac.isa.de> Keywords: RTF, ISO Latin, mapping, character sets I wrote yeserday: >I'm looking for some sort of table that lists all the RTF hexadecimal >codes -- the ones that look like \\'\<xx> and are replaced by special >characters and "Umlaute" according to The Microsoft Way when you >display the document -- and their corresponding ISO 8859 entities. >(No, I don't regularly work with RTF ... :-) It has been pointed out to me that this is quite obvious from the fact that I didn't indicate which version of RTF and which word processor I am working with ... (God I'm starting to hate RTF, don't you?). OK, so it's the 1.4 spec of RTF that I have here, and I need the map primarily for a German MS Word 6, but English would be nice, too. In the meantime, I'm trying to transcribe those stupid RTF characters manually and find the entities, and I have stopped wondering why the rtftohtml filters that I have seen did mostly a bad job ... :-) Christian. -- Christian Taube -- \<A HREF="http://www.isa.de/~taube">taube@isa.de\</A> Any opinions expressed are mine, and mine only! Thought for today: These are the children of the race "Enterprise." Her 5-year mission - to logically go where no Romulan has gone before. </message> <message id="<199511241214.MAA05230@sun1.cup.cam.ac.uk>" date="3026203554" seqno="11543"> From: Paul Treadaway \<pt100@cup.cam.ac.uk> Newsgroups: comp.text.sgml Subject: Re: SGML -> HTML Date: Fri, 24 Nov 95 12:05:54 GMT Organization: Cambridge University Press Message-ID: <199511241214.MAA05230@sun1.cup.cam.ac.uk> References: <48g54e$8d3@news.INbe.net> <48i0qh$16e@news.Belgium.EU.net> <48jgka$fb5@ixnews2.ix.netcom.com> In-Reply-To: <48jgka$fb5@ixnews2.ix.netcom.com> In article <48jgka$fb5@ixnews2.ix.netcom.com>, Mark Klamerus says... >Jacques Deseyne \<jad@sema.be> wrote: >>"M.T. Carrasco Benitez" \<carrasco@innet.lu> wrote: >>>I am looking for SGML->HTML filter, or a pointer to the FAQ where this >>>kind of answer could be. Mail prefered. >Actually making the required filters in UNIX scripts is very, very >easy. There is no "generic" filter that I know of since every DTD is >different, but if you're looking for a "tool" you might consider >Exoterica, which is excellent. Exoterica is the name of a company. I assume you're referring to one of their products (e.g. Omnimark, their scripting language for SGML conversions). Omnimark has a 4GL style interface which drives a fairly brute force processing model, but with some nice extras such as back- patching for more complicated tasks. Another option is AIS's Balise, which has a more conventional interface (C++ style syntax) but comes with a library of translation functions which operate by directly manipulating an ESIS tree. </message> <message id="<199511241354.OAA25691@cygne.ais.berger-levrault.fr>" date="3026210042" seqno="11544"> Newsgroups: comp.text.sgml Date: Fri, 24 Nov 1995 14:54:02 +0100 From: St�phane Bornnerand \<sbor@Berger-Levrault.fr> Message-ID: <199511241354.OAA25691@cygne.ais.berger-levrault.fr> References: <47tp7u$p2j@dub-news-svc-2.compuserve.com> \<DHtKuM.Mq4@oasis.icl.co.uk> Subject: Re: SGML ----> RTF [Jeff Webster] | Anyone know of a tool for converting SGML to RTF. Thanks in advance. [Alfie Kirkpatrick] | I don't think you'll find an off-the-shelf tool for an arbitrary DTD. | Tools like Omnimark make fairly light work of it though, and the RTF | spec is reasonably well documented (available from the MS Web site, | http://www.microsoft.com). There are two ways to proceed for converting SGML to RTF. The first one is as you proposed is to use a product such Omnimark or Balise to convert a SGML file into a RTF file. This tool works with the format. It is transform a representation into another. The problem here is first that SGML instance is dependent on a DTD and, second, SGML does not include all the data requires by RTF. You should overcome this gap by additional programming or declaration (or both). Note: Balise(TM) is an SGML document processing language developed and shipping by AIS. The second way is to use a Page-composition gateway to transform an SGML instance into an RTF file by composing and formatting SGML instance in order to generate RTF (or other formats). AIS is developing such a product called DynaPage(TM). How does it work ? - take an SGML instance and its DTD - create a stylesheet from the DTD in order to define the value of the attributes of RTF styles - compile SGML instance, DTD, and Stylesheet - generate RTF Advantages - Never require programming - Interactive stylesheet editor (simple text editor could be enough) - Batch conversion - Reliable process - Target RTF styles customisation For further information, contact me or Isabelle Bornier: phone: [+33] 1 46 40 84 15 FAX: [+33] 1 46 40 84 10 email: ibor@ais.berger-levrault.fr Regards, Stephane Bornerand. -- _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/ Stephane Bornerand phone: [+33] 1 46 40 84 06 _/ _/ AIS S.A. FAX: [+33] 1 46 40 84 10 _/ _/ 35, Rue du Pont email: sbor@ais.berger-levrault.fr _/ _/ 92200, Neuilly, FRANCE WWW: http://ais.berger-levrault.fr _/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ </message> <message id="<494l4q$1tq@news.Belgium.EU.net>" date="3026212442" seqno="11545"> From: Jacques Deseyne \<jad@sema.be> Newsgroups: comp.text.sgml Subject: Re: Mapping RTF hex codes to ISO Latin entities? Date: 24 Nov 1995 14:34:02 GMT Organization: SEMA Group Belgium Message-ID: <494l4q$1tq@news.Belgium.EU.net> References: <492g3n$p9d@dirac.isa.de> <494a66$qas@dirac.isa.de> taube@isa.de (Christian Taube) wrote: >I wrote yeserday: >>I'm looking for some sort of table that lists all the RTF hexadecimal >>codes -- the ones that look like \\'\<xx> and are replaced by special >>characters and "Umlaute" according to The Microsoft Way when you >>display the document -- and their corresponding ISO 8859 entities. > >>(No, I don't regularly work with RTF ... :-) > >It has been pointed out to me that this is quite obvious from the fact >that I didn't indicate which version of RTF and which word processor I >am working with ... (God I'm starting to hate RTF, don't you?). OK, >so it's the 1.4 spec of RTF that I have here, and I need the map >primarily for a German MS Word 6, but English would be nice, too. > I'm not sure whether the spec version makes much difference, but the character set defined at the start of a document will. It can be "ansi", "mac", "pc" (IBM PC code page 437) or "pca" (code page 850). With MS-Word 6, you have probaly only the "ansi" character set in your RTF files. The meaning may also differ according to the font which is in use. A single symbol may even be represented by a field like {\\field{\\*\\fldinst SYMBOL 36 \\\\f "ZapfDingbats" \\\\s 12}{\\fldrslt\\f12\\fs24}} There are many other caveats which don't come to mind now. Long live vendor-independent representations! Best regards, ------------------------------------------------------ Jacques Deseyne \<jad@sema.be> Sema Group Belgium - Stallestraat 96 - B-1180 Brussels ------------------------------------------------------ </message> <message id="<DIJyDJ.Gnq@oasis.icl.co.uk>" date="3026213190" seqno="11546"> Newsgroups: comp.text.sgml From: ak@oasis.icl.co.uk (Alfie Kirkpatrick) Subject: Re: Mapping RTF hex codes to ISO Latin entities? Message-ID: \<DIJyDJ.Gnq@oasis.icl.co.uk> Sender: news@oasis.icl.co.uk Reply-To: ak@oasis.icl.co.uk Organization: ICL, Bracknell, UK References: <492g3n$p9d@dirac.isa.de> Date: Fri, 24 Nov 1995 14:46:30 GMT Christian Taube (taube@isa.de) wrote: : Sorry to bug you guys again with stupid RTF stuff, but I need to work : this out ASAP ... :-( Sorry this isn't a definitive answer ... :-) : I'm looking for some sort of table that lists all the RTF hexadecimal : codes -- the ones that look like \\'\<xx> and are replaced by special : characters and "Umlaute" according to The Microsoft Way when you : display the document -- and their corresponding ISO 8859 entities. : Does anyone know if and where such a beast exists? I don't think it will be that simple even if such a table does exist. Remember that some of the ISO entities need a different font in addition to a character code, so it won't be a straight hex code to entity name mapping; you will also need to know the font in effect when the RTF \\'xx occurs. You could construct your own table(s) from the entity sets given an SGML processing tool (bread and butter for Omnimark) and then read them back in when doing the RTF conversion. EBT have done some work on the ISO entities to define text, decimal (&#ddd;) and SDATA equivalents. For the SDATA case, the format is something like: \<!ENTITY mdash SDATA "font=symbol charset=fontspecific code=190"> From this you can construct the right RTF sequence to change font and output the \\'xx code. Incidentally, does anyone know if this format for SDATA entities is a standard/convention or something EBT devised for themselves? Hope this helps. Converting RTF to SGML - rather you than me! Alfie. -- +-Professional Publishing Services---------------+ | Alfie Kirkpatrick ICL | | external: +44 1344 472500 Lovelace Road | | internal: 7263 2500 Bracknell, Berks | | mail: ak@oasis.icl.co.uk RG12 8SN | +------------------------------------------------+ </message> <message id="<DIJyoE.GvB@oasis.icl.co.uk>" date="3026213581" seqno="11547"> Newsgroups: comp.text.sgml From: ak@oasis.icl.co.uk (Alfie Kirkpatrick) Subject: Re: Mapping RTF hex codes to ISO Latin entities? Message-ID: \<DIJyoE.GvB@oasis.icl.co.uk> Sender: news@oasis.icl.co.uk Reply-To: ak@oasis.icl.co.uk Organization: ICL, Bracknell, UK References: <492g3n$p9d@dirac.isa.de> Date: Fri, 24 Nov 1995 14:53:01 GMT Christian Taube (taube@isa.de) wrote: : I'm looking for some sort of table that lists all the RTF hexadecimal : codes -- the ones that look like \\'\<xx> and are replaced by special : characters and "Umlaute" according to The Microsoft Way when you : display the document -- and their corresponding ISO 8859 entities. : Does anyone know if and where such a beast exists? Just ignore my previous reply - I didn't read the subject line properly (about Latin entities). Here is a file mapping the ISOlat1 entity set to decimal characters. \ \ \<!ENTITY aacute "\á"--=small a, acute accent--> \<!ENTITY Aacute "\Á"--=capital A, acute accent--> \<!ENTITY acirc "\â"--=small a, circumflex accent--> \<!ENTITY Acirc "\Â"--=capital A, circumflex accent--> \<!ENTITY agrave "\à"--=small a, grave accent--> \<!ENTITY Agrave "\À"--=capital A, grave accent--> \<!ENTITY aring "\å"--=small a, ring--> \<!ENTITY Aring "\Å"--=capital A, ring--> \<!ENTITY atilde "\ã"--=small a, tilde--> \<!ENTITY Atilde "\Ã"--=capital A, tilde--> \<!ENTITY auml "\ä"--=small a, dieresis or umlaut mark--> \<!ENTITY Auml "\Ä"--=capital A, dieresis or umlaut mark--> \<!ENTITY aelig "\æ"--=small ae diphthong (ligature)--> \<!ENTITY AElig "\Æ"--=capital AE diphthong (ligature)--> \<!ENTITY ccedil "\ç"--=small c, cedilla--> \<!ENTITY Ccedil "\Ç"--=capital C, cedilla--> \<!ENTITY eth "\ð"--=small eth, Icelandic--> \<!ENTITY ETH "\Ð"--=capital Eth, Icelandic--> \<!ENTITY eacute "\é"--=small e, acute accent--> \<!ENTITY Eacute "\É"--=capital E, acute accent--> \<!ENTITY ecirc "\ê"--=small e, circumflex accent--> \<!ENTITY Ecirc "\Ê"--=capital E, circumflex accent--> \<!ENTITY egrave "\è"--=small e, grave accent--> \<!ENTITY Egrave "\È"--=capital E, grave accent--> \<!ENTITY euml "\ë"--=small e, dieresis or umlaut mark--> \<!ENTITY Euml "\Ë"--=capital E, dieresis or umlaut mark--> \<!ENTITY iacute "\í"--=small i, acute accent--> \<!ENTITY Iacute "\Í"--=capital I, acute accent--> \<!ENTITY icirc "\î"--=small i, circumflex accent--> \<!ENTITY Icirc "\Î"--=capital I, circumflex accent--> \<!ENTITY igrave "\ì"--=small i, grave accent--> \<!ENTITY Igrave "\Ì"--=capital I, grave accent--> \<!ENTITY iuml "\ï"--=small i, dieresis or umlaut mark--> \<!ENTITY Iuml "\Ï"--=capital I, dieresis or umlaut mark--> \<!ENTITY ntilde "\ñ"--=small n, tilde--> \<!ENTITY Ntilde "\Ñ"--=capital N, tilde--> \<!ENTITY oacute "\ó"--=small o, acute accent--> \<!ENTITY Oacute "\Ó"--=capital O, acute accent--> \<!ENTITY ocirc "\ô"--=small o, circumflex accent--> \<!ENTITY Ocirc "\Ô"--=capital O, circumflex accent--> \<!ENTITY ograve "\ò"--=small o, grave accent--> \<!ENTITY Ograve "\Ò"--=capital O, grave accent--> \<!ENTITY oslash "\ø"--=small o, slash--> \<!ENTITY Oslash "\Ø"--=capital O, slash--> \<!ENTITY otilde "\õ"--=small o, tilde--> \<!ENTITY Otilde "\Õ"--=capital O, tilde--> \<!ENTITY ouml "\ö"--=small o, dieresis or umlaut mark--> \<!ENTITY Ouml "\Ö"--=capital O, dieresis or umlaut mark--> \<!ENTITY szlig "\ß"--=small sharp s, German (sz ligature)--> \<!ENTITY thorn "\þ"--=small thorn, Icelandic--> \<!ENTITY THORN "\Þ"--=capital THORN, Icelandic--> \<!ENTITY uacute "\ú"--=small u, acute accent--> \<!ENTITY Uacute "\Ú"--=capital U, acute accent--> \<!ENTITY ucirc "\û"--=small u, circumflex accent--> \<!ENTITY Ucirc "\Û"--=capital U, circumflex accent--> \<!ENTITY ugrave "\ù"--=small u, grave accent--> \<!ENTITY Ugrave "\Ù"--=capital U, grave accent--> \<!ENTITY uuml "\ü"--=small u, dieresis or umlaut mark--> \<!ENTITY Uuml "\Ü"--=capital U, dieresis or umlaut mark--> \<!ENTITY yacute "\ý"--=small y, acute accent--> \<!ENTITY Yacute "\Ý"--=capital Y, acute accent--> \<!ENTITY yuml "\ÿ"--=small y, dieresis or umlaut mark--> -- +-Professional Publishing Services---------------+ | Alfie Kirkpatrick ICL | | external: +44 1344 472500 Lovelace Road | | internal: 7263 2500 Bracknell, Berks | | mail: ak@oasis.icl.co.uk RG12 8SN | +------------------------------------------------+ </message> <message id="<199511241627.RAA01609@cygne.ais.berger-levrault.fr>" date="3026219276" seqno="11548"> Newsgroups: comp.text.sgml Date: 24 Nov 1995 17:27:56 +0100 From: St�phane Bornnerand \<sbor@berger-levrault.fr> Message-ID: <199511241627.RAA01609@cygne.ais.berger-levrault.fr> References: \<jfritchDHzuJD.DDs@netcom.com> Subject: Re: SGML and Page Composition [Jeanne Fritch] | I am looking for people using SGML and still have important Page Layout | requirements. The questions I have are: Here, I will describe a Page-composition gateway which is able to generate input to page makeup systems from SGML data. The page layout could then be precisely designed and printed under these systems. Requirements addressed by the Page-compositon gateway include: 1) a style sheet mechanism with "propagating" styles and SGML context sensitivity; 2) a powerful set of data manipulation functions for data filtering, reordering, and generatione; 3) a database implementation, capable of scaling to very large documents without any concern for document size; and 4) sufficient typographic sophistication in composition/layout algorithms (in the page makeup system). | 1. Describe the process from SGML to output. Let me take FrameMaker as page makeup system. Preparation - take an SGML instance and its DTD - create a stylesheet from the DTD in order to define the value of the attributes of RTF styles - compile SGML instance, DTD, and Stylesheet - generate MIF - open MIF with FrameMaker - Enhance style definitions and define page layout under FrameMaker - save FrameMaker stylesheet Production (automating the process) - compilation of SGML instance, its DTD, and its "basic" stylesheet - generatation of MIF - Loading MIF and its "enhanced" stylesheet under FrameMaker and - Printing or doing what you need ! | 2. What tools ae you using? In order to handle the complete Page-Composition layout process from SGML data as detailled in 1, you should use : - DynaText for compiling SGML - DynaPage for generating MIF file - FrameMaker for composing and printing AIS is developing the product DynaPage(TM). DynaPage is a compagnon product of DynaText from EBT. | 3. Is it a batch process or does a lot of work need to be done on | the information at publishing time. Yes. it is a batch process on production (at publishing time !). | 4. What was the process of setting up the environment? Please explain. | 5. Do you deal with A-pages and how? If the question is: Do you deal with A3, A4, ... ? The answer is: Yes. How: under FrameMaker. | 6. Tips you can suggest for someone starting the setup process. If I can answer to 4 than I will be able to suggest tips. A Page-composition gateway based on DynaPage(TM) product could generate input to popular document composition systems, including FrameMaker(TM), Interleaf(TM) and MS-Word(TM). DynaPage is running under Unix. I can send by E-mail the electronic form of the DynaPage documentation on request. Don't hesitate to ask for. For further information, contact me or Isabelle Bornier: phone: [+33] 1 46 40 84 15 FAX: [+33] 1 46 40 84 10 email: ibor@ais.berger-levrault.fr Regards, Stephane Bornerand. -- _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/ Stephane Bornerand phone: [+33] 1 46 40 84 06 _/ _/ AIS S.A. FAX: [+33] 1 46 40 84 10 _/ _/ 35, Rue du Pont email: sbor@ais.berger-levrault.fr _/ _/ 92200, Neuilly, FRANCE WWW: http://ais.berger-levrault.fr _/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ </message> <message id="<495306$3g8@news.mulberry.com>" date="3026228550" seqno="11549"> From: voodoo@mulberry.com (Dennis Stein) Newsgroups: comp.text.sgml Subject: {{{{{{{{{{{ Your Web Page! }}}}}}}}}}} Date: Fri, 24 Nov 1995 19:02:30 GMT Organization: Uniquest On-Line Services Message-ID: <495306$3g8@news.mulberry.com> Send us an email with your rough ideas of what you want on a web page, and we'll e-mail you a sample page of what your business will look like on the WWW. We also do Graphic Design, Data Recovery, Searches, Tourguiding, etc. all by E-mail!!! Send E-mail to: voodoo@mulberry.com HyperNet Information Systems </message> <message id="<49535v$3g8@news.mulberry.com>" date="3026228733" seqno="11550"> From: voodoo@mulberry.com (Dennis Stein) Newsgroups: comp.text.sgml Subject: {{{{{{{{{{{ Your Web Page! }}}}}}}}}}} Date: Fri, 24 Nov 1995 19:05:33 GMT Organization: Uniquest On-Line Services Message-ID: <49535v$3g8@news.mulberry.com> Send us an email with your rough ideas of what you want on a web page, and we'll e-mail you a sample page of what your business will look like on the WWW. We also do Graphic Design, Data Recovery, Searches, Tourguiding, etc. all by E-mail!!! Send E-mail to: voodoo@mulberry.com HyperNet Information Systems </message> <message id="<1995Nov24.181139.24412@sq.com>" date="3026225499" seqno="11551"> Newsgroups: comp.text.sgml From: lee@sq.com (Liam R. E. Quin) Subject: Re: SGML book? Message-ID: <1995Nov24.181139.24412@sq.com> Organization: SoftQuad Inc., Toronto, Canada References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI973L.9CE@mv.mv.com> <1995Nov20.045420.29244@sq.com> <4940fd$anb@usenet1.interramp.com> Date: Fri, 24 Nov 1995 18:11:39 GMT Charles F. Goldfarb \<Goldfarb@InterRamp.com> wrote: >lee@sq.com (Liam R. E. Quin) wrote: > I would appreciate hearing about any errors in README.1ST or other > books in my series as I am not aware of any. E.g. saying that \<P> is a paragraph separator in HTML. The whole HTML section at the back is based on a draft DTD dated from 1993 that was a discussion piece, and is not the correct DTD for HTML. (I realise that RFC1866 wasn't published in time for the book, of course) I didn't find any errors in the SGML part (and didn't claim to) -- but I have only read the HTML part in detail. I do think most of what I have read is well-written, but being two years out of date when you're writing about the Word Wide Web is not a good idea :-( Actually, I should probably praise README.1ST for including anything at all about the Web -- Travis and Waldt included much less in their book, and (I think misleadingly) called HTML an `sgml-like language' [p.55] and also says `HTML is not really SGML' [p.56]. But I don't want to get into details of the content -- my earlier article (as it said) was about first impressions, and I wrote it partly to alert people to more SGML books, and partly to try & get some discussion going. > Please write to Goldfarb@SGMLSource.com. (I've cc'd that mail address) Lee -- Liam Quin, SoftQuad Inc +1 416 239 4801 lee@sq.com \<URL:http://www.sq.com/> HexSweeper NeWS game;OPEN LOOK+XView+mf-fonts FAQs;lq-text unix text retrieval </message> <message id="<49532v$it1@kannews.ca.newbridge.com>" date="3026226719" seqno="11552"> From: Paul Madsen \<paul_madsen@qmail.newbridge.com> Newsgroups: comp.text.sgml Subject: CALS DTD table attributes Date: 24 Nov 1995 18:31:59 GMT Organization: Newbridge Networks Corporation Message-ID: <49532v$it1@kannews.ca.newbridge.com> I'd like to get some info on the CALS table (and its subelements) attributes. Is there an archive defining the purpose of each of the attributes somewhere? Thnaks for any help Paul Madsen Online Documentation Newbridge Networks Corp </message> <message id="<497b1k$21j@news.xs4all.nl>" date="3026300404" seqno="11553"> From: jang@xs4all.nl (Jan Grootenhuis) Newsgroups: comp.text.sgml Subject: Re: HyTime and OpenDoc and OLE Date: 25 Nov 1995 15:00:04 GMT Organization: XS4ALL, networking for the masses Message-ID: <497b1k$21j@news.xs4all.nl> References: \<DIG9zL.uy@sunsrvr6.cci.com> Keywords: HyTime,OpenDoc,OLE rs@cci.com (Rick Silterra) writes: >I have been asked the question: How do HyTime, and OpenDoc relate? I propose you read Liora Alschuler's beautiful book: ABCD... SGML, a user's guide to structured information, International Thomson Computer Press, ISBN 1-850-32197-3 It compares SGML-WYSIWYG-DTP-Databases-GML-ODA-Opendoc-OLE-RTF-PostScript- PDF-HTML-HyTime on p. 31-38. Enjoy! -- Jan Grootenhuis Kralenbeek 1873 1104 KJ AMSTERDAM The Netherlands Tel/fax (+31)20-699.89.66 Internet jang@xs4all.nl </message> <message id="<497bi1$29p@news.xs4all.nl>" date="3026300928" seqno="11554"> From: jang@xs4all.nl (Jan Grootenhuis) Newsgroups: comp.text.sgml Subject: Re: Mapping RTF hex codes to ISO Latin entities? Date: 25 Nov 1995 15:08:48 GMT Organization: XS4ALL, networking for the masses Message-ID: <497bi1$29p@news.xs4all.nl> References: <492g3n$p9d@dirac.isa.de> <494a66$qas@dirac.isa.de> Keywords: RTF, ISO Latin, mapping, character sets taube@isa.de (Christian Taube) writes: >so it's the 1.4 spec of RTF that I have here, and I need the map >primarily for a German MS Word 6, but English would be nice, too. Did you check your local MicroSoft BBS? They normally that sort of stuff. Tschuess, -- Jan Grootenhuis Kralenbeek 1873 1104 KJ AMSTERDAM The Netherlands Tel/fax (+31)20-699.89.66 Internet jang@xs4all.nl </message> <message id="<497nt4$eem@toads.pgh.pa.us>" date="3026313572" seqno="11555"> From: tore@lis.pitt.edu (Tore Joergensen) Newsgroups: kaikow@standards.com,comp.text.sgml Subject: Re: SGML book? Date: 25 Nov 1995 18:39:32 GMT Organization: University of Pittsburgh Message-ID: <497nt4$eem@toads.pgh.pa.us> References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI625t.F6M@curia.ucc.ie> \<DI973L.9CE@mv.mv.com> kaikow@standards.com wrote: : Is the Alschuler book better than the Practical SGML book by Van : Herwijejnen? If so, how? : In article \<DI625t.F6M@curia.ucc.ie>, Peter Flynn \<pflynn@www.ucc.ie> wrote: : >Dixit pyachnes@ucs.indiana.edu (paul alan yachnes): : >> Can someone recommend the best book for learning SGML? : >The best one I can recommend is "ABCD... SGML: A User's Guide : >to Structured Information" by Liora Alschuler (ISBN: 1-850-32197-3) "ABCD...SGML" is not a book for learning SGML. The explanation of the SGML syntax is very brief. This book tries to explain the big picture, rather than to learn you how to write SGML. "Practical SGML" by Eric van Herwijnen is a book for learning SGML, but in my opinion not a very good one. It is to much like a reference manual... 'This is everything you can do with this,... now let's see on the next structure in the language'. I don't think the explanation of everything is as good as it should be either. "The SGML Handbook" by Charles F. Goldfarb is a very complete book. I don't know how it is to learn from, but it is very good as a reference manual. My impression (after reading a few sections :-) is that it probably is easier to learn from, than "Practical SGML". "The Complete SGML"-CD from Exoterica is the ISO-standard with some explanations, and many example documents. In my opinion the interface sucks. I would not try to learn from this. There are a few other books, and SoftQuad have a cheap CD, but I haven't seen any of them and can of course not comment on them. -- +-------------------------+-------------------------------------------+ | Tore B. Joergensen | e-mail : tore@lis.pitt.edu | | Centre Court Villa | web : http://www.pitt.edu/~tojst1 | | 5535 Centre Avenue # 6 | | | Pgh, PA 15232, USA | Norwegian MSIS-student at Univ. of Pgh. | +-------------------------+-------------------------------------------+ </message> <message id="<4981lf$qcj@agate.berkeley.edu>" date="3026323567" seqno="11556"> From: jmcd@info.sims.berkeley.edu (Jerry McDonough) Newsgroups: comp.text.sgml Subject: Content Model Blues Date: 25 Nov 1995 21:26:07 GMT Organization: School of Information Management and Systems, UC Berkeley Message-ID: <4981lf$qcj@agate.berkeley.edu> I've got what I thought was a fairly simple content model problem, that's turning uglier every time I look at it. I suspect I'm missing something fundamental here. If anyone can shed some enlightment on me, I'd appreciate it. The Problem: I've got an element (let's call it HAIRY) consisting of three subelements, a, b, & c. Element a is optional, and may only occur once (e.g., a?). Element b is mandatory and may only occur once (e.g, b), and element c is optional and may occur as many times as it wants (e.g, c*). The subelements may occur in any order they want, so the following: \<HAIRY> \<c>data\</c>\<a>data\</a>\<c>data\</c>\<c>data\</c>\<b>data\</b>\<c>data\</c> \</HAIRY> should be perfectly valid. Getting the constraints I want on the repeatability of the subelements is pretty straightforward if I impose a specific order on them, like \<!ELEMENT HAIRY - - (a?, b, c*)> but how do I get the constraints on repeatability *without* restricting the order in any fashion? Thankful for any help I can get, Jerry McDonough -- jmcd@info.SIMS.Berkeley.EDU | (......) UCB School of Information Management & Systems | \\ * * / | \\ <> / "Don't worry. I know what I'm doing!" | \\ -- / SGNORMPF!!! -- From the Famous Last Words file | |||| </message> <message id="<NEWTNews.2051.817324880.bhoffman@cerfnet.cerfnet>" date="3026309628" seqno="11591"> From: bhoffman@cerfnet.com Newsgroups: comp.text.sgml Subject: SGML Contracts/Perm. Jobs Date: Sat, 25 Nov 95 10:33:48 PDT Organization: CERFnet Message-ID: \<NEWTNews.2051.817324880.bhoffman@cerfnet.cerfnet> Greetings - I have several career opportunities for SGML experts. My client is seeking several talented, skilled individuals to join a very dynamic, SGML-based publishing team. On the contract side, we are interested in individuals who can provide expertise in use of Electronic Book Technologies (EBT) tools. On the permanent side, we are looking for Content Management and Production Management people. For more information and consideration, please contact me: netWorkers, inc. 415 254-2350 - Bob Susskind </message> <message id="<49a4gh$idn@usenet.ucs.indiana.edu>" date="3026392017" seqno="11558"> From: pyachnes@ucs.indiana.edu (paul alan yachnes) Newsgroups: kaikow@standards.com,comp.text.sgml Subject: Re: SGML book? Followup-To: kaikow@standards.com,comp.text.sgml Date: 26 Nov 1995 16:26:57 GMT Organization: Indiana University Message-ID: <49a4gh$idn@usenet.ucs.indiana.edu> References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI625t.F6M@curia.ucc.ie> \<DI973L.9CE@mv.mv.com> <497nt4$eem@toads.pgh.pa.us> I began this thread way back when, before I had done any research on the subject. Subsequently, I discovered theat QUE will be publishing "Using SGML" at the end of January (listed in Books in Print as Oct. '95). Since I was looking for a "how to" book for the purpose of self-instruction, I look forward to seeing this. The books in this series are consistently pretty good. I thought others might be interested. Paul Yachnes </message> <message id="<BXCFjXC.liora@delphi.com>" date="3026393722" seqno="11559"> From: liora@delphi.com Newsgroups: comp.text.sgml Subject: Re: SGML book? Date: Sun, 26 Nov 95 11:55:22 -0500 Organization: Delphi (info@delphi.com email, 800-695-4005 voice) Message-ID: \<BXCFjXC.liora@delphi.com> References: <480i5v$hsp@usenet.ucs.indiana.edu> \<DI973L.9CE@mv.mv.com> <1995Nov20.045420.29244@sq.com> <1995Nov20.214942.10037@sq.com> \<DIJ8sB.1Lq@mv.mv.com> \<kaikow@standards.com> writes: >I agree, but a few bookstores stock SGML books. Others will get copies of >most books for one to look at with no obligation to buy, however, most >will only do that if the book is available thru their supplier. Alas, >Practical SGML is not so available. Excerpts from ABCD... SGML are available at: http://www.thomson.com/itcp/liora/LA_HOME.HTML These are in HTML. There will soon be excerpts (the entire Resource Guide) available in SGML as a WebExtra at www.thomson.com/itcp/itcp.html Please sample. Liora </message> <message id="<JBW.95Nov26144724@bigbird.bu.edu>" date="3026404044" seqno="11560"> From: jbw@bigbird.bu.edu (Joe Wells) Newsgroups: comp.text.sgml Subject: Re: Content Model Blues Date: 26 Nov 95 19:47:24 GMT Organization: Boston University Computer Science Department Message-ID: \<JBW.95Nov26144724@bigbird.bu.edu> References: <4981lf$qcj@agate.berkeley.edu> In-reply-to: jmcd@info.sims.berkeley.edu's message of 25 Nov 1995 21:26:07 GMT In article <4981lf$qcj@agate.berkeley.edu> jmcd@info.sims.berkeley.edu (Jerry McDonough) writes: I've got an element (let's call it HAIRY) consisting of three subelements, a, b, & c. Element a is optional, and may only occur once (e.g., a?). Element b is mandatory and may only occur once (e.g, b), and element c is optional and may occur as many times as it wants (e.g, c*). The subelements may occur in any order they want, so the following: \<HAIRY> \<c>data\</c>\<a>data\</a>\<c>data\</c>\<c>data\</c>\<b>data\</b>\<c>data\</c> \</HAIRY> should be perfectly valid. Hmm, nice puzzle. The elements can occur in these numbers in any order: a: 0 or 1 b: 1 c: 0, 1, 2, ... If elements a, b, and c have declared content CDATA, RCDATA, or EMPTY, then we can just do this and be done: \<!ELEMENT HAIRY (b & a?) +(c)> However, using an inclusion exception to allow element c to appear anywhere can screw you up if one of the elements a, b, or c allow included elements to appear. You could get around this by having exclusion exceptions for a, b, and c, but this can be a problem as well. The more general solution goes as follows. The way the elements can appear is described by this state machine: _c_ _c_ / \\ / \\ | | | | \\ V V / --a0,b0 ---b---> a0,b1-- | | | | a a | | | | V V --a1,b0 ---b---> a1,b1-- / ^ ^ \\ | | | | \\ / \\ / -c- -c- I've given the states names that indicate how many a or b elements have been seen so far. The state "a0,b0" is the start state and the states with "b1" in their name are accept states. That corresponds to this regular expression: c*((ac*bc*)|(bc*(ac*)?) I'll need to play with this a bit to put it in SGML syntax. This should be a valid content model for what you want: (c*, ((a, c*, b, c*) | (b, c*, (a, c*)?))) By the way, for your amusement, the proposed DTD for HTML 3.0 faced the same issue and handled it wrong. As a result, the following is not valid HTML 3.0: \<HTML> \<HEAD> \<META NAME="foo" CONTENT="bar"> \<TITLE> Here is the title. \

', it is an >abomination. OK. I see that my intuition was basically on target. I sense that this 'abomination' is a coming train. >The main problem is that the values that the 'CLASS' attribute can take >are not part of the HTML standard at all, therefore they are arbitrary, >and therefore they should be only used to specify rendering >possibilities, not structure (for structure _must_ be >document-independent). A very slippery slope. This is such a fine distinction that I think we're going to have a lot of trouble. Just look at the level of knowledge (WRT HTML) of the average reader of this group (ciwah). People are going to start using CLASS for things that really should be semantic (e.g., Appendix). For more evidence, see the discussion in a different thread about character markup vs heading tags... >Stan> Yes - with the caveat that they *also* provide a way to specify a >Stan> reasonable default behavior. >Let me add: a reasonable _HTML_ default behaviour. Because one of the >important consideration in the HTML3 design has been backwards >compatibility with HTML2 implementations. We could (?) try the same trick that Netscape uses for frames. You have a head section that contains the content for browsers capable of handling mappings from, say, \ to \

. Generally, that trick is a coin that we could probably spend once to make a big leap in capability in HTML and still preserve reasonable behavior in existing browsers. I'm not sure about this... >This has obvious benefits in terms of continuity, as compared to "OK, >HTML was small and fun, good to start with, but now there is no problem >in switching to full SGML-capable browsers". >The obvious snag is that one is then encouraged to do things like: >Stan> That is using, or instance, \

says that this >Stan> is an appendix title, and that it can be safely considered to be >Stan> equivalent to a second level header without doing damage to the >Stan> document content. >Which, as argued above, strictly speaking, are horrors. Agreed. But this may just be a fact of life as HTML can only evolve -- I doubt revolution (say full SGML) is likely to work broadly... >Stan> Saying \ says this is an appendix, and who the heck knows >Stan> how to treat it if it cannot be treated as an appendix title. >Which would be the right thing to do, _if_ \ were part of >HTML's content model. Right. >Stan> This difference - the mechanism for attaching default meaning - is >Stan> the critical one that makes it a useful and safe means of >Stan> extending HTML. >Again, the point is that _really_ 'CLASS' should not imply any new >"meaning", just customization of an existing thingie. >Stan> The lack of this is why \ was such a bad idea. What could >Stan> this man to a browser that doesn't have this extension? On the >Stan> other hand using \
at least says to treat the text >Stan> as a paragraph even if centering is not supported. >This works because 'ALIGN' has no semantic connotation. '\
' is a >paragraph however it is aligned. But consider instead > \
> \
> \
See above. >If this usage spreads (which in some measure is inevitable), then HTML >could be replaced by HTML(-1), with a single > \ >construct. :-) I used to point out to people that they could easily collapse all their object models down to: Object-related-to-Object. Voila! You've covered everything. It's useless, of course... --------------------------------------------- David Ornstein BrowserCaps: http://objarts.com/bc Outbreak: http://ichiban.objarts.com/outbreak-unreg Personal Info: http://objarts.com/davido From: Bruce Hunter \ Newsgroups: comp.text.sgml Subject: PRODUCT ANNOUNCEMENT - SGML processing language with free tools Date: 27 Nov 1995 02:18:51 GMT Organization: Direct Connection (Call 0181 297 2200 for info) Message-ID: <49b76b$a4v@newsgate.dircon.co.uk> SGMLC - AN SGML DOCUMENT PROCESSING LANGUAGE, WITH FREE COMPILER AND RUNTIME PROCESSOR SGMLC is an SGML document processing language which is based extensively upon the C programming language. It is the language at the heart of a suite of products, developed by SGML Systems Engineering, which enable the creation of powerful and flexible SGML document processing systems. These may range from simple transformation tools, for example to convert an SGML document into another form, to complex IETMS's and batch page composition systems. Programs, or "rules files", are written in SGMLC using any text editor. They are then compiled into object form by the SGMLC compiler. There are several versions of the SGMLC compiler, ranging from a basic free one to the complete SGMLC Application Development Environment, which incorporates an interactive debugger, C language DLL interface, and system customisation facilities. The resulting object file may then be freely distibuted to anyone who has the free SGMLC runtime processor. An SGML document is processed by simply opening it in the runtime processor and applying the relevant rules file. The free version of the compiler, together with the free runtime processor, will be all that is needed to create stand-alone SGML document transformation applications. It is hoped that the free availability of these tools will remove the need for hybrid, multi-pass applications such as, for example, using perl to process the output of the sgmls parser. For commercial organisations who require a supported product, but who do not need the full Application Development Environment, there will be an enhanced, supported version of the free compiler. The additional facilities will include enhanced regular expression patterm matching together with the ability to attach the inbuilt SGML parser to any stream rather than just the input stream. This will enable, for example, complex up-translations where the parser is used to parse the generated output. The output parser data structures are then available for making conditional decisions whilst processing the input. A range of SGMLC add-on libraries will also be available. These include the SGMLC-View library, which provides the functionality for creating complex SGML browser applications, including support for complex tables and equations. The SGMLC-CALSTAB library provides pre-written viewer support for processing CALS tables. It conforms to the recommendations of the SGML Open White Paper on Tables Interoperability. The range of products, which will be made generally available after the SGML 95 conference for the MS Windows operating systems, may be seen at the SGML Systems Engineering booth in the exhibition area at SGML 95. A limited number of diskettes, containing the free products and examples of some sophisticated viewer applications, will be available there. After SGML 95 the free SGMLC compiler and runtime processor will be made publicly available via the usual SGML archive ftp sites. regards, Bruce Hunter SGML Systems Engineering bruce@sgml.dircon.co.uk From: jbw@bigbird.bu.edu (Joe Wells) Newsgroups: comp.text.sgml Subject: Re: Content Model Blues Date: 27 Nov 95 02:32:40 GMT Organization: Boston University Computer Science Department Message-ID: \ References: <4981lf$qcj@agate.berkeley.edu> \ In-reply-to: jbw@bigbird.bu.edu's message of 26 Nov 95 19:47:24 GMT In article \ jbw@bigbird.bu.edu (Joe Wells) writes: The HTML 2.0 DTD botched the handling of this issue in a different way. The following is valid by the HTML 2.0 DTD: \ \ \ Here is \<META NAME="foo" CONTENT="bar"> the title. \ \ \ \
Here is the body. \
\ \ I would like to correct myself. My statement was based on the August 4, 1995 version of the HTML 2.0 DTD. In that version, the TITLE element was declared as follows: \ It occurred to me that this problem might have been fixed in the final version in RFC 1866. I just checked and indeed the problem had been fixed. The declaration of the TITLE element now reads like this: \ What is a bit confusing to me is why a declared content of RCDATA was not used instead of a content model with an exclusion exception. Is there some disadvantage to a declared content of RCDATA? (Yes, I know about the problems with CDATA, I am asking about RCDATA.) -- Joe Wells \ From: kshapiro@cyberus.ca (Kivi Shapiro) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: hdp:// Date: 26 Nov 1995 22:55:37 -0500 Organization: Graduate School of Library and Information Science, UWO Message-ID: <49bcrp$qq4@cyberhome.cyberus.ca> References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48vf7f$nuc@dove.nist.gov> \ Reply-To: kshapiro@pobox.com (Kivi Shapiro) Summary: The time is ripe. In article \, Piercarlo Grandi \ wrote: > > \ > >(hdp of course is the "hardware download protocol") seems rather hard to >do. But it fills a need! Can you imagine the usefulness of this protocol? I'll help write the RFC, if you like. The Web will never be the same. - Kivi, who thinks the rec.arts.startrek.tech folks might be able to help -- kshapiro@julian.uwo.ca or kivi@pobox.com (Kivi Shapiro) All original material in this message is in the public domain. Newsgroups: comp.text.sgml,alt.fan.bill-gates,alt.fan.mozilla,alt.fan.newt-gingrich,alt.politics.usa.republican From: poulosio@netcom.com (Mad Max) Subject: Re: Slight flame on SGML Message-ID: \ Organization: Troll Busters Ltd. Melbourne VIC. References: <9511171331.AA12385@fly.HiWAAY.net> <48qmfi$14r@bell.maths.tcd.ie> Date: Mon, 27 Nov 1995 04:32:47 GMT Sender: poulosio@netcom16.netcom.com Timothy Murphy (tim@maths.tcd.ie) wrote: : * 25% of World Wide Web users have incomes of more than $80,000. If that is anywhere true, they all hang out in alt.politics.* newsgroups, which could be rmgrouped and replaced with alt.flame.the-poor.get.off.your.lazy.arses.and.get.a.bloody.job . Of course it would be named in American-made English instead, which is the M$-DOG of human language, of course. -- "When I collect two solar masses of AOL Disks, I will use them to detonate the Sun" Web Page URL: ftp://ftp.netcom.com/pub/po/poulosio/poulosio.html Mad Max, LT, TRES Corps, Kappa. Shut the immigration inlet valve NOW! Linux: You get what you pay for. From: pepper@falch.no (Steve Pepper) Newsgroups: comp.text.sgml Subject: Re: Numeric Tags and Parsing Errors Date: Mon, 27 Nov 1995 12:56:11 LOCAL Organization: Falch Infotek AS, Norway Message-ID: \ References: \ \ In article \ sjd@ebt.com (Steven J. DeRose) writes: |In article \, |tim@michelob.wustl.edu (Tim Rand) wrote: | |> We have recently run across a problem were codes that begin with numerics |> are passed through parsing, yet are invalid. |>... |> Example: <3a> passed through and it isn't defined in the DTD. | | [...] | |In the SGML declaration you thoughtfully supplied the usual naming rules |are in effect, so "3" is a name character but not a name start character |(that is, names like GIs can't start with it). | |So in "<3a>" the "<" is not a STAGO in context, and is just data. Assuming |#PCDATA is allowed at the points in question, that won't produce an error. | | [...] | |If you want <3a> as a tag, you'd need to add digits to the list of name |start characters, and then declare the element type as usual. Unfortunately this is one of the few areas in which the SGML syntax isn't configurable. [189] naming rules states that "A _character_ assigned to LCNMCHAR, UCNMCHAR, LCNMSTRT, or UCNMSTRT cannot be an LC Letter, UC Letter, Digit, RE, RS, SPACE, or SEPCHAR." In other words: You can't just add digits to the list of name start characters. There is no way of turning <3a> into a valid tag. Steve Steve Pepper, SGML Architect pepper@falch.no Falch Infotek a.s, Postboks 130 Kalbakken, N-0902 Oslo, Norway http://www.falch.no/ tel://+47 2290 2733 fax://+47 2290 2599 "The Whirlwind Guide": http://www.falch.no/~pepper/sgmltool/ From: tonyk@cybercom.net (Antoun Kanawati) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: hdp:// Date: Mon, 27 Nov 1995 07:59:02 -0500 Organization: at home Message-ID: \ References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48vf7f$nuc@dove.nist.gov> \ <49bcrp$qq4@cyberhome.cyberus.ca> In article <49bcrp$qq4@cyberhome.cyberus.ca>, kshapiro@pobox.com (Kivi Shapiro) wrote: > > \ > > > >(hdp of course is the "hardware download protocol") seems rather hard to > >do. > > But it fills a need! Can you imagine the usefulness of this protocol? > I'll help write the RFC, if you like. The Web will never be the same. Sounds like trouble. PC's have this tendency to behave rather strangely if you don't exactly the right mixture of versions (OS, Windows, DLL's, etc...) Things are bad enough with some installers overwriting DLLs indiscriminately. Getting what appears to be device drivers loaded thru WWW sounds like a major security hole, and a nice way to crash many delicately configured platforms. -- Antoun (Tony) Kanawati tonyk@cybercom.net http://www.cybercom.net/~tonyk/ From: liora@delphi.com Newsgroups: comp.text.sgml Subject: Re: SGML book? Date: Mon, 27 Nov 95 08:42:53 -0500 Organization: Delphi (info@delphi.com email, 800-695-4005 voice) Message-ID: \ References: <480i5v$hsp@usenet.ucs.indiana.edu> \ <1995Nov20.045420.29244@sq.com> <1995Nov20.214942.10037@sq.com> \ \ tore@lis.pitt.edu (Tore Joergensen) wrote: >kaikow@standards.com wrote: > >: Is the Alschuler book better than the Practical SGML book by Van >: Herwijejnen? If so, how? > >: In article \, Peter Flynn \ wrote: >: >Dixit pyachnes@ucs.indiana.edu (paul alan yachnes): >: >> Can someone recommend the best book for learning SGML? >: >The best one I can recommend is "ABCD... SGML: A User's Guide >: >to Structured Information" by Liora Alschuler (ISBN: 1- 850-32197-3) > >"ABCD...SGML" is not a book for learning SGML. The explanation of the >SGML syntax is very brief. This book tries to explain the big picture, >rather than to learn you how to write SGML. > > "Practical SGML" by Eric van Herwijnen is a book for learning SGML,... Quite true, if by "learning SGML" Mr. Joergensen means learning SGML syntax. There is little overlap in function between my book and van Herwijnen's or the other books on SGML syntax mentionned in the same post. An earlier post also distinquished between the two types of book saying that those who are learning to write SGML syntax needn't explore the choices for the "big picture" type of book. Also a statement that I must agree with -- assuming that these syntax writers have the big picture already in mind, garnered through other sources, before they sit down to write code. I have one other clarification on the position of my book in terms of the others mentionned in this thread and that is that unlike at least one of the two recent books that attempt to put SGML into a broad perspective, my book does not deal with the Web as if it should just go away. Two of the book's case studies are of Web sites, one in HTML, one in SGML, and the book comes bundled with Web-ready Panorama files and a custom build of SoftQuad's Panorama Web viewer for SGML... and much, much more, as they say. Finally, I want to correct the Web addressses I posted earlier. At the current time, the HTML excerpts from ABCD... are available on: http://www.thomson.com/itcp/liora/LA_HOME.HTM NOTE: No final "L" (hope to have this fixed soon) The SGML excerpts will be posted to http://www.thomson.com/itcp.html under WebExtras. NOTE: No intermediate /itcp/ directory. Sorry if this caused any inconvenience. Regards, Liora Liora Alschuler The Word Electric POB 177 East Thetford, VT 05043 802/785-2623 liora@delphi.com Newsgroups: comp.text.sgml From: jbottoms@world.std.com (John W Bottoms) Subject: Content Markup & Blackbird Message-ID: \ Organization: The World Public Access UNIX, Brookline, MA Date: Mon, 27 Nov 1995 16:37:36 GMT There are too many areas to keep up with. I was reading in "Wired" that Blackbird will support content markup with Microsoft Markup Language (MSML?) which is based on SGML. Has any information been published on how MS will determine the content structure for the various communities of interest? Also, BillG has a book out called "The Road Ahead" which may may reference to these future trends. -jb Newsgroups: comp.text.sgml From: mstewart@sq.com (Mike Stewart) Subject: SGML, The New World Tour CD-ROM Message-ID: <1995Nov27.205343.12374@sq.com> Sender: news@sq.com (News Administrator) Organization: SoftQuad Inc, Date: Mon, 27 Nov 1995 20:53:43 GMT About a year ago SoftQuad Inc. released a CD-ROM (SGML, The World Tour) full of stuff about SGML. The purpose of the CD was to be a reference for and to promote the use of SGML. We're doing it again and this time we're looking for your input. Our updated version, which will be titled, "SGML, The New World Tour" will be available for free. We are, therefore, limited to royalty-free content. We are targetting the content towards three levels of experience: those who have never encountered SGML, those who have some familiarity with it, and those who are already involved with SGML but wish to expand their knowledge. We are organizing materials under the following headings: "Starting Out in SGML"; "Tutorials and Demos"; "Who's Using SGML (The Web, Industry, Experts and Accessibility)"; "Resources and References" (including history and F.A.Q.s); and "SoftQuad, Inc.". Contributors (if we use your stuff) are entitled to free copies of the CD when it comes out. We are looking for marked up contributions (we'll be using SoftQuad's Panorama to browse the material). We're also open to suggestions and pointers to neat stuff (like SGML freeware and cool projects). Email us at worldtour@sq.com and we can work out the details of how to submit and when. Oh yeah, by the way we're looking for final submissions by December the 15th. I know, I know, it's close. If this were an ideal world we'd have got on it sooner (but then again publishing dates wouldn't move up either). P.S. The old World Tour CD is available, contact us for details. ==================== Mike Stewart SGML, The New World Tour SoftQuad Inc. (416) 239-4801 xt 203 mstewart@sq.com From: lazio@spacenet.tn.cornell.edu (T. Joseph W. Lazio) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: hdp:// Date: 27 Nov 1995 21:18:01 GMT Organization: Department of Astronomy, Cornell University Sender: jl26@cornell.edu (Verified) Message-ID: \ References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48vf7f$nuc@dove.nist.gov> \ <49bcrp$qq4@cyberhome.cyberus.ca> \ Reply-To: lazio@spacenet.tn.cornell.edu In-reply-to: tonyk@cybercom.net's message of Mon, 27 Nov 1995 07:59:02 -0500 >>>>> "AK" == Antoun Kanawati \ writes: In article <49bcrp$qq4@cyberhome.cyberus.ca>, kshapiro@pobox.com (Kivi Shapiro) wrote: > > \ > > > >(hdp of course is the "hardware download protocol") seems rather hard to > >do. > > But it fills a need! Can you imagine the usefulness of this protocol? > I'll help write the RFC, if you like. The Web will never be the same. AK> Sounds like trouble. PC's have this tendency to behave rather AK> strangely if you don't exactly the right mixture of versions (OS, AK> Windows, DLL's, etc...) Things are bad enough with some AK> installers overwriting DLLs indiscriminately. Getting what AK> appears to be device drivers loaded thru WWW sounds like a major AK> security hole, and a nice way to crash many delicately configured AK> platforms. Ah, but I'm sure that with 50000 bit secure encrypting combined with the appropriate salts, transfer of the drivers can be accomplished. Seems to me the more difficult issue will be whether the bandwidth is sufficient to handle the the necessary encoding to transfer the actual hardware. -- Cornell knows I exist?!? | e-mail: lazio@spacenet.tn.cornell.edu Sgt. Lazio, HTML police | http://astrosun.tn.cornell.edu/students/lazio/ STOP RAPE | ICBM: 42:20:08 N 76:28:48 W 305 m alt. sci.astro FAQ at http://astrosun.tn.cornell.edu/students/lazio/sci.astro.html Newsgroups: comp.text.sgml Message-ID: <9511280033.AA0192@notes.microstar.com> From: Ken Holman \ Date: 27 Nov 1995 16:31:42 GMT References: \ Subject: Re: What SGML products are using OLE? [Lori Bolen] > I was wondering which SGML products are using OLE? Are there any? The NEAR & FAR Author for Microsoft Word product (now shipping V1.0) will allow external entities (referenced either as general entities in content or the entity attribute of an element) to be linked (not embedded) OLE-servicable objects. \ \ \ \ \ \ ]> \ \ This is a general entity reference to an OLE object: \&sheet; \ This is an element reference to an OLE object:\ \ The object is displayed within the Word environment; double-clicking on the entity construct, or selecting the menu item Edit/LinkedObject on the element construct, will invoke the appropriate OLE server. Configuration files allow any SYSTEM identifier to be mapped to OLE servers through WIN.INI field values, and the catalog file maps PUBLIC identifiers to SYSTEM identifiers. ........... Ken G. Ken Holman Tel: +1 613 596-CADE(2233) /\\ /\\ Chief Technology Officer Fax: +1 613 596-5934 \\/ \\/ Computer Microstar Software Ltd. WATS: 1 800 267-9975 /\\ /\\ Aided 3775 Richmond Road Mail: gkholman@microstar.com \\/ \\/ Document Nepean Ontario Info: cade@microstar.com /\\ /\\ Engineering CANADA K2H-5B7 Web: http://www.microstar.com \\/ \\/ * Please Note New Address * From: bmoore@neonramp.com (Bill Moore) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: hdp:// Date: Mon, 27 Nov 1995 17:19:30 +0200 Organization: Digital Revolution Message-ID: \ References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48vf7f$nuc@dove.nist.gov> \ <49bcrp$qq4@cyberhome.cyberus.ca> In article <49bcrp$qq4@cyberhome.cyberus.ca>, kshapiro@pobox.com (Kivi Shapiro) wrote: > In article \, > Piercarlo Grandi \ wrote: > > > > \ > > > >(hdp of course is the "hardware download protocol") seems rather hard to > >do. > > But it fills a need! Can you imagine the usefulness of this protocol? > I'll help write the RFC, if you like. The Web will never be the same. > > - Kivi, who thinks the rec.arts.startrek.tech folks might be able to help > -- > kshapiro@julian.uwo.ca or kivi@pobox.com (Kivi Shapiro) > All original material in this message is in the public domain. Sorry folks, but I missed something. hdp hardware download protocol ? Anyone care to explain? ( I gotta read this group more often!) Bill -- Bill Moore Internet Publishing & Design Editor, Sr. Partner Digital Revolution Publishers of MacWare Revue 399 N 117th,Suite 301 http://www.piazza.com Omaha, NE 68154 MacO/S Freeware/Shareware Reviews \MacWare Revue\ From: jsbell@acs.ucalgary.ca (Joshua Sean Bell) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: hdp:// Date: 28 Nov 1995 02:28:18 GMT Organization: University Computing Services (The University of Calgary) Message-ID: <49ds42$ri8@ds2.acs.ucalgary.ca> References: <47d5ro$o3j@hamilton.maths.tcd.ie> <49bcrp$qq4@cyberhome.cyberus.ca> \ \ T. Joseph W. Lazio \ wrote: >>>>>> "AK" == Antoun Kanawati \ writes: > >In article <49bcrp$qq4@cyberhome.cyberus.ca>, kshapiro@pobox.com (Kivi >Shapiro) wrote: >> > \ >> > >> >(hdp of course is the "hardware download protocol") seems rather hard to >> >do. >> >> But it fills a need! Can you imagine the usefulness of this protocol? >> I'll help write the RFC, if you like. The Web will never be the same. > >AK> Sounds like trouble. > > Ah, but I'm sure that with 50000 bit secure encrypting combined with >the appropriate salts, transfer of the drivers can be accomplished. >Seems to me the more difficult issue will be whether the bandwidth is >sufficient to handle the the necessary encoding to transfer the actual >hardware. We'll have to state in the RFC that browsers are required to fetch hdp: method URLs, and ignore any attempts at content negotiation. I mean, imagine the poor user who will have to sit through the microseconds necessary for this: Accept: Macintosh/PowerPC-NuBus; q=0.9, Macintosh/PowerPC-ISA; q=0.1 Content-type: Wintel/PentiumPro-ISA Accept: Macintosh/PowerPC-NuBus; q=0.9, Macintosh/PowerPC-ISA; q=0.1 Content-type: I SAID Wintel/PentiumPro-ISA Accept: Macintosh/PowerPC-NuBus; q=0.9, Macintosh/PowerPC-ISA; q=0.1 Content-type: Fuck you, Macintosh sucks! Accept: Macintosh/PowerPC-NuBus; q=0.9, Macintosh/PowerPC-ISA; q=0.1 Content-override: munition/x-hydrogen-bomb; v=10.1MT Obviously, the Macintosh should recognize its obselecense and gracefully commit sepuku (sp?) before the Mightyness That Is Bill. Joshua -- _.----~~~----._ | MIME: jsbell@acs.ucalgary.ca ,--------.-.,-'-------------------` | WWW: http://www.ucalgary.ca/~jsbell/ `--------"-'-.,---`~-----~' '---'-._____/ --===( Note: the above is humor. Hu-mor. Look it up.) From: K. McLennan <71431.2255@CompuServe.COM> Newsgroups: comp.text.sgml Subject: ISO12038 Math viewer advice needed Date: 28 Nov 1995 00:20:56 GMT Organization: CompuServe, Inc. (1-800-689-0736) Message-ID: <49dkl8$nc0$2@mhadg.production.compuserve.com> I'm looking for a PC utility capable of displaying WYSIWYG, or near WYSIWYG, representations of ISO12038 Mathematics. Any advice would be greatly appreciated. Thanks in advance, - Karl From: islander@msen.com (Michael Bauser) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: Slight flame on SGML Date: 28 Nov 1995 02:32:04 -0500 Organization: Vagreangvbany Fbpvrgl sbe gur Cerfreingvba bs ebg13 Message-ID: \ References: <47d5ro$o3j@hamilton.maths.tcd.ie> \ <47o0tp$fjb@icarus.convex.com> <19951108T121248Z@naggum.no> <47tn8b$m7r@news.iastate.edu> <487tql$mus@dove.nist.gov> <48bv00$79@doc.jmu.edu> \ <48q6u9$5d3@dove.nist.gov> \ <48t627$fr6@voyager.Internex.NET> Reply-To: islander@msen.com (Michael Bauser) -----BEGIN PGP SIGNED MESSAGE----- Following-up from \ In article <48t627$fr6@voyager.Internex.NET>, davido@apocalypse.org (David Ornstein) wrote: > smishra@gaia.cc.gatech.edu (Sunil Mishra) wrote: > At the risk of stepping out having not read all the important > documents in detail, let me ask about style sheets and markup. It has > seemed to me that the CLASSes being added to the various tags (e.g., > \) are really a poor-man's way to add new tags (e.g., > \). Is this basically right? > > Ideally (of course?) browsers would support displaying custom tags > defined in DTDs that are specific to the documents being viewed. I > know they don't do this (I assume because it's a lot of work and they > have too many other things on thier minds), Poppycock. Balderdash. Foolishness. They don't do this because they know that most people couldn't/wouldn't learn enough SGML to write their own DTD. For God's Sake, Marc "Excuse me while I take over the Web" Andreesen can't create SGMLable extensions -- what makes you think all his followers could? > but why didn't the style > sheet folks design something that was a bit more in line (like > allowing the style sheet information to define the new tags for a > browser using stuff in the header -- thus avoiding the full complexity > of DTDs)? The Cascading Style Sheet proposal does have a limited ability to define new tags, although it's not the primary thrust of the proposal. (I wouldn't really recommend using it, myself -- suddenly defining new tags isn't very backward compatible, after all.) Besides, DTDs are about syntax, not representation. I can create a DTD that tells a parser where \ is usable, but I can't really order the display to blink. SGML is not a command language. -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBMLfGycRHZFQbZSuZAQHlywP/QUb6d4eQpiSGTmRoXm15zF27maEGvnp4 93wh3tTzwK5fAq2LHzdsIyHqRVQllpbOzhlhB/NAjMRDG/5B90Ehj9THIMAL1zx6 7QZZ/VXcoR8IC6SFT1Niwb18q4uSgle0qJsojVuDXNLeRtz4ngncRDYfHEYOUa2R WOK88TqeC2I= =wbLl -----END PGP SIGNATURE----- -- Michael Bauser \ 42 07 30 N, 83 08 30 W \ ---- For real, now. Finger islander@msen.com@PublicKey.com for PGP public key. I pronounce HTML as "Netscape-free markup". Live with it. From: HARTMUT BECKER \ Newsgroups: comp.text.sgml Subject: Re: SGML -> HTML Date: 28 Nov 1995 16:47:19 GMT Organization: University of Rostock Message-ID: <49feen$18i@taiwan.informatik.uni-rostock.de> References: <48g54e$8d3@news.INbe.net> To: carrasco@innet.lu "M.T. Carrasco Benitez" \ wrote: >I am looking for SGML->HTML filter, or a pointer to the FAQ where this >kind of answer could be. Mail prefered. Try SGMLS.pm ("a perl5 class library for handling output from the SGMLS and NSGMLS parsers"); URL: http://www.uottawa.ca/~dmeggins/SGMLSpm/sgmlspm.html (You have to use SGMLS/SGMLS parser for parsing your SGML documents; URL for NSGMLS: http://www.jclark.com/sp.html) With SGMLS.pm there is a special script (sgmlspl) coming which is taking a so-called specification file to handle all the elements of your document. For each start tag and end tag there is an event where you can specify what to do when the tag is occuring. You can even produce a specification file template with sgmlspl. You just have to fill the events with Perl5 code. I have been using SGMLS.pm/sgmlspl for one week. It seems to be quite a good tool. At the same time I started learning Perl. (What an awful language; why is there no well-designed C++ class library available for this?!) Hartmut Becker (hbecker@informatik.uni-rostock.de) From: mwm@contessa.phone.net (Mike Meyer) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: hdp:// & sdp:// (Was: Slight flame on SGML) Date: Tue, 28 Nov 95 09:32:37 PST Organization: Missionaria Phonibalonica Message-ID: <19951128.75B52E8.8D50@contessa.phone.net> References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48vf7f$nuc@dove.nist.gov> \ In \, pcg@aber.ac.uk (Piercarlo Grandi) wrote: > What do you do in this case? > \ > (hdp of course is the "hardware download protocol") seems rather hard to > do. Ah, a solution to my problem with "Download NetScape Now"): \ Download \NetScape Now,\ and if you're foolish enough not to own a PC, get \one of those\ with your choise of OS: \Windows NT\, \, Windows 95\ or \Unix.\ \ From: Martin Morrison \ Newsgroups: comp.text.sgml Subject: Re: SGML, The New World Tour CD-ROM Date: 28 Nov 1995 18:25:17 GMT Organization: Shared Medical Systems Message-ID: <49fk6d$epe@bdt.bdt.com> References: <1995Nov27.205343.12374@sq.com> To: pmarks@allegra.com -- Martin Morrison Sr. Technical Writer/Documentation Coordinator Shared Medical Systems - Turnkey Systems Division From: jenglish@crl.com (Joe English) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml Subject: Re: Slight flame on SGML Followup-To: comp.text.sgml Date: 28 Nov 1995 10:06:35 -0800 Organization: Mentally ill venomous purists with strange emotional drives Message-ID: <49fj3b$5ef@crl2.crl.com> References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48q6u9$5d3@dove.nist.gov> \ <48t627$fr6@voyager.Internex.NET> David Ornstein \ wrote: >[...] It has >seemed to me that the CLASSes being added to the various tags (e.g., >\) are really a poor-man's way to add new tags (e.g., >\). Is this basically right? Exactly. The "usual" way to accomplish this in SGML is to define a "meta-DTD" or "architecture" for a particular application (e.g., online delivery), and then map elements in the source DTD ("CHAPTER", "SECTION") to architectural forms in the meta-DTD ("DIV"). This can be done with a link process definition or by adding #FIXED attributes to the source DTD: \ ... ]> This way the element types in the source DTD reflect what's important semantically ("this is a chapter"), and an attribute tells a specific application (e.g., a Web browser) how to process them ("this is a hierarchical division"). HTML 3 with the CLASS attribute is an SGML architecture turned inside out. With this approach, the _element types_ tell the application how to process each element, and an _attribute_ specifies the semantically important information: \
This is a nice compromise that doesn't require too much SGML sophistication on the part of authors or Web browsers, but it allows at least some semantic information to be retained when converting other SGML documents into HTML. >Ideally (of course?) browsers would support displaying custom tags >defined in DTDs that are specific to the documents being viewed. I >know they don't do this (I assume because it's a lot of work and they >have too many other things on thier minds), but why didn't the style >sheet folks design something that was a bit more in line (like >allowing the style sheet information to define the new tags for a >browser using stuff in the header -- thus avoiding the full complexity >of DTDs)? With SGML, it's not possible for a stylesheet to define new element types ("tags"); that can only be done in the DTD. At least one earlier version of HTML 3/HTML+ (circa late 1993) provided a mechanism whereby authors could define new element types. If memory serves, it looked something like: \ \ \ \ \ \ ]> \ \... \ \
You can use \person, \place, and \thing tags in this document. ... \\ This scheme was eventually rejected. I expect that HTML 4 (if and when that happens) will be defined as an SGML architecture instead of an SGML DTD. --Joe English jenglish@crl.com From: mchepeka@microstar.com () Newsgroups: comp.text.sgml Subject: Microstar announces release of NEAR & FAR Author Date: Tue, 28 Nov 1995 21:14:21 GMT Organization: HookUp Communication Corporation, Oakville, Ontario, CANADA Message-ID: <49fu4r$cup@noc.tor.hookup.net> FOR RELEASE IMMEDIATELY MICROSTAR ANNOUNCES RELEASE OF NEAR & FAR(r) AUTHOR, STRUCTURED AUTHORING ENVIRONMENT FOR MICROSOFT(r) WORD 6.0 OTTAWA, CANADA, November 27, 1995 -- Microstar Software Ltd. today announced the release of NEAR & FAR Author, its structured, graphical authoring product for Word. Designed as a Word add-on, NEAR & FAR Author enables writers to create structured documents based on the international standard, SGML (Standard Generalized Markup Language) right in their familiar Word environment with no knowledge of SGML required. "No other software today lets authors create and edit structured documents quickly and easily with no knowledge of SGML required -- all inside their Microsoft Word environment," said Shaun McEwan, Microstar's President. NEAR & FAR Author uses a graphical document model to act as the document's structure guide. Work can be stored as a Word document or easily saved as an SGML file. Writers can import and export SGML documents and format or print the documents using Word's extensive feature set. NEAR & FAR Author's key user-driven features include: Quick startup Graphical tree map of document model Intuitive interface On-line author's guide Continuous validation Requires no knowledge of SGML syntax Interactive DTD view WYSIWYG on-screen display Peter Jordan, Microstar's Chairman and CEO, explained that SGML documents are "intelligent documents." "Intelligent documents save businesses time and money. Information can be standardized and reused, at the same time that issues like document conversion are eliminated. With NEAR & FAR Author for Word, corporations are taking the first step towards enterprise-wide document management." "Not many writers know SGML, but almost everyone knows Microsoft Word. That gives us a bigger pool of resources to draw from and we can eliminate the time spent training writers how to produce SGML-compliant documents," said Brian Fitzpatrick, of US-based Information Handling Services, an information access solutions provider. Jordan also pointed out, "With the introduction of NEAR & FAR Author, businesses that want structured documents do not need to invest heavily in tools and retraining. They also gain the added advantage of a wider variety of publishing destinations including CD-ROM and the World Wide Web." * Microstar's introductory pricing for NEAR & FAR Author is $249 (US) per copy, 30% off the regular retail cost until January 31, 1996. -30- Microstar Software Ltd. is a publicly traded software and services company listed on The Toronto Stock Exchange (MSS). NEAR & FAR and C.A.D.E. are registered trademarks of Microstar Software Ltd. Microsoft is a registered trademark of Microsoft Corporation. For more information please contact: Mandy Chepeka Marketing Communications Specialist Microstar Software Ltd. 3775 Richmond Road Nepean, Ontario Canada K2H 5B7 Ph: 613-596-2233 Fx: 613-596-5934 Internet: mchepeka@microstar.com OR Cathy Browne Account Director Hill & Knowlton 55 Metcalfe St., Suite 1300 Ottawa, Ontario Canada K1P6L5 Ph: 613-786-9951 Fx: 613-238-8642 Internet: cbrowne@ott.hillknowlton.ca From: Michael Hilton <100016.3262@CompuServe.COM> Newsgroups: comp.text.sgml Subject: How many elements do you have? Date: 28 Nov 1995 21:54:02 GMT Organization: - Message-ID: <49g0dq$309$1@mhafc.production.compuserve.com> The DTD I use has over 100 elements & about 20 PIs. Is that a lot? It certainly makes it harder writing conversions to other things. -Michael -- Michael Hilton From: John Lamp \ Newsgroups: comp.text.sgml Subject: Re: Is there a thing such as http://www.sgmlopen.home? Date: 29 Nov 1995 01:33:01 GMT Organization: University of Tasmania, Department of Computer Science Message-ID: <49gd8d$lbj@franklin.cc.utas.edu.au> References: \ kaikow@standards.com wrote: > >Or something like that? Yes, have a look at the links on http://lamp.cs.utas.edu.au/net.html Cheers John -- _--_|\\ John Lamp, originating in Hobart, Tasmania / \\ Phone: 002 20 2375 - Fax: 002 20 2913 \\_.--._/ email: John.Lamp@cs.utas.edu.au v <--<< http://lamp.cs.utas.edu.au/jw_lamp.html Information Systems Group, Dept of Comp Science, Uni of Tasmania Newsgroups: comp.text.sgml From: jjr@panini.att.com (Jeff Rocca) Subject: SGML_PATH (sgmls vs. nsgmls) Message-ID: \ Sender: news@nntpa.cb.att.com (Netnews Administration) Organization: AT\&T Bell Laboratories Date: Wed, 29 Nov 1995 00:55:00 GMT I have been using the SGML_PATH environment variable with sgmls to specify where the entity manager should look for files. For example, if my DTD file contains the following lines \ %foo; and I export SGML_PATH=/usr/home/entities/%N prior to running sgmls, sgmls correctly finds the file named "foo" in the directory named /usr/home/entities. I have now migrated to nsgmls and discovered that the SGML_PATH environment variable is no longer available. How can I achieve the same results with nsgmls? Thanks, Jeff Rocca jjr@pez.att.com From: j_mcarthur@BIX.com (Jeffrey McArthur) Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 29 Nov 1995 00:59:50 GMT Organization: \ Message-ID: <49gba6$qrv@news2.delphi.com> References: \ Marcy Thompson (Netcom Online Communications Services (408-241-9760 login: guest)) wrote: >Do not be deceived into believing that document analysis is the act of >"recording what is there". It is the act of expanding your vision to >see what is there and what is not and then creatively designing an >architecture that supports both your notions of what is there and >your ideas of what is not there. Designing a DTD is a profoundly >creative act, and if you approach it as the task of "tagging what >I see", you will get the wrong DTD, I promise. My opinion is simple: read Deming. Dig out a copy of "Sample Design in Business Research" ISBN 0-471-52370-4. Think about applying Deming's standard design philosophies to DTD development. ---- Jeffrey M\\kern-.05em\\raise.5ex\\hbox{\\b c}\\kern-.05emArthur a.k.a. Jeffrey McArthur email: j_mcarthur@bix.com home: (410) 290-6935 The opinions expressed are mine. They do not reflect the opinions of my employer. My access to the Internet is NOT paid for by my employer. My access to the Internet is on my own time at my own expense. Newsgroups: comp.text.sgml Date: Wed, 29 Nov 1995 10:48:33 +0100 From: St�phane Bornnerand \ Message-ID: <199511290948.KAA22985@cygne.ais.berger-levrault.fr> Subject: SGML'95 - AIS AIS will be at the GCA SGML '95 Conference and Expo, Sheraton Boston Hotel and Towers, Boston, December 4-7. AIS will present the new release of the Balise(TM) product. Balise is an SGML document processing language with usual features to handle text and specialized features to handle SGML structures under various representations such as events and trees. The main characteristics of Balise could be listed as follows: � Compiled/interpreted language � Automatic memory management � Portable (source and binary level) � Modular and Extensible Architecture (5 set of APIs) � C-like syntax with object-oriented declaration � General purpose with SGML knowledge buit-in at the core � Mixed SGML document processing mode (event-driven, Tree navigation) � SGML Context handling based on both backward and look-ahead � Source level debugger The Balise V3 supports the new features: New embedded parser This new Balise release is based on James Clark's new SP parser, which supports all SGML features (except CONCUR) and allows for any SGML declaration. Support for double-byte characters Balise is now based on Unicode for internal character representation and fully supports double-byte encoding systems, including for text search and regular expression matching functions. JIS, Shift-JIS and EUC are accepted as input/output coding systems. Built-in ESIS-tree manipulation primitives The direct tree-manipulation primitives available through the Polypus library in previous versions are now part of the Balise kernel, allowing for improved performance, reduced memory consumption, and seamless integration with standard Balise programming. Built-in source-level debugger Balise now provides a full-blown source-level debugger as part of the Balise program itself. Large programs can now be debugged easily. New DLL based architecture On most platforms (Unix, Windows, Windows NT), Balise is now implemented as a set of dynamic libraries, allowing easy substitution of some components (such as the SGML parser) and easy integration with other software. Balise SDK The Balise Software Developer's Kit draws upon the DLL architecture to provide programmers with a full set of APIs to extend Balise and/or to use it as an embedded engine. New supported platforms Balise V3 is available on: DOS (32 bits), Windows 95, Windows NT, SUN/Sun-OS, SUN/Solaris, HP9xxx/HP-UX, RS6000, SGI, DEC alpha/OSF, Solaris X86, Unixware. In Europe, AIS will also partcipate to Documation '95 Conference and Expo, Hotel Meridien Etoile, Paris, France, December 12 and 13. AIS will demonstrate DynaPage(TM) product which is based on Balise. DynaPage is a Page Composition Gateway which fills in the gap between SGML and page makeup systems such as MS Word, FrameMaker, and Interleaf. For further information, contact Isabelle Bornier: phone: [+33] 1 46 40 84 15 FAX: [+33] 1 46 40 84 10 email: ibor@ais.berger-levrault.fr Regards, Stephane Bornerand. _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ _/ Stephane Bornerand phone: [+33] 1 46 40 84 06 _/ _/ AIS S.A. FAX: [+33] 1 46 40 84 10 _/ _/ 35, Rue du Pont email: sbor@ais.berger-levrault.fr _/ _/ 92200, Neuilly, FRANCE WWW: http://www.ais.berger-levrault.fr _/ _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/ From: LSBUMGAR@vax1.acs.jmu.edu (lee s. bumgarner) Newsgroups: comp.text.sgml,comp.infosystems.www.authoring.html,alt.culture.www Subject: Re: Content Markup & Blackbird Date: 29 Nov 1995 03:02:33 GMT Organization: Undertoad Hypermedia Publishing Co. Sender: LSBUMGAR@jmu.edu Message-ID: <49gig9$qjb@doc.jmu.edu> References: \ In-Reply-To: jbottoms@world.std.com's message of Mon, 27 Nov 1995 16:37:36 GMT In \ jbottoms@world.std.com writes: > There are too many areas to keep up with. I was reading in "Wired" > that Blackbird will support content markup with Microsoft Markup > Language (MSML?) which is based on SGML. Has any information been > published on how MS will determine the content structure for the > various communities of interest? > > Also, BillG has a book out called "The Road Ahead" which may may > reference to these future trends. The mere mentioning of Blackbird makes my bloodboil...Here are my questions: 1)Seriously, how likely is it that Blackbird will take over the Web and we will have to wait for Blackbird98 to update the tech? 2)What is Hytime vis-a-vis Hyper-G... 3)How likely is it that the current http://www.xxxx.com setup, etc will be changed to such an extent that the average user will have to learn a new URL setup? (I personally find it somewhat unlikely that the current setup is going to change dramaticlly. please email response.... -l --- -----> Undertoad (under construction) http://falcon.jmu.edu/~bumgarls/ <-------- REALITY.SYS corrupted. Reboot universe (Y/N/Q)? | "Usenet is Frosty The Snowman committing suicide with a flame thrower."-- Kibo |Somebody visited a.r.k and all I got was this lousy .sig | M$'s Blackbird is EVIL, do not support it! | ##30## From: edwardj@interramp.com (Edward Jung) Newsgroups: comp.text.sgml Subject: Re: Content Markup & Blackbird Date: Wed, 29 Nov 1995 08:43:12 GMT Organization: PSI Public Usenet Link Message-ID: <49h6hb$agp@usenet1.interramp.com> References: \ jbottoms@world.std.com (John W Bottoms) wrote: >There are too many areas to keep up with. I was reading in "Wired" >that Blackbird will support content markup with Microsoft Markup >Language (MSML?) which is based on SGML. Has any information been >published on how MS will determine the content structure for the >various communities of interest? That markup language is defined by a specific SGML DTD. The DTD is published to allow content providers to write translators. As I understand it, this DTD was defined by the Blackbird designers (and reviewed by some of their content partners). ===== My thoughts are my own ===== Edward Jung, edwardj@interramp.com Newsgroups: comp.text.sgml From: ak@oasis.icl.co.uk (Alfie Kirkpatrick) Subject: Re: How many elements do you have? Message-ID: \ Sender: news@oasis.icl.co.uk Reply-To: ak@oasis.icl.co.uk Organization: ICL, Bracknell, UK References: <49g0dq$309$1@mhafc.production.compuserve.com> Date: Wed, 29 Nov 1995 12:06:59 GMT Michael Hilton (100016.3262@CompuServe.COM) wrote: : The DTD I use has over 100 elements & about 20 PIs. Is that a lot? : It certainly makes it harder writing conversions to other things. : -Michael 100 elements doesn't sound too many. The real question is whether your typical document uses them all. Most conversions out of a DTD don't cater for every element/element context. I tend to write the conversion scripts for our document set and a couple of other docs if I can find them. Then write in clear warnings when an element is passed over so that if an author uses a previously uncatered for element, they can come and discuss it with you. You can then decide whether to cater for it or tell them not to use it! I do have the luxury of being able to discuss these issues with authors (I also do some authoring myself) and am able to change the processing whenever required. I've never heard of processing instructions being specified in the DTD (seems like a failing of the DTD design to me), but maybe that's just my limited experience... Hope this helps. Alfie. -- +-Professional Publishing Services---------------+ | Alfie Kirkpatrick ICL | | external: +44 1344 472500 Lovelace Road | | internal: 7263 2500 Bracknell, Berks | | mail: ak@oasis.icl.co.uk RG12 8SN | +------------------------------------------------+ From: tfb@ed.ac.uk (Tim Bradshaw) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: hdp:// Date: 29 Nov 1995 12:33:47 GMT Organization: AIAI, University of Edinburgh Message-ID: \ References: <47d5ro$o3j@hamilton.maths.tcd.ie> <48vf7f$nuc@dove.nist.gov> \ <49bcrp$qq4@cyberhome.cyberus.ca> \ In-reply-to: bmoore@neonramp.com's message of Mon, 27 Nov 1995 17:19:30 +0200 * Bill Moore wrote: > Sorry folks, but I missed something. hdp hardware download protocol ? It's an attempt to update and standardise the facilities provided by the well-known UUMT system (unix-unix matter transfer) provided with BSD 4.2 and later releases. --tim From: Bob Steele \ Newsgroups: comp.text.sgml,ba.jobs.offered Subject: Wanted: Software Engineers: SGML, OODBMS, DMS Date: Wed, 29 Nov 1995 11:40:35 -0800 Organization: Technology Deployment International Message-ID: <30BCB733.1EE2@tdiinc.com> CC: bsteele Wanted: Software Engineers: SGML, OODBMS, DMS Technology Deployment International (TDI) is looking to expand its expertise in object-oriented databases to SGML and Document Management Systems. We have a number of openings for engineers with backgrounds in document management and object database internals. We see that as corporations move from paper based document systems to electronic repositories, the need for sophisticated storage and retrieval will be great. Due to the complex structure of SGML documents, we believe OODBMS's are well suited to this task. We also believe the WWW will play a major role in providing the delivery of document services. This is an opportunity to influence this industry. TDI is a growing software consulting house with a great deal of expertise in object-oriented databases. Many of us have strong academic credentials in addition to our professional careers. Our web site, http://www.tdiinc.com will give an indication of our corporation and its direction. Please contact: Bob Steele, Vice President of Multimedia Information Systems Technology Deployment International 4100 Moorpark Ave, Suite 116 San Jose, CA, 95117 -- Bob Steele, TDI Inc., 4100 Moorpark Ave, Suit 116, San Jose, CA, 95117 (408) 345 9157 mailto:bsteele@tdiinc.com http://www.tdiinc.com From: kathy_graham@corp.dialog.com (Kathy Graham) Newsgroups: comp.text.sgml Subject: Re: CALS DTD table attributes Date: Wed, 29 Nov 1995 21:28:28 GMT Organization: Knight-Ridder Information, Inc. Message-ID: <49iii9$aj5@iserv.dnt.dialog.com> References: <49532v$it1@kannews.ca.newbridge.com> Paul Madsen \ wrote: >I'd like to get some info on the CALS table (and its subelements) attributes. >Is there an archive defining the purpose of each of the attributes somewhere? >Thnaks for any help >Paul Madsen >Online Documentation >Newbridge Networks Corp Another option is to get the complete CALS specification. You can get to it electronically via the Navy WEB page at: navysgml.dt.navy.mil It is a long specification, but all of the gory details are there. Hope this helps... Kathy Graham From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: How many elements do you have? Date: 29 Nov 1995 10:01:21 -0800 Organization: Mentally ill venomous purists with strange emotional drives Message-ID: <49i75h$opm@crl12.crl.com> References: <49g0dq$309$1@mhafc.production.compuserve.com> Michael Hilton <100016.3262@CompuServe.COM> wrote: >The DTD I use has over 100 elements & about 20 PIs. Is that a lot? >It certainly makes it harder writing conversions to other things. It may help to use #FIXED attributes in the DTD. For example, one of my current projects -- a series of MIL-STD-2167a documents -- has very strict preparation requirements regarding the order of paragraphs and subparagraphs. The approach I've taken is to declare an element type for each section, and use the SGML parser to validate that everything is present and in the right order. However, the processing is in most cases identical for every section-like element. For the benefit of post-processors, each section has a #FIXED attribute 'FORM=SECTION': \ \ \ \ \ [ ... ] \ \ This way, the conversion scripts only need to check the FORM attribute value instead of dozens of different element types. --Joe English jenglish@crl.com Newsgroups: comp.text.sgml Date: 29 Nov 1995 14:17:05 -0500 From: "Eve L. Maler" \ Message-ID: <199511291910.OAA21341@village.doctools.com> References: <48ihmj$bge@newsbf02.news.aol.com> <48nd2s$25j@news2.delphi.com> <1995Nov20.051628.29522@sq.com> Subject: Re: Program to take SGML source and produce a DTD lee@sq.com (Liam R. E. Quin) wrote: | SGML DTDs can serve two disparate functions. The first, and most | common, is a set of prescriptive rules, saying what is allowed where. | The second is to define a set of descriptive markup in which any | restrictions on content models are merely for convenience in later | processing, since the authoring is already done. : | If you do have editorial control, though, it's not so good, as it | turns out that any errors you make in your tagging will cause the DTD | (that is later generated automatically) to allow those errors... | | In this sort of case, another approach is incremental. I sometimes | define a simple DTD, use Author/Editor to enter (or edit) some text, or | perhaps even vi.... and then type markup into Author/Editor. I then | change the DTD to allow the new markup, save and re-import the file | (`open' uses the old rules file from the in-memory cache) and proceed | in a loop until tired. : | Still, two approaches to DTD-writing, and two kinds of DTD, points at | the ends of spectra, perhaps defining a rectangular co-ordinate | space... I can respect that people have found lots of ways to do DTD development, but... I've spent a lot of time working on and thinking about exactly this topic, and every time I've been involved in a project where there was a rush to "make the DTD," by whatever means, it's been a mess. Usually it's because (sorry, I'm probably about to offend somebody) the hacker type in charge of the DTD was too enthusiastic about coding to stop and listen to the people who will need to use the DTD. I can see where it could be useful to spot-tag a document for content analysis and then back-form a DTD for it, but only if: you already know what is to be considered important, you're a DTD expert with experience at balancing all the inherent tradeoffs (and a parser in your head :-), and the DTD has a single consumer: you. Otherwise you'll be sorry when it comes time to deploy the DTD in any real way: no specification, no rationale management, no documentation, and often no "portability" (suitability for general use) to SGML environments other than the one for which you developed it originally. There can be many motivations for tagging particular pieces of information, and many ways to do it. It's not a reductionist task of "identifying the data so it doesn't get lost." There's an art to defining what (in your particular case) counts as "data" (being sensitive to your goals and tools for creation, management, processing, etc.); naming the element types and attributes; choosing methods for identifying the data (multiple elements? same element in multiple contexts? attribute value?); and so on. Implementing a chosen DTD design is easy, whether it's in vi (my usual weapon of choice) or Near\&Far; doing the design is not a trivial task. Regards, Eve -- \ \ \ \ From: jsbell@acs.ucalgary.ca (Joshua Sean Bell) Newsgroups: comp.infosystems.www.authoring.html,comp.text.sgml,alt.fan.mozilla Subject: Re: hdp:// Date: 30 Nov 1995 00:57:19 GMT Organization: University Computing Services (The University of Calgary) Message-ID: <49ivhf$100s@ds2.acs.ucalgary.ca> References: <47d5ro$o3j@hamilton.maths.tcd.ie> <49bcrp$qq4@cyberhome.cyberus.ca> \ \ Tim Bradshaw \ wrote: >* Bill Moore wrote: >> Sorry folks, but I missed something. hdp hardware download protocol ? > >It's an attempt to update and standardise the facilities provided by >the well-known UUMT system (unix-unix matter transfer) provided with >BSD 4.2 and later releases. While we waste bandwidth (hey, Bill Gates says it'll be unlimited soon; I guess anyone who pays to download this can send the... invoice... to him. :) XEROX ANNOUNCES HYPER-ETHERNET SAN FRANCISCO, CA., Jan 7, 2010 -- Xerox today announced Hyper-Ethernet, its fourth-generation local area network. In addition to its ability to transmit text, data and images, Hyper-Ethernet enables the transmission of people. "People transmission over Hyper-Ethernet," according to Michael Liddle, V.P. of Office Systems, "will greatly reduce elevator congestion and eliminate the need for video conferencing." Order taking for Hyper-Ethernet will begin next month. Installation will start in Los Angeles in the Third Quarter. In a related announcement, Wang Labs, headquartered in Hoboken, New Jersey, announced Super-Hyper Wangnet, its twelfth generation local area network. According to Freddie Wang, President of Wang Labs, "Super-Hyper Wangnet will not only transmit people over the Wangnet, but will also transmit furniture and buildings over the interconnect and utility bands. These additional capabilities of Super-Hyper Wangnet are vital to the emerging office of the future." Order taking for Super-Hyper Wangnet will begin next month. Installation has already occurred worldwide. IBM Corporation, which has been rumored to be about to announce a local area network since 1980, was not available for comment. Digital Responds to Hyper-Ethernet TEWKSBURY, MA, April 1, 2010 -- Digital Equipment announced today its new DECNet Phase XVIII Architecture. In response to recent Xerox and Wang improvements to Ethernet that provide people- and facility- transportation across inter-node links, DEC's latest DECNet provides these capabilities as well as providing for the creation of virtual facilities and even countries. These capabilities are provided by breakthroughs in communications technology that actually uses the Ether as a communications medium. Through the use of a new dedicated NANO-PDP-11/E99 gateway processor system, ETHERGATE, DECNet users can access anywhere in the Ethereal Plane. This development obsoletes teleconferencing, since meeting groups can create their own common conference rooms and cafeterias, thus resolving space, travel, and dining problems. There may be a few bugs left, as some of the dissenting DECNet Review Group members have not been seen since the last meeting held in such a virtual conference facility. This breakthrough was brought about by a team of the Distributed Systems Software and Hardware engineering teams in an effort to improve on their Tewksbury, Massachusetts, facility. In a compromise decision, Distributed Systems will maintain an ETHERGATE in TWOOO but it will connect directly to their new home somewhere in the Shire of their newly defined Middle Earth reality. Despite some difficulties, the scenery, windows, tax breaks, pool, and racquetball courts made the relocation go quite smoothly. Engineering Network topology will not change, as all forwarding will be done by the TWOOO Ethereal Plane Router residing in the crater at the former building site. Utility packages such as Ethereal Person Transfer (EPT) and Ethereal Facility Transfer (EFT) provide appropriate capabilities for casual users. Sophisticated users can create ($CREATE), access ($OPEN) and delete ($NUKE) ethereal entities transparently from high level languages using the Ethereal Management System (EMS) package and the Ethereal Access Protocol (EAP). An ETHERTRIEVE utility for easy interactive use will be available shortly. DECNet Phase XVIII follows on the success of the Phase XVI ability to access everyone's Digital Professional Wristwatch computer system. The lead to the current Phase XVII architecture, which has routing capabilities that allow direct communications with the entire Earth population's Atari home video games. Distributed Systems architects are hard at work on the next phase of DECNet that will include multi-plane existence network management (using the NIECE protocol) and galaxy level routing using 64K-bit addresses. Digital will continue to support its Gateway products into the Prime Material Plane. These products include an IBM ANA (Acronym-based Network Architecture) Gateway, the TOLKIEN product that allows control of all ring based networks, and our Mega-broad-jump-band hardware Sorry, I don't have an attribution for the above (I'd love one, though). It's at least 5 years old, one of the first things I came across and saved when I first ventured onto the Internet. The fact that the competitor is Digital not Novell certainly dates it, but it remains funny. :) Joshua -- _.----~~~----._ | MIME: jsbell@acs.ucalgary.ca ,--------.-.,-'-------------------` | WWW: http://www.ucalgary.ca/~jsbell/ `--------"-'-.,---`~-----~' '---'-._____/ From: jamesr@magna.com.au (James Robertson) Newsgroups: comp.text.sgml Subject: Re: Content Markup & Blackbird Date: 30 Nov 1995 00:38:01 GMT Organization: Desktop Law Message-ID: <49iuda$7k9@kettle.magna.com.au> References: \ <49h6hb$agp@usenet1.interramp.com> Reply-To: jamesr@desklaw.com.au In article <49h6hb$agp@usenet1.interramp.com>, edwardj@interramp.com says... > >jbottoms@world.std.com (John W Bottoms) wrote: > >>There are too many areas to keep up with. I was reading in "Wired" >>that Blackbird will support content markup with Microsoft Markup >>Language (MSML?) which is based on SGML. Has any information been >>published on how MS will determine the content structure for the >>various communities of interest? > >That markup language is defined by a specific SGML DTD. The DTD is >published to allow content providers to write translators. As I >understand it, this DTD was defined by the Blackbird designers (and >reviewed by some of their content partners). > > Has anyone actually seen the DTD? Is it on the Microsoft Web site? If so, where? Any word on how compatible it is with HTML 2.0? Does it contain a lot of horrific Word-isms, or have they done proper job? Yours curiously, J -- ---------------------------------------------------------------------------- James Robertson * jamesr@desklaw.com.au Coding * Design * Layout * Newton * Windows "Beyond the idea" From: hayashia@dps.ntts.co.jp (Akihiro Hayashi) Newsgroups: comp.text.sgml Subject: elisp for mule Date: 30 Nov 1995 03:47:48 GMT Organization: NTT Software Corp., Yokohama, Japan Message-ID: <49j9h4$cmc@nttsgw.yh.ntts.co.jp> Hi, Netters! Do you know some elisp which helps me make sgml, DTD files, like cmutex.el or html-helper-mode.el... ? -------- Akihiro Hayashi \ From: Dathan@delta.pp.fi Newsgroups: comp.text.sgml Subject: HTML and SGML Authoring Date: 30 Nov 1995 11:59:22 GMT Organization: DELTA COMPUTERS LTD OY Message-ID: <49k6aq$3kn@idefix.eunet.fi> Reply-To: Dathan@delta.pp.fi Summary: FREE DEMOS Keywords: HTML,SGML We are pleased to announce the use of our WWW home page for FREE fully featured working demonstrations of the SGML TagWizard and WebWizard products for authoring HTML and SGML. URL:http://www.jsp.fi/delta From: bkline@cortex.nlm.nih.gov (Bob Kline) Newsgroups: comp.text.sgml Subject: Q: what's a caron? Date: 30 Nov 1995 12:22:01 GMT Organization: National Library of Medicine Message-ID: <49k7l9$lj3@lhc.nlm.nih.gov> I've been scrounging around trying to track down the meaning of the word "caron" which I came across in one of the ISO documents on character sets (ISO 8879:1986) while looking for the official SGML element names for the uppercase and lowercase "Polish L" characters (as the USMARC standard refers to them). The ISO document doesn't refer to anything as a "Polish L" but does have two possible candidates (that is, two each for the lowercase and uppercase versions of Polish L), referred to as "L caron" and "L stroke." The "L stroke" seems possible because the characters in question do have a short stroke through the middle of the vertical, from lower left to upper right (and in fact, the older USMARC documents refer to the characters as "L slash" which seems close enough to "L stroke"). But I can't rule out "L caron" as a possibility because I can't determine what kind of mark "caron" refers to. Unfortunately, the ISO documents themselves (or at least the versions to which I have access) don't provide the graphic represntation of the characters, and I don't have any really specialized reference sources to turn to. Things like the Oxford English Dictionary (both the original and new editions) or Webster's International Unabridged (both 2nd and 3rd editions) don't have anything. I even tried a Polish dictionary, but since I don't know anything about Polish, I'm afraid I wasn't able to make much intelligent use of that source. Can anyone tell me what a "caron" is? Email response if possible, please. Thanks! Bob Kline bkline@cortex.nlm.nih.gov bob_kline@stream.com From: kmoran@riatax.com (Kevin J. Moran) Newsgroups: comp.text.sgml Subject: Re: SGML book? Date: Thu, 30 Nov 1995 16:59:42 GMT Organization: Research Institute of America, New York, NY Message-ID: <49ko8g$4cj@news.tpp.com> References: <480i5v$hsp@usenet.ucs.indiana.edu> \ <1995Nov20.045420.29244@sq.com> <1995Nov20.214942.10037@sq.com> \ \ Reply-To: kmoran@riatax.com liora@delphi.com wrote: \ >Excerpts from ABCD... SGML are available at: > >http://www.thomson.com/itcp/liora/LA_HOME.HTML > Try http://www.thomson.com/itcp/liora/LA_HOME.HTM instead. From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: 30 Nov 1995 11:14:52 -0800 Organization: Mentally ill venomous purists with strange emotional drives Message-ID: <49kvrc$rva@crl2.crl.com> References: <48ihmj$bge@newsbf02.news.aol.com> <48nd2s$25j@news2.delphi.com> <1995Nov20.051628.29522@sq.com> <199511291910.OAA21341@village.doctools.com> Eve L. Maler \ wrote: >[...] every time I've been involved in a project where there was a >rush to "make the DTD," by whatever means, it's been a mess. Usually it's >because (sorry, I'm probably about to offend somebody) the hacker type in >charge of the DTD was too enthusiastic about coding to stop and listen to >the people who will need to use the DTD. > >I can see where it could be useful to spot-tag a document for content >analysis and then back-form a DTD for it, but only if: you already know >what is to be considered important, you're a DTD expert with experience at >balancing all the inherent tradeoffs (and a parser in your head :-), and >the DTD has a single consumer: you. Otherwise you'll be sorry when it >comes time to deploy the DTD in any real way: no specification, no >rationale management, no documentation, and often no "portability" >(suitability for general use) to SGML environments other than the one for >which you developed it originally. Hear, hear! I can attest to this personally. (And wish I had heard your advice when I first learned the mechanics of SGML...) Of course one should not underestimate the usefulness of hacking together DTDs for personal use; there's no better way to gain the requisite experience. Every DTD I write is for my use alone; I figure after another year's practice or so I'll be ready to do it for real :-) --Joe English jenglish@crl.com From: jeffrey@hks.com (Glenda Jeffrey) Newsgroups: comp.text.sgml Subject: Marked sections inside entities? Date: 30 Nov 1995 18:25:23 GMT Organization: Hibbitt, Karlsson & Sorensen, Inc. Message-ID: <49ksuj$8qd@argo.hks.com> I hope someone can suggest a way to do what I'm trying to do here. I need to set up a bunch of entity definitions (hundreds, actually) that look something like this: \The Title\, Section \\" > I'd like for these definitions to be compiled into my DTD. (I'm using ArborText Adept.) Adept doesn't seem to like this situation unless I define %booka; and %bookb; before defining myxref. I tried adding these definitions to my DTD: \ \ but then, of course, Adept just replaces %booka; and %bookb; with "INCLUDE" -- not much use. I've read about the MS keyword, as in: \ but this setup means that my entire entity definition has to be a single marked section -- not what I'm after. Does anyone have any clever ways of doing what I'm trying to do here? A little background, in case it's not already clear what I'm trying to do: I have two books which are produced from the same SGML source. I do the cross referencing within the books using an autogenerated list of entities, like %myxref; above. (I could autogenerate them on the fly, but this would mean I'd have to print the entire book every time -- impractical at best.) I'd like to use marked sections %booka; and %bookb; to select which cross reference will get picked up when I print. Thanks in advance for any suggestions! -- Glenda Jeffrey Email: jeffrey@hks.com Hibbitt, Karlsson & Sorensen, Inc Phone: 401-727-4200 1080 Main St. Fax: 401-727-4208 Pawtucket, RI 02860 From: amgreene@ATHENA.MIT.EDU (Andrew Marc Greene) Newsgroups: comp.text.sgml Subject: Re: Content Markup & Blackbird Date: 30 Nov 1995 22:55:15 GMT Organization: Massachusetts Institute of Technology Message-ID: <49lcoj$365@senator-bedfellow.MIT.EDU> References: \ In article \ jbottoms@world.std.com (John W Bottoms) writes: >There are too many areas to keep up with. I was reading in "Wired" >that Blackbird will support content markup with Microsoft Markup >Language (MSML?) which is based on SGML. Has any information been >published on how MS will determine the content structure for the >various communities of interest? Blackbird Markup Language (BBML or BML, the folks at Microsoft can't seem to make up their minds :-) is basically HTML minus forms minus nested lists plus OLE embedding. To claim that Blackbird is SGML is as inaccurate as to claim that "the Web" is SGML. Both Blackbird and "Web browsers" use SGML document types; Blackbird actually enforces its DTD (as opposed to Netscrape, for instance); but neither Blackbird nor any of the mainstream HTML-based browsers are capable of handling any DTD except the one hardocded into their calcified little arteries. - Andrew Greene [Disclaimer: I co-authored the original Blackbird Word import and export filters about six months ago. For that I humbly apologize to the SGML community and to the world at large.] -- Andrew M. Greene \ http://www.mit.edu:8001/people/amgreene Newsgroups: comp.text.sgml From: robinw@andyne.on.ca (Robin Wynne-Edwards) Subject: SGML 95 - Document Databases Message-ID: \ Sender: -Not-Authenticated-[4949] Xdisclaimer: No attempt was made to authenticate the sender's name. Organization: Andyne Computing Ltd. Date: Thu, 30 Nov 1995 20:17:22 GMT Andyne Computing Limited will be sponsoring two focus groups on The Future of Structured Document Databases during SGML 95 in Boston. The focus groups are part of an Andyne research initiative, code named Rosetta, that will be launching an object oriented document database in the third quarter of 1996. The Agenda is: Introduction What Should a Document Database Do (and Not Do) Non-SGML File Support Issues Document Query Wish List The Ideal SGML Product Suite The Focus Groups will be held in The Clarendon Room, Third Floor, Sheraton Boston on Tuesday December 5 at 3.00pm and again at 4.00pm. SGML Users and VARs are invited to attend. If you are interested in attending please look for the sign up sheets on bulletin boards at the conference or respond by eMail. Space is limited. Newsgroups: comp.text.sgml From: pbg (Paul Grosso) Subject: Re: Is there a thing such as http://www.sgmlopen.home? Message-ID: <1995Nov30.152635.15058@arbortext.com> Sender: news@arbortext.com Organization: ArborText, Inc. 1000 Victors Way Ann Arbor, Michigan 48108 Date: Thu, 30 Nov 1995 15:26:35 GMT > From news Wed Nov 29 18:45:45 1995 > Newsgroups: comp.text.sgml > From: kaikow@standards.com > Subject: Is there a thing such as http://www.sgmlopen.home? > Date: Wed, 22 Nov 1995 22:28:18 GMT > > Or something like that? I'm not sure why this is dated Nov 22, but someone on the 29th just gave some unexpected response, and I haven't seen the right response, so excuse me if this is a repeat. It's http://www.sgmlopen.org paul Paul Grosso Chief Technology Officer, SGML Open paul@arbortext.com =================================== From: Francis Cave \ Newsgroups: comp.text.sgml Subject: Re: Program to take SGML source and produce a DTD Date: Thu, 30 Nov 95 23:22:54 GMT Organization: Pira International Message-ID: <817773774snz@pira2.demon.co.uk> References: <48b8rc$qpd@manuel.anu.edu.au> Reply-To: cave@pira2.demon.co.uk There's a clever gadget called the DREAM Parser that might interest you, but I'm not sure how available it is, if at all. It was developed at GMD-IPSI in Darmstadt and, as I understand it (which, as a non-user, is necessarily only to a limited extent), is a tool for converting to SGML with or without a pre-existing DTD. The rules that one writes to perform the conversion are declaration-like and can be used to generate a DTD "post hoc", or so I believe. I suppose this implies that the person writing the conversion must have the DTD in their head to start with. Being lazy, I find the idea of a tool that would enable me to write a DTD _and_ the conversion program at one go very appealing. I too like the statistical approach to DTD derivation from tagged text. When a colleague of mine was recently faced with working out a DTD suitable for coding 30 years-worth of journal articles - having decided that none of the public DTDs was quite right and with no clear picture as to what coding would be likely to prove important in the future - it was decided that the best policy was to write a DTD that describes what is there and leave the worrying about what else might be needed until the need arises. -- Francis Cave