Newsgroups: alt.hypertext,comp.text.sgml Date: Tue, 13 Feb 1996 23:14:31 -0600 Message-ID: <9602140514.AA03483@fly.HiWAAY.net> From: Len Bullard \ References: <31164A18.25C4@passage.com> <4fnqa5$p8t@Venus.mcs.com> \ <4fqbaf$rl4@Mercury.mcs.com> Subject: Re: Strongly-Typed Hyperlinks [Jorn Barger] >To argue, as Len did in a quote I've reposted *many* times, that no >progress could be made on the WWWeb because it lacked strong types, is >*arrogant bullshit* and bad science, because there's clearly no proof. Yes it is ... whatever you said. >Len is a massive bullshit artist. I've proved this in enough detail to >call a spade a spade. Yes I am. Whatever you say. Now, the issue was strongly-typed hyperlinks. Paul Burchard \ sends me the following with permission to quote him here: >By "session handles" I mean that the URL would contain some kind of >globally unique session+transaction IDs, and that part of the algorithm >for retrieving the resource would be that the client must be in the >designated state in order to properly use the rest of the location info in >that URL. Of course such URLs would have *very* limited lifetime and use, >basically being offered by a stateful server for immediate and >non-repeatable consumption by a specific client. >The approaches in the Z39.50 URL draft yield resources that are less >ephemeral, and more useful. The first approach is to specify how to >initialize a stateful session, but to let the user perform the subsequent >stateful interactions (like telnet: URLs). The second approach covers >resources which can be accessed without knowledge of state, even though a >stateful protocol happens to be required to accomplish that access (like >ftp: URLs). >P.S. Feel free to forward this info as you think appropriate. So, my original question about Z39.50 is pertinent. What is the expense of using it? Tomorrow I will post a MID relationship arch form for declaring strongly typed links based on the HyTime ilink architectural form. I would very much appreciate comments on how it can be implemented in Z39.50 or HTTP. This would be most helpful in our IETM research. IETMs typically run in standalone environments, but the ability to use the existing protocols would greatly reduce the cost of development and fielding as well as further consensus in our community. thanks, Len From: Bj|rn-Helge Mevik \ Newsgroups: comp.text.sgml Subject: Re: SGML -> RFT Converter Date: 14 Feb 1996 10:30:42 GMT Organization: University of Oslo, Norway Message-ID: <4fsdki$g3j@ratatosk.uio.no> References: <31206E03.5F9D@prl.research.philips.com> In-reply-to: "M.Bearne"'s message of Tue, 13 Feb 1996 10:54:59 GMT If I am not mistaken, the documentation of linuxdoc-sgml-1.5 says that it converts instances of the linuxdoc document type to RTF. I haven't tried it, though. B/H -- Bjørn-Helge Mevik \ From: jenglish@crl.com (Joe English) Newsgroups: comp.text.sgml Subject: Re: SGML without DTDs? Date: 13 Feb 1996 16:08:29 -0800 Organization: Tagheads Message-ID: <4fr95t$d7b@crl2.crl.com> References: <4fr1pj$lm7@murphy.servtech.com> Steven R. Newcomb \ wrote: >* Should SGML be amended so that DTDs can allow *totally unmodelled* > structure to occur in documents? I think this is already possible: \ \ Or, if you want to allow structured parts inside the unstructured part: \ Or, if you want to allow an overlapping element structure: \ \ \ Or, for _completely_ unstructured data: \ >* Does SGML's requirement that everything be modelled before it can be > expressed actually weaken the expressive power of the language? Not really, since a model of "no prescribed structure" can already be specified as it is. >* Is it a good idea to allow DTD designers the power to allow portions > of documents to be *entirely* arbitrarily structured?" > >* Is arbitrary (or chaotic) structure still a kind of structure, > worthy of being modellable? There are a couple cases where this might be useful. The grove abstraction could be a powerful tool for analyzing otherwise unstructured data. Converting existing documents with no (known) structure to SGML would enable external linking and annotations with HyTime. A "tag soup" model might be appropriate for some word-processing applications where a rigorous definition of the allowable document types is not feasible. Not to mention the Web :-) >Without voicing my own opinion on this question (which opinion is, in >any case, incompletely formed), I'd be interested to hear any thoughts >on this. I have good and practical reasons for wanting to know the >arguments pro and con. I'll gladly share my reasons if anybody is >really interested and after some opinions are expressed. Please do! --Joe English jenglish@crl.com From: Bernstein@eastgate.com (Mark Bernstein) Newsgroups: alt.hypertext,comp.text.sgml Subject: Re: Strongly-Typed Hyperlinks Date: Tue, 13 Feb 1996 20:16:00 -0500 Organization: Eastgate Systems, Inc. Message-ID: \ References: <31164A18.25C4@passage.com> <4fnqa5$p8t@Venus.mcs.com> \ <4fqbaf$rl4@Mercury.mcs.com> In article <4fqbaf$rl4@Mercury.mcs.com>, jorn@MCS.COM (Jorn Barger) wrote lots of contemptible stuff. He does raise one point which might be worth mentioning. Earlier, I had written of Barger's indexing proposal that, > > >As is his custom, Barger didn't mention that this isn't a new or original > >idea Barger ironically comments: > Which, in Mark's value system, is obviously a capital crime. This is much nearer the mark than Barger conceives. Shoddy scholarship -- which includes passing off well-known ideas as your own because you can't be bothered to read anyone else's work, or are embarassed to acknowledge anyone else's contributions -- ought not, I think, to be taken lightly. Barger's faith in the web -- which extends to a disclination to consult anything published elsewhere -- is a charming (albeit misguided) testament of faith in the power of hypertext. Personally, I think people who are interested in hypertext theory might enjoy reading some good hypertexts, a few fine books on hypertext design and hypertext criticism, and some relevant research papers on the subject. I know that when I started in the field, back before 1987, I wish that some of this existed. It didn't then, it exists now, and knowing it will help us all build better systems and write more eloquent hypertext than we can imagine. -- Mark Bernstein Bernstein@eastgate.com Eastgate Systems, Inc voice: (800) 562-1638 +1 (617) 924-9044 134 Main Street fax: (617) 924-9051 Watertown MA 02172 USA http://www.eastgate.com/ Newsgroups: comp.text.sgml Date: Wed, 14 Feb 1996 08:43:23 -0600 Message-ID: <9602141443.AA32310@fly.HiWAAY.net> From: Len Bullard \ Subject: RE: Strongly-typed links This is another paragraph from Paul Burchard's postings on the VRML list on the subject of Internet protocols. The context is a discussion between Paul and J. Gwinner on the design of address models. Note the CIS protocol Paul mentions is Compuserve. As with the last post, Paul has given permission to post this to CTS persuant to the discussions on strongly-typed links. There are requirements for the VRML system for alternate resources for link targets, e.g., "if a resource at one target is unavailable, here is an alternate". The request was made that the address model of the language not be "WWW" or Internet-constrained to enable VRML to be applied applications that do not require network aware engines. This issue affects all notations to support instances that must be interoperable, portable, and reusable while maintaining testable validity. [Paul Burchard] >Please note: URL != Internet. >A "URL scheme" is nothing more than a hierarchical naming system for a >(typically networked) resource. The definition of each URL scheme must >include an algorithm for accessing the resource associated to such a name, >if there is one. The algorithm could e.g. include making TCP/IP >connections to Internet hosts, but it could just as well involve >transactions in CIS' HMI protocol. While HTTP happens to be stateless, >there is no requirement for statelessness. Len Bullard From: censign@interserv.com (Chet Ensign) Newsgroups: comp.text.sgml Subject: Looking for contact for North Calif. SGML ug Date: Wed, 14 Feb 1996 19:09:14 GMT Organization: I-2000 Inc. - Internet Services Message-ID: <4ft0us$7dr@i-2000.com> All, Does anyone know who acts as a contact for the northern california SGML users group? I had a name but lost it when my PC crashed. You can just send email directly to me at censign@lds.com. Thanks, /chet Newsgroups: comp.text.sgml Date: Wed, 14 Feb 1996 15:48:24 -0600 Message-ID: <9602142148.AA23939@fly.HiWAAY.net> From: Len Bullard \ References: <9602122117.AA15266@fly.HiWAAY.net> \ Subject: Re: Facts, folks. We need facts. [Bernd Kreimeier] In article <9602122117.AA15266@fly.HiWAAY.net> Len Bullard \ writes: >> Cross-posted to comp-text-sgml Beg your pardon - from where? The VRML mailing list. >I've unsubscribed from the VRML lists two months ago (drowning in electric >squirrels). :-) When the windmills ran down. >Care to summarize what happened, and how exxactly it is relevant to SGML? Sure, but with this preface. My posting was the classic case of an idiot (me) getting to the SEND key before the mind engaged to stop it. I apologize to the CTS list for that. >> Well, I have read the SGI posting on the subject. >Nice. What posting. Background: The VRML standard is similar to the HTML work. A large volunteer list from all over the world has been working very hard for two years to create a virtual reality (read 3D graphics metafile format) for the Internet. SGI has been very involved and in fact, the current VRML 1.0 is actually a net-enabled version of SGI's Open Inventor. It works nicely. The current effort is focused on a more powerful version which will enable behaviors (animations, electric squirrels, etc.). This is a consensus list. No one adds things without a vote basically and that is supposed to head off the kinds of things that happened to HTML. Issue: As always, when the gig looks lucrative, some start throwing their weight around in various ways. That posting was my "mad as Hell" response when the rumors surfaced that Netscape was about to buy the best VRML browser for the PC market (WebFX) and issue a press announcement with the signatures of others in the industry (the biggest players) endorsing one of the proposals. This turned out to be true. The press release seem to indicate that Netscape, by its support of native VRML (just as they do for HTML) made other proposals moot. Mark Pesce, the List Moderator and leader of the self-appointed architecture group that honchos the process, published an article on ZDNet that appeared to label the proposal (Moving Worlds) a fait accompli and was quoted at the beginning of the press release on the SGI page. This coincided nicely with the beginning of polling on the proposal. After a year of "we don't need the ISO, we have the Net, we are a community, we can be fair", it looked very much like a hijacking was in progress. I've been beaten about the ears for the last 24 hours for my post, so, I got the point. That said, the Moving Worlds proposal IS the one that has the most involvement of the list members and it IS a technically sound piece of work. But other proposals are also sound, were promised fair consideration, and in my opinion, just took a well-aimed and well-planned shafting. Business as usual. As I have pointed out on the VRML list, ISO standards and policies are set up to make this sort of thing difficult to do. My experience with Dr. Goldfarb in SGML made me consider his refusal to endorse products the norm. That was naive of me. I applied a good example to a bad situation. I am used to industry types twisting SGML to suit their short term plans for profits, but I have never seen them get to the principals of the standards groups to achieve it. Consortiums, are where that is best done. Business. I have no problem with that. So that posting was as I said, an idiot getting to the send key too quickly. My only worthless excuse is the last weeks of being polite to Jorn finally blew a circuit and I went ballistic in another direction. Again, an idiot move, so, you're right, Barger. I'm an idiot. >Btw., I have been able to get the 400K+ Moving worlds proposal within a >minute. I have yet to get even the intro on ActiveVRML, www.microsoft.com >being about as communicado as last year when I headed for the RTF specs. Don't know. I downloaded it. Have you looked at the Apple proposal? >Besides being lost on what happened, I catch your drift. You care to >enlighten us? On the SGML aspect, that is... and where stampede happened >that threw the dust in here. OK. Rant follows. Non-bozo though. The stakes are getting too high for humor at this instant. Look carefully at how the standards for the Internet and the rest of the SGML applications are being defined. Look very carefully at the processes. In the complaint that "standards are too slow for the swiftly moving technologies" are also the seeds that enable small groups of information product providers to effectively dominate markets and eliminate alternatives. Note carefully when some of the proposals are actually watered down versions of the standard that gut the best parts and concepts without attribution, then claim credit for serving you up a hindquarter for a few cents less when you already owned the cow. Be very aware of answers to questions about why a standard can't be used that consist of "Yuck! Ugly!" All of the things an application have to do have to be considered. The efficiency of the parse and the burden on the programmer are just two. Don't throw away DTDs because you are too impatient to let a file parse completely. You lose a lot if you do that. SGML works BECAUSE of DTDs although a post-parse run time optimized transport format is a good idea. But if you lose the part that enables you to build models by a consensus process, check those models, create strongly-typed relationships among the models and document the rationale for the agreements they represent, you will have screwed the pooch in a way that you won't recover from for a generation. Do it once right, or keep paying to have it redone. Ask anyone in the conversion business about this. As I have been told, "It's Your Data, Dammit!" SGML is one standard whose originators carefully planned to protect that right. A lot of people out there want to kill it. You want to debate the future of SGML? Fine. It's your future you are debating. Ask your browser vendor representative just WHY they can't serve SGML correctly. Then buy a server that can. You won't regret it. Len Bullard From: whatis@yyz.com Newsgroups: comp.text.sgml Subject: Help! I don't understand this nsgmls error! Date: 14 Feb 1996 20:01:41 GMT Organization: You wish. Message-ID: <4ftf35$gnj@news.cerf.net> Originator: whatis@yyz.com Here's an SGML fragment that reproduces a larger problem that I'm having inside of my real DTD. It involves my first attempt at using parameter entities to encode often-used patterns of tags. ---------------------------BEGIN DOCUMENT----------------------------- \ \ \ \ \ ]> \\Duke Nukem 3D!\\ ----------------------------END DOCUMENT------------------------------ When I run this through nsgmls, I get: ----------------------------BEGIN OUTPUT------------------------------ trash.sgm:3:43:E: delimiter `(' invalid: only delimiter `&', delimiter `|', delimiter `,', delimiter `)' and token separators are allowed trash.sgm:7:1:E: DTD did not contain element declaration for document type name trash.sgm:9:7:E: element `TRASH' undefined trash.sgm:9:50:E: no document element (TRASH (SOMETAG -Duke Nukem 3D! )SOMETAG )TRASH -----------------------------END OUTPUT------------------------------- Why am I getting this bizarre error? Help please! Steven Boswell whatis@yyz.com From: jorn@MCS.COM (Jorn Barger) Newsgroups: comp.text.sgml Subject: Re: Strongly-Typed Hyperlinks Date: 14 Feb 1996 15:50:12 -0600 Organization: The Responsible Party Message-ID: <4ftlek$220@Venus.mcs.com> References: <9602132124.AA02210@fly.hiwaay.net> In article <9602132124.AA02210@fly.hiwaay.net>, Len Bullard \ wrote/quoted: >>For an example of a WWW system that uses link types in a similar way (to >>clarify where the user can look for different types of information), see: >> >> http://www.ils.nwu.edu/~e_for_e Heh! Perhaps I should have mentioned that I'm co-author of the tech report on which these linktypes are based. See \ j From: asengupt@indiana.edu (Arijit Sengupta) Newsgroups: comp.text.sgml Subject: Re: Help! I don't understand this nsgmls error! Date: 14 Feb 1996 22:14:27 GMT Organization: Indiana University, Bloomington Message-ID: <4ftms3$pg9@usenet.ucs.indiana.edu> References: <4ftf35$gnj@news.cerf.net> whatis@yyz.com wrote: : Here's an SGML fragment that reproduces a larger problem that I'm : having inside of my real DTD. It involves my first attempt at using : parameter entities to encode often-used patterns of tags. : : ---------------------------BEGIN DOCUMENT----------------------------- : \ : \ ---------------------------------------^---- error is here : ----------------------------BEGIN OUTPUT------------------------------ : trash.sgm:3:43:E: delimiter `(' invalid: only delimiter `&', delimiter `|', delimiter `,', delimiter `)' and token separators are allowed Actually, sp does a very good job of pinpointing the error - unfortunately it is a little off-standard, so the emacs C-x ` can't catch the error message - maybe James can fix that in the next release? Anyhow - that is off the point here. look at sp's output - line 3, col 43 has an open bracket immediately following a member of a content model and a space - which is illegal in sgml - you need to put a ,/|/&/) as sp suggested. So a correct version would be: \ Jit. -- _|_|_|_|_| _|_|_| _|_|_|_|_| | There cannot be a crisis next week, _| _| _| | My schedule is already full. _| _| _| | _| _| _| _| | - Henry Kissinger _|_|_| _|_|_| _| | From: jbazuzi@vt.edu (Jay Bazuzi) Newsgroups: comp.text.sgml Subject: Correct SGML comments and other correctnesses Date: 14 Feb 1996 23:28:36 GMT Organization: Blacksburg, VA.'s unemployed programmers Message-ID: <4ftr74$99b@solaris.cc.vt.edu> Summary: What's the correct comment format? Where do I find this out? Keywords: SGML comment HTML compliance RTFM -----BEGIN PGP SIGNED MESSAGE----- Hello folks, I'm sure that a bunch of you know the answer to this one. In order to reduce bandwidth I think you should email your replies directly to me and I'll summarize. What's the correct format for SGML comments? We all know that few HTML browsers get it right. Even people who complain don't always get it right. I've seen the following in the \KGV FAQ\: o A comment declaration begins with the character sequence `', and contains zero or more comments. o Comments begin and end with the character sequence `--', and may not contain `--'. Whitespace is allowed after any comment, but not before the first comment. Non-whitespace characters are not allowed before or after comments. o A `>' inside a comment does not close the comment declaration. However, Arena sometimes thinks comments like this are unclosed: \ And I've seen people say this is correct. To what internet resource can I refer to get the final word on the matter? Where is TFM I can R for such answers to SGML questions? - -- Jay Bazuzi jbazuzi@vt.edu http://neilyoung.async.vt.edu -----BEGIN PGP SIGNATURE----- Version: 2.6.2 iQCVAwUBMSJwIyVKqTw4CcPhAQGYswP/bh0zuS/kt6DEiAAX/tugKOChEpCuPrQ4 mc7Lok2PG2W/4rEjHl7kD1UuShanOrUIy/fQYGN8Q5pYMoaNfUhD/HhVciOQf69l tC1L4A18s7Wa8FU7/deI+jSPCjqhFRQwVgtxe9q+RH9D392iC7FVoG827b8sDQgV VKZxy3SpxoA= =zJlh -----END PGP SIGNATURE----- From: "Arijit Sengupta" \ Newsgroups: comp.text.sgml Subject: Re: HTML Paragraph Date: Wed, 14 Feb 1996 13:37:45 -0500 (EST) Organization: Computer Science, Indiana University Message-ID: <2656@824323070> References: <3121102D.7CD3@poly2.nist.gov> In article <3121102D.7CD3@poly2.nist.gov>, Frederick R. Phelan Jr. \ wrote: > >I am trying to make single spaced lines of html text. >When I use the \

operator, the browser automatically >puts a space between the paragraph's ... how can I stop >this behavior? anyone know? > Why would you not want space between paragraphs? If you want to force line breaks, use \
instead of \

..\

But note that if the lines are too long, some browsers will not render them correctly. Jit. -- _|_|_|_|_| _|_|_| _|_|_|_|_| | asengupt@indiana.edu _| _| _| http://www.cs.indiana.edu/hyplan/asengupt.html _| _| _| | Computer Science, LH215 _| _| _| _| | Indiana University, Bloomington IN47405 _|_|_|ust _|_|_|n _|ime | (812) 855-4318 / (812) 334-2695 From: I._Calixa-Lavallee@TeleGraphique*MTL.utlglc.upc.qc.ca (I. Calixa-Lavallee) Reply-To: I._Calixa-Lavallee@TeleGraphique*MTL.utlglc.upc.qc.ca Newsgroups: comp.text.sgml Subject: Re: SGML without DTDs? Date: 14 Feb 1996 21:07:19 GMT Message-ID: <1900510.1388241@utlglc.upc.qc.ca> Organization: TeleGraphique LC inc. well done