OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] SGML on the Web

[ Lists Home | Date Index | Thread Index ]

Liam Quin still can't post to xml-dev, but there's some great food for
thought in his messages!  

====== Forwarded Message ======
Date: 10/4/02 4:21 PM
Received: 10/4/02 4:21 PM
From: liam@w3.org (Liam Quin)
To: simonstl@simonstl.com (Simon St.Laurent)

[private reply, but feel free to forward it to xml-dev if you want,
 as I can't post there - Liam]

> XML seems to have been conceived as three parts: Syntax, Styling, and
> Linking.  In the SGML world, these three were largely represented by
> SGML itself, DSSSL, and HyTime, but all three of those were widely
> as overly complicated and were only rarely implemented completely.  

I am not aware of *any* complete implementations of DSSSL or HyTime.
As for SGML, I am not sure that James' SP is complete, because I don't
think it does all of link processing, concur and data tag, but I
might be mistaken. The SEMA/Yard parser claimed to be complete, and
it's possible omnimark's was/is complete or pretty close.

I think it's fair to say that DSSSL and HyTime were not completely
implemented in 1995/6, although I know Fujitsu were showing some
work they had done, as were a couple of other groups.  The complexity
of SGML was (and is) such that writing a parser was a major engineering
effort once you'd bought the standard and the Handbook.  James said
it took him 3 solid months just to understand the spec enough to
start coding, and he's anything but slow.

SGML came out of work done in the 60s, primarily for document
processing and typesetting, and has a lot of practical, hands-on
features (e.g. datatag, shortref, omittag, shorttag) that may have
helped adoption early on, but in the 1990s were a handicap.

> XML directly on the Web seems to have fallen victim to the notion that
> XML needed a transformative style approach,

This was the approach taken by DSSSL too.

> XLink just kind of stalled.
Agreed. I think it's interesting to note that HyTime took a *long* time
to become an ISO standard, and is vastly overcomplex.  SMDL, the
standard music description language, came out of HyTime, too;
HyTime isn't just about linking, but also about time-based multimedia.
It's a shame Eliot Kimber's book was never published, I think --
the spec is inaccessible, and for those of us who worked with
(partial) implementations, the lack of good documentation was a
big drawback.  SoftQuad Panorama (a web-based SGML browser)
represented HyTime links internally as TEI Pointers.

I think I'd describe XSLT as wildly successful, albeit often in
unforseen ways.  The ability to transform XML into something you know
how to process, whether it's XHTML, HTML, text, or just different
XML, is a very important one for many people.

In particular, if you don't have editorial control over your documents,
but have to format them, and the markup wasn't really designed for
formatting, transformation is often a real win.

Generating HTML, or some other form of easy-to-style markup, is an
essential part of XSLT's purpose in life, there's nothing ironic there.

> Though I don't think XLink was the right answer, we still need a
> hypertext answer (AFs? CSS per Opera?) to get XML on the Web.

I agree that XLink isn't the end to all hypertext research.

I think XML Query will bring a new twist to hypertext in XML, since
there's a fine line between hard-coded links, runtime queries, and
even text retrieval and automatically-supplied links.


Liam Quin, W3C XML Activity Lead, liam@w3.org

====== End Forwarded Message ======

Simon St.Laurent - SSL is my TLA
http://simonstl.com may be my URI
http://monasticxml.org may be my ascetic URI
urn:oid: is another possibility altogether


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS