coming clean with the SGML crowd (was re: namespaces)

David Megginson david at
Fri Sep 11 13:49:43 BST 1998

Richard L. Goerwitz III writes:

 > XML right now is, well, only slightly more than a dream.  I have a
 > hard time even finding valid XML document instances on the net.
 > Its APIs are almost all done in Java (itself a moving target which,
 > despite the hype, has precious few major applications to its
 > credit).  And its supporters have fallen to squabbling.

PRONOUN REFERENCE ALERT!  Who's squabbling -- Java's supporters or
XML's supporters?  Not that we ever squabbled about SGML...

Richard is certainly right about the lack of valid (or even just
well-formed) XML document instances on the net.  You'll find the XML
spec itself, Jon's religious and Shakespeare texts (which need
fixing), the XML Heart of Darkness on my personal (Sprynet) web site,
a few converted Sun docs, etc. -- certainly nothing to get excited
about.  If it weren't for the TEI, however, SGML wouldn't be doing
much better than XML on the Web, outside of HTML itself.  Only a tiny
fraction of user systems have the technology to view rendered versions
of XML *or* general SGML documents, so it is not surprising that
people aren't bothering to publish directly in SGML or XML.

As for Java, the comparison might not be fair -- Java is very heavily
used as middleware and on the server side (the most likely place for
XML to be deployed), and the core non-graphical APIs for Java are
relatively stable since 1.1 (modulo a few new classes in java.util).
Furthermore, nearly any user who bothers to go on the Internet
encounters several Java applets/session.

When the Java hype started, people could think of Java only in
conventional terms: vendors would write giant applications that users
would install and run on their systems, just as they did using C++.
Of course, you can still do that if you really want to, but it turns
out that instead of competing directly with C++ (etc.), Java helped
bring along along a new, distributed information model, where users
see small applets on web pages, and servers plug in small servlets to
do the work.

When the XML hype started, people could think of XML only in
conventional terms: authors would write XML documents that users would
view rendered in browsers, just as they did using HTML.  Of course,
you can still work towards that if you really want to...

 > I don't really know what to develop for right now, what with
 > confusion over our design goals, about namespaces, about XSL, SGML
 > compatibility, and virtually everything else.  Perhaps if we could
 > all agree on a few basics, we could set- tle things down.

Develop for XML 1.0.  If you don't need namespaces, don't bother with
them, or at least wait until the bigger parts of the market decide
whether they'll use them.

XML found an unoccupied niche and moved into it.  Now, XML is breeding
and producing dozens of derived and related specs, like tadpoles --
since food is scarce and predators are fierce, most of the offspring
will die immediately, and most of the remainder will not make it to
adulthood.  This is a standard situation, both in nature and in
standards bodies.

All the best,


David Megginson                 david at

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list