xml-dev - Re: W3C, responsibility (Re: Why the Infoset?)

[Rick JELLIFFE <ricko@geotempo.com>]

"Winchel 'Todd' Vincent, III" wrote:
 
> Some of the W3C's big lies (bad behavior) follow:
> 
> 1. XML is backwards compatible with SGML
> 
> DTDs are no longer supported by the W3C, in favor of XML-Schemas.
> XML-Schemas are not backwards compatible with SGML.  So, as a practial
> matter, the statement "XML is compatible with SGML" really isn't true,
> although this was the W3C promise in 1998 and much of 1999.

Incorrect. SGML allows documents with no DTDs.  See ISO 8879 Annex K.

> 2. XML is nearly as powerful and a lot simpler than SGML
> 
> This was the promise and the marketing hype.  We've all read the following .
> . . SGML has been around for a long time, but has always been too
> complicated for the masses.  HTML is very simple and, therefore, gained
> wide-spread acceptance among the masses.  XML, because it simplifies SGML,
> yet is technically superior to HTML, is the next step in the development of
> the web.
> 
> As it turns out . . . we have . . . Namespaces, XML-Schemas, RDF, XLink,
> XPointer, XPath, XSL, XSLT, XSLT-FO, XML Query, XML-Signatures, Canonical
> XML, SVG, Infoset, etc. . . . whew! . . . and all of these are consistent,
> by the way, and work well together, and the W3C has done a great job to
> explain how they all fit together in harmony . . . hmmmm . . . where is the
> simplicity??

Incorrect.  You have switched from "XML" the parser language to "XML"
the family 
of technologies mid-thought. XML the language is simpler than SGML: it
is a
profile. XML the family of technologies will of course become more
complex 
than SGML alone (even SGML extened by the Extended Facilities and
General 
Architecture and HyTime and DSSSL and RAST).
 
W3C has never made any representation that
XML+Namespaces+XML-Schemas+RDF+
XLink+Xpointer+SVG+Infoset is simpler than SGML alone.

> 3. Namespaces
> 
> Consensus on this list is they they don't work.  Certainly, DTD validation
> and Namespaces don't work.  Nowhere is this stated in the W3C Namespace
> Recommendation.

Incorrect.  DTD validation does work with namespaces. Namespaces are a
layer interpreted after validation. That is clear from the namespaces
spec. It is not
the consensus on this list that namespaces don't work: I use them every
day
and they work fine for my use.  (The namespaces spec is underspecified
as far
as URIs go, and does not give us the semantic web that some people think
it 
should, and I think some of the non-normative appendixes are botched,
but
none of those things mean it does not work.)
 
> 4. XML-Schemas (and the tools) are right around the corner
> 
> Good to know, since XML-Schemas purportedly solve all problems and DTDs are
> no longer supported.
> 
> Microsoft gets flamed, among other things, for such behavior -- so should
> the W3C.

Incorrect. There are several draft tools for XML Schemas. These are
tracking the 
draft: it is a legitimate decision for a developer to skip releases that
track drafts,
in the interests of giving their uses tools that are stable for longer
than
a month.  The decisions about XML Schemas development and draft release
dates are up
to the WG--we are not going to release such an important spec unless it
is correct.

Where is a W3C spec  or press release that promised XML Schemas by a
certain date or promises it will solve all problems?  However, I
certainly concede that somehow from somewhere people have the
expectation that it will be a universal typing or constraining framework
that will meet all their needs: I don't think W3C or the stakeholder
companies have been remotely vigourous to dispell this. In part, of
course, it is difficult to say "XML Schemas won't do such and such" when
so many things in schema languages are judgement calls with tradeoffs
that are altered by flow-on effects.  And don't forget that the Schema
WG are only human: we need to keep a positive attitude about what we are
doing, and any encouragement from the general community is appreciated.

I know of more than 5 tools under development.   I comment negatively
more
about XML Schemas than anyone: I think I am entitled to because I also
comment
positively on it more than anyone, and because I have an alternative
schema
language that I think is pretty good and I want to encourage discussion
on
the merits of particular features (which will help the Schema WG), and
because
I am a member of the W3C Schema WG and so I want to get responses that
will
help remove groupthink and to help me represent non-corporate interests 
on the Schema WG faithfully, to the best of my ability (I am on the
other side
of the world, so my ability to network personally is 0). 

Let me defend W3C in another way. Of course W3C is mostly paid by big
business:
they need a neutral forum where they can bash out technical solutions so
that
don't cut off their noses despite their faces.  The schemas and queries
groups
are two which have big stakes riding on them: they affect the core
implementations
of database products. But the existence of the W3C also allows the
companies to
agree on approaches to internationalization and disability access that
otherwise
there might not be.  Of course, there is a commercial or legal interest
at play
here too. But W3C has been exemplorary in its support for
internationalization and
accessability.  Look at IETF now: they have not supported
internationalization 
concerns adequately in DNS and the whole thing is falling apart as Asian
countries
are going their incompatible way!  I am an invited expert to the W3C
internationalization group, so I see the things they talk about, and
their
concern for much more than individual corporate goals.

The W3C is not the enemy. Tim Berners-Lee is not the enemy. Namespaces
are not
the enemy. 

If you don't like the direction of XML technology, are W3C trying to
stop anyone making alternatives?  I have had nothing from them against
Schematron (which is not just a competitive schema language but a
competetive schema paradigm): in fact, Dan Connolly (a long-time W3C
insider) recently used it to put up a test online validator for web
accessability.  When I wrote yesterday about ".org fronts" for large
companies, if one considers, say, biztalk.org and w3c.org, one is
probably more connected to big business than the other. I hope I haven't
contributed to anti-W3C bigotry by my comments; that would disappoint
me.

There is so much useful software and ideas that we could be spending our
time creating: even adding comments to free code so that non-native
English speakers can understand it could result in the world being
better for programmers in third-world countries. Lets not waste our
presious emotional energy.  W3C will put out specs: some specs we will
like or dislike; some specs they will do a bad job at; some specs could
be bettered by an individual under inspiration; some specs will be
skewed to the commercial positions occupied by the companies who pay the
people to create the specs. 

(Please, no accusations of censorship or Lee Kuan Yu-ery. :-)  I suppose
from a kind of Marxist-Leninist position, W3C could be seen as a sign
that industry is reaching the stage of oligarchy capitalism.)

Rick Jelliffe