xml-dev - Re: XML Schemas: Needs Marketing?

["Simon St.Laurent" <simonstl@simonstl.com>]

At 04:47 PM 2/16/00 +0000, Edd Dumbill wrote:
>Experience indicates that vendors may well implement as much of a
>specification as their customers insist upon. It is therefore in
>the interests of W3C et al. to make as much as possible of their
>specifications implementable. If they don't provide for 80/20 -- and
>in doing so preserve some sympathy with the "100%" aims of the specs
>-- then API vendors will probably make the 80/20 cut themselves, not
>necessarily with the same sympathy for the integrity of the specification.

This is where customers lose out on interoperability, and XML's reputation
starts tarnishing rapidly.  The XML 1.0 spec already included some gotchas
that are quite legal - like non-validating parsers not loading external
resources - but which can make interoperability tougher.  Add the rest of
the specifications to this list, and even rose-colored glasses start seeing
a darker world.

>I occasionally wonder whether the current model of standards development
>is sustainable given the growth in the XML marketplace.  There is a real
>prospect of a "proprietary standard" achieving sufficient mindshare before
the
>W3C finishes the process of a developing the "official" one.  Perhaps
>different models of development are required that are more responsive to
>the commercial environment. Layers, as Simon suggested, are one
>alternative. Don Park also suggested multiple, smaller, groups and
>previously, if I remember correctly, small "atomic" specs.  Food for
>thought, anyway. I wonder if there have been parallel examples in other
>areas we might learn from here?

I've argued a few times for modular standards, whether they be layered or
composed of smaller independent parts.  Watching the progress on XHTML 1.1,
where they're more or less tearing down HTML to rebuild it, has been very
interesting. XHTML is really only a vocabulary, so the issues are simpler,
but they're still very complex.  A core piece, the rules for negotiating
which modules can be used for various transmissions, is still missing, and
I think that's going to be a tough bit to sort out cleanly.

I'd like to think that there's a core of XML that's reliably interoperable
and on which we can build/rebuild the rest.  I'm not talking about starting
over, but I'm definitely talking about cleaning up and clarifying what we
have.

Simon St.Laurent
XML Elements of Style / XML: A Primer, 2nd Ed.
Building XML Applications
Inside XML DTDs: Scientific and Technical
Cookies / Sharing Bandwidth
http://www.simonstl.com

***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/threads.html
***************************************************************************