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


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: How to redesign W3C XML Schema (Was: Remembering the original XML vi

[ Lists Home | Date Index | Thread Index ]
  • To: xml-dev@lists.xml.org
  • Subject: Re: How to redesign W3C XML Schema (Was: Remembering the original XML vision)
  • From: Jochen Wiedmann <joe@ispsoft.de>
  • Date: Mon, 17 Feb 2003 09:27:50 +0100 (CET)
  • In-reply-to: <1045441502.7668.ezmlm@lists.xml.org>
  • References: <1045441502.7668.ezmlm@lists.xml.org>
  • User-agent: IMP/PHP IMAP webmail program 2.2.6

Quoting Jonathan Robie:

> But I would only support it if it had a prayer of being implemented and 
> used. Remember that XML was a marketing coup as much as a technical coup - 
> if we wanted to establish another schema language, we would have to find 
> markets that desperately need it, and for whom existing solutions are 
> painful enough to make them willing to change, and to abandon already 
> accepted standards. Building a market like that takes time and energy.

I understand the issue. However, much that has been said in this
thread about SGML reminds me very much about my current experiences
with XML Schema.

IMO a good specification is a small specification. For example, the
old RFC's have typically been *very* small, compared to nowadays
nightmares. Standards like SMTP have of course evolved in time,
adding features over time. As David wrote:

> I spent three months on and off trying to write an SGML parser in
> Java and never made much progress; I had the first working draft
> of AElfred done in an evening, and a quite usable parser in a couple
> of days (on top my regular work responsibilities at the time)

I think the same can be said for the early versions of SMTP, POP3,
HTTP, and all that stuff. (At least for the protocol, of course it
is a different task to write an SMTP protocol handler than writing
Sendmail or MS Exchange.)

IMO this can be said partially for XML (at least not for the DTD
related stuff), but not at all for XML Schema. I am currently working
since months on an implementation of JAXB (http://java.sun.com), which
is a fairly big specification in itself. But I can realistically *never*
expect to implement all the details of XML Schema, on which JAXB is
sitting, not even at the parser level.

I am not so keen on the idea of supporting Relax NG, Schematron, and
all that stuff. But a simplification of XML Schema would make quite
some sense, much in the way of DocBook with its spinoff DocBook Lite.



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

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