W3C home > Mailing lists > Public > www-tag@w3.org > May 2009

RE: XML Schema usage statistics (WAS: Draft minutes of 2009-05-12 TAG weekly)

From: Larry Masinter <LMM@acm.org>
Date: Mon, 18 May 2009 18:58:37 -0700
To: <noah_mendelsohn@us.ibm.com>, <ashok.malhotra@oracle.com>, "'Tim Berners-Lee'" <timbl@w3.org>
CC: "'T.V Raman'" <raman@google.com>, <www-tag@w3.org>, "'David Ezell'" <David_E3@VERIFONE.com>, <cmsmcq@w3.org>, <holstege@mathling.com>, "'Michael Kay'" <mike@saxonica.com>, <sandygao@ca.ibm.com>, <ian@w3.org>, <shh@us.ibm.com>
Message-ID: <000001c9d825$5302ece0$f908c6a0$@org>
My experience with the schema controversy came when editing
RFC 3470, http://tools.ietf.org/html/rfc3470. This is also an IETF
"best current practice" document; as a BCP, it is the official IETF
policy for XML languages defined by IETF. The wording of section 4.7
was carefully written after a lengthy debate:

   There is ongoing discussion (and controversy) within the XML
   community on the use and applicability of various validity constraint
   mechanisms.  The choice of tool depends on the needs for
   extensibility or for a formal language and mechanism for constraining
   permissible values and validating adherence to the constraints.
   ... (See actual document for some lengthy guidelines).... 

I would suggest (from a TAG point of view) that we consider whether
W3C policy and IETF policy on XML languages should be any different,
and, if so, why? Certainly more than 5 years have passed -- should
the policy change?

I find the increasing isolation of W3C from non-W3C-developed
technologies to be disturbing. To reject XML Schema at this
point would be disruptive and harmful, but to acknowledge other
mechanisms and even (depending on resources and member interest)
sponsor development of alternatives might be useful, if they
can be introduced in a non-disruptive way. XML Schema and
Relax NG seem to coexist without difficulty except for some
redundancy of work if you are producing both.

Although we don't want to wantonly encourage alternatives and different
ways of accomplishing "the same thing", it does seem reasonable to
at least investigate having more than one alternative.

One side note: I wonder if it would be possible to define a
validity description that might be used to define HTML5 and
CSS, both important web languages that are not well-defined
by XML Schema, RelaxNG, BNF, or any other mechanism.

Received on Tuesday, 19 May 2009 02:00:14 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:28 UTC