W3C home > Mailing lists > Public > xml-editor@w3.org > July to September 2002

RE: [xml-dev] The version number in XML documents

From: Michael Kay <michael.h.kay@ntlworld.com>
Date: Mon, 22 Jul 2002 10:07:11 +0100
To: "'John Cowan'" <jcowan@reutershealth.com>
Cc: <xml-dev@lists.xml.org>, <xml-editor@w3.org>
Message-ID: <002b01c2315f$29f918c0$6401a8c0@pcukmka>

> 
> > You mean you want to abuse the errata process to make a 
> retrospective 
> > change to the spec that is not actually an erratum.
> 
> Do you have a definition of "erratum"?  We do not make errata 
> that involve changes to the definition of well-formedness.  
> This one is marginal because <?xml version="bluberry"?> is 
> allowed to generate a fatal error as if it were well-formed.

The erratum process was invented to allow factual and typographic errors
to be corrected, and ambiguities to be clarified. It was not invented in
order to allow W3C to change its mind 4-and-a-half years after releasing
a specification. 

The process should not be used in a way that makes parsers that are
unambiguously conformant with the original specification suddenly become
non-conformant, and that forces new releases of those parsers to reject
documents that they previously accepted. 

The specification as originally approved clearly allows different
parsers to make different choices here. You might wish it had been
written differently, but trying to fix it now would be a worse error, if
only because of the damage to W3C's credibility as a supplier of stable
specifications.

Michael Kay
Software AG
home: Michael.H.Kay@ntlworld.com
work: Michael.Kay@softwareag.com 
Received on Monday, 22 July 2002 05:05:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:32 GMT