Fwd: Re: namespace prefixes within attribute values

I also sent this to XML-Dev; I think c14n may have to reexamine its
discarding namespace prefixes in the light of other W3C activity (schemas
and XLink) that need that information to process attribute values using
those prefixes correctly.

I'm aware that we're past last call, but hopefully this is at least worth
addressing.

>Date: Thu, 09 Mar 2000 22:26:41 -0500
>To: Tim Bray <tbray@textuality.com>, Ray Whitmer <ray@xmission.com>
>From: "Simon St.Laurent" <simonstl@simonstl.com>
>Subject: Re: namespace prefixes within attribute values
>Cc: XML-Dev Mailing list <xml-dev@xml.org>
>
>At 04:17 PM 3/9/00 -0800, Tim Bray wrote:
>>At 04:52 PM 3/9/00 -0700, Ray Whitmer wrote:
>>>Perhaps I missed something, but would this type of value become shredded
if 
>>>you save canonical XML, which reassigns all the prefixes but presumably 
>>>does not know about namespaces in values represented by prefixes?
>>>
>>>Is that a problem?
>>
>>Yes... in fact, if you were only thinking about the namespace spec, you
ought
>>to feel free to rewrite namespace attributes any old time you want. For
>>example in canonical XML.  Ones that are hidden away in attribute values
>>won't get rewritten and hence will break.  One of the reasons I was against
>>this.  However, schemas, if there's a QName type, will help. -T.
>
>Hmmm... but aren't schemas one of those core document types where
canonicalization and signature processing would be very very helpful?  I'm
thinking back to last week's "When XML gets ugly" keynote by David Megginson.
>
>It sounds like the c14n spec (2.15) needs to include some kind of
information about prefixes, or we won't be able to use it with either
Schemas or documents that use XLink's external linkbase capabilities.
>
>To me, that's a real loss, never mind the philosophical issues.
> 
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

Received on Thursday, 9 March 2000 22:37:33 UTC