RE: Exclusive C14N Error: Using #default in NMTOKENS

Hi Joseph,

The prime motivation is that two XForms editors independently said they were curious why the DSig group did not use the more formal specification.  Since it is possible to precisely say what we mean, I had no technical answer.  But if the answer is consistency with XSLT's definition, then I would be fine with that.

A secondary motivation is that, while I don't see a really good reason for being more precise, a fairly lengthy amount of implementation experience has shown me that the decision to be less precise often has a way of coming back to provide some really painful experiences.  Still, in this case, the fact that we are somewhat constrained by people who have only DTD and not schema provides some reasonable justification for not going that extra mile.  Consistency would be another reason (i.e. if such a level of precision does not match what was done everywhere else in the DSig related products, then it's probably not a good idea to introduce that in this one case).

John Boyer, Ph.D.
Senior Product Architect
PureEdge Solutions Inc.
 

-----Original Message-----
From: Joseph Reagle [mailto:reagle@w3.org]
Sent: Friday, December 06, 2002 9:06 AM
To: John Boyer; Gregor Karlinger
Cc: w3c-ietf-xmldsig@w3.org
Subject: Re: Exclusive C14N Error: Using #default in NMTOKENS


On Friday 06 December 2002 11:32 am, John Boyer wrote:
> Yes, I agree too.  The DTD says CDATA, but should the schema be
> xsd:string, or should it be the longer and more precise content
> specifier I posted yesterday?

What is the motivation for the longer/precise defintion? My preference is 
not to create a new data type in the errata, but just rely upon string 
since that what schema and xslt seem to do for their "#foo" tokens.

Received on Friday, 6 December 2002 12:44:00 UTC