Re: Another updated xml:id specification

Norman Walsh scripsit:

> I've updated the xml:id specification per our discussions at the XML
> Core WG telcon today:
> 
>    http://www.w3.org/XML/Group/2004/xmlcore/xmlid/xml-id.html

Thanks for the quick turnaround, Norm.  A few editorial notes:

1) The must-within-must and must-within-should wording in section 4
   is confusing.  I suggest the following:

An xml:id processor must assure that the following statements hold for
all xml:id attributes:

    * The normalized value of each xml:id attribute is a valid NCName.
    * If the type of an xml:id attribute is specified using a validation
      technology, the type is 'ID'.

An xml:id processor should assure that the following statement holds
for xml:id attributes:

    * All values of type )B�ID� within a document are unique.  

2) The "Strictly speaking" in the definition of validation is a hangover from
   earlier drafts, and the definition of validation isn't actually marked up as
   a definition.

-- 
He played King Lear as though           John Cowan <jcowan@reutershealth.com>
someone had played the ace.             http://www.ccil.org/~cowan
        --Eugene Field                  http://www.reutershealth.com

Received on Wednesday, 27 October 2004 17:23:05 UTC