Re: [xml-dev] IDs without DTD/Schema , Is there a way ?

[Daniel Veillard <veillard@redhat.com>]

On Mon, Oct 29, 2001 at 11:37:05AM -0400, Elliotte Rusty Harold wrote:
> At 10:18 AM -0500 10/29/01, Daniel Veillard wrote:
> Yes, but the problem is that foo can't have an ID attribute if the document does not have a DTD.

  Hum, sounds fairly strongly worded. If an XML Schemas can define
IDs, IMHO the XML specification itself could also do this.

> No, if the document does not have a DTD, then no attribute can have type ID; i.e. there are no ID type attributes without a DTD.

  Schemas ???

> I don't think we need to change XML 1.0 to make this happen.
> Tim Bray's xml:id or ID namespace would allow us to define a second
> space of unique element identifiers encoded in attributes that are
> separate from the normal ID type attributes of the document. These
> attributes might even have type CDATA in documents without DTDs, just
> like all other attributes in such documents. Nonetheless, they could
> have all the characteristics we need for reliable, name-based linking.

  My point of view is:
  - XML Core group handling the XML spec is alive
  - XML Linking WG is not active
  
  - making my parser detect xml:id as an ID even in the absence
    of a DTD, is a 5 line patch, relatively clean
  - making yet another new class of attribute "pseudo" type for
    the intent of linking only, requires framework changes, and would
    be completely limited in scope, i.e. an architectural bad hack.

 From my point of view, I don't see your suggestion being:
   - easier to process
   - more likely to get accepted
   - more likely to get implemented

 than a proposal based on an extension of ID at the XML specification
level. It's just my point of view...

Daniel

-- 
Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
veillard@redhat.com  | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/