Re: Dependency of XPointers upon DTD

Hello again Eric-- We'll certainly list this as an issue and discuss 
it.  FWIW, here are a few reactions (speaking only for myself, as a busy 
editor)...

At 04:00 PM 9/1/00 +0200, Eric van der Vlist wrote:
>I'd like to point out the level of dependency of the current XPointers
>CR on DTDs which IMHO should be neutral to the schema system used.

I'm not sure this is appropriate right now.  DTDs are already a 
well-established and well-supported part of a Recommendation (the XML 
spec), whereas XML Schema is not yet standardized.  I'm reluctant to add 
examples -- even non-normative ones -- that could be obsoleted by syntactic 
changes to XML Schema in the future.

>The XPointers CR makes an intensive (non normative) usage of DTDs as
>examples of a way to simplify the authoring of documents using XLink
>attributes through implied values.
>
>Couldn't some examples using XML Schemas be provided as well to show
>that alternative usages are possible ?

Actually, some of us are working on a writeup that shows how the future of 
XLink might go with XML Schemas in the picture, so this aspect hasn't been 
ignored.  I don't know quite when this will be made public, though.

>There is another more pernicious usage of DTDs.

I'm not sure "pernicious" is a fair word to use...  In what sense is the 
usage "high injurious or destructive"?  (www.m-w.com)

>My understanding is that the "bare names" scheme [1] relies on the XPath
>unique IDs [2] which themselves relies on DTD's ID typed attributes.
>
>This scheme, being the easiest to write and more robust scheme will
>probably be widely used and I wonder if it's not dangerous to rely on
>DTDs which present well know limitations especially with the namespaces
>which usage is mandatory with XLink.

DTDs are not being relied on; there are always other means to point to a 
particular element if you wish, and you don't need to default your 
attribute values.  And an internal subset could be used in a published 
document to "mark" an attribute as being of type ID even if the document is 
only well-formed; an XPointer into a document without such a "marker" 
wouldn't want to rely on the fact that the attribute is typed.

I agree that the namespace aspect is somewhat problematic (what if the 
attribute is "global" and has a namespace prefix on it? what about the 
element it's on?).  However, the problem here would be one of 
adding/maintaining the "markers" in the target document, not changing an 
XPointer into that document (if the XPointer is written correctly).

When XML Schema is stable and standardized, a later version of XPointer 
(and, I assume, the XPath spec on which it is based) could take advantage 
of schemas and the post-schema-validation infoset to detect which elements 
have which IDs.

>This scheme relies on a unique ID mechanism which is also implemented in
>XML Schemas [3].
>
>Is it possible to add amongst the XPointers extensions to XPath that a
>XML Schemas ID should play the same role ?

Until XML Schema is a Recommendation, as I mentioned above, I think it 
would be complicated to add features that rely on it.

         Eve
--
Eve Maler                                          +1 781 442 3190
Sun Microsystems XML Technology Center    eve.maler @ east.sun.com

Received on Wednesday, 6 September 2000 14:01:20 UTC