If you read through the specification you'll notice that the assignment of IDs is conceptually done by a component separate from the parser called an xml:id processor [1]. Xerces has never provided an implementation of such a processor, so the answer to your question is no it won't work unless you've declared xml:id in a DTD. This isn't a bug. It's just a spec we've never implemented and have never claimed to support.
[1] http://www.w3.org/TR/xml-id/#dt-xml-id-proc
"An xml:id processor is a software module that works in conjunction with an XML processor (i.e. parser) to provide access to the IDs in an XML document."
Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: mrglavas@ca.ibm.com
E-mail: mrglavas@apache.org
Mike Sokolov <sokolov@ifactory.com> wrote on 09/16/2009 02:50:16 PM:
> Yes - I was hoping you would pick up this thread. So, I'm not sure I
> understood your response. I guess my basic question is whether it is
> feasible to use xpointer="element(<id-value>)" with Xerces (as it seems
> from the FAQ), and if so, what needs to be in place for that to work. I
> understand the ids needs to be recognized as such, but I'm a little
> unclear on how to make that happen.
>
> I did think (naively, perhaps), that xml:id was just automatically
> recognized as having type "ID" by virtue of its special nature.
>
> Then I thought perhaps it is necessary to validate (with a DTD) in order
> to get that fact identified (I think this seems to be what you're
> saying), so I enabled validation. But my xpointer references are still
> failing. If you can confirm that this *should* work, then I will assume
> the parser isn't actually validating as I thought.
>
> -Mike
>
> Michael Glavassevich wrote:
> >
> > john.l.clark@gmail.com wrote on 09/16/2009 02:20:49 PM:
> >
> > > Mike,
> > >
> > > On Wed, Sep 16, 2009 at 2:04 PM, Mike Sokolov <sokolov@ifactory.com>
> > wrote:
> > > > Question: is xml:id considered a "DTD-determined ID" according to
> > the XProc
> > > > spec? I read it, and found this sentence:
> > > >
> > > > [Definition: An attribute information item is a DTD-determined ID
> > if and
> > > > only if it has a [type definition] property whose value is equal
> > to ID.]
> > > >
> > > > I think that's true of xml:id, but would like it if someone more
> > > > knowledgeable could confirm or deny
> > >
> > > It sure is. See <http://www.w3.org/TR/xml-id/#dt-id-assignment>,
> > > which I think is pretty clear on that.
> >
> > ... only if an xml:id processor did that assignment. The attribute
> > itself doesn't magically have that property.
> >
> > > Take care,
> > >
> > > John L. Clark
> > >
> > > _______________________________________________________________________
> > >
> > > XML-DEV is a publicly archived, unmoderated list hosted by OASIS
> > > to support XML implementation and development. To minimize
> > > spam in the archives, you must subscribe before posting.
> > >
> > > [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> > > Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
> > > subscribe: xml-dev-subscribe@lists.xml.org
> > > List archive: http://lists.xml.org/archives/xml-dev/
> > > List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
> >
> > Michael Glavassevich
> > XML Parser Development
> > IBM Toronto Lab
> > E-mail: mrglavas@ca.ibm.com
> > E-mail: mrglavas@apache.org