Re: Preparing for XPointer

> * Mark Baker wrote:
> >In working on the registration of application/xhtml+xml, one of the
> >issues was how to best handle a possible transition to XPointer.
> >
> >During that thought experiment, it occurred to me that we should
> >be watching +xml registrations to ensure that, where possible,
> >"#foo" syntax is reserved for identifying fragments labelled with
> >an ID (or http://www.w3.org/2001/XMLSchema#id) type attribute of
> >value "foo".
> 
> XPointer does this, as I've said, in Section 4.2.2, "Bare Names".

You misunderstood.

My point is that if somebody goes out and registers application/bar+xml
tomorrow, and it defines "#foo" to mean something other than ID-based
fragment idenfication, then 3023 will have difficulty integrating XPointer
later because it cannot state that all +xml types will have the same
fragment identification behaviour.

Like I say, the chances are slim of this happening, but it's still,
IMO, an important enough issue to raise (assuming you like XPointer 8-).

MB
-- 
Mark Baker, CSO, Planetfred.
Ottawa, Ontario, CANADA.
[email protected]