Re: [xml-dev] Re: determining ID-ness in XML

[Rob Lugt <roblugt@elcel.com>]

Bob Hutchinson wrote
>
> I am going to suggest that the id's listed in the <?xml-typeinfo?>
> should be unioned with the currently in-force set of ids.
>

I presume you are advocating unioning the IDs from the DTD with those
created by a <?xml-typeinfo?> PI.  A number of people have voiced concern
about what happens when the PI (or xml:id) contradicts or conflicts with the
DTD.  As Elliotte has pointed out, we only need xml:id in the absence of a
DTD, so why don't we say that xml:id (or <?xml-typeinfo?>) only takes effect
in the absence of a DTD?

Perhaps this is too simplistic, because some xml instances may well make use
of an internal subset to declare entities.  Perhaps a more refined
suggestion would be "use xml-typeinfo only when there is no ATTLIST for the
given element".  Hey, we could even add a validation warning to check that
xml-typeinfo information agrees with the DTD when xml instances are
validated.

p.s.  Even though I'm advocating <?xml-typeinfo?> as the solution, I agree
with Len that we haven't really defined the problem ;-)

~Rob