Re: [xml-dev] IDs considered harmful or why keys might be better thanIDs...

["Jonathan Borden" <jborden@mediaone.net>]

David Carlisle wrote:

>... The reasons why the ID attributes which are in the DTD
> need to be redeclared in an internal subset, and in fact what an
> internal subset is, are things that one shouldn't ever have to know
...

Why would one need to redeclare an ID in the internal subset which is
already declared in the external DTD. Its already an ID right?

Again, my point about using internal declarations of IDs is for documents
which don't have an external DTD. If you are using an external DTD you can
always declare a short DTD driver which sucks in docbook+mathml + whatever
and defines the new IDs your particular application desires. The point being
that if you are declaring an ID _and_ going to the trouble to validate in
the first place, then you probably want to be sure the ID is really an ID
... i.e. that the validating software (e.g. validating XML parser) processes
it. Having _two_ mechanisms to define IDs by disparate techniques would mean
rewriting every validating XML parser to understand that _xml:id_ or
_xml:idatts_ or whatever means.

Since we are going to have to deploy _some_ new type of software which
understands and implements _xml:id_ or whatever, and perhaps rewrite _every_
XSLT and DOM implementation so that it understands the new behavior of id()
and getElementById() (!!!) it would seem far easier to fix XSLT so that it
can properly handle internal subsets (i.e. write them as well as read them).

Jonathan