document vs non-document entity (was Re: CORBA's not boring yet. / XML in an OS?)

[james anderson]

This raises the question: "What is the extent of an entity binding?"

Assume that is possible to qualify entity names. Since these qualifications
are already lexically scoped, why would one need to introduce local entity
definitions? The effective entity is already determined by a binding with a
lexical scope.

One reason could be that the form actually makes explicit a dynamic extent for
"local" entity definitions and is set in contrast to an indefinite extent for
"global" definitions within the dtd (or elsewhere). I have gathered from
discussions, that contemporary implementation techniques already enforce a
dynamic extent for entity definitions - they "live" exactly as long as the
document. Unless this changes, then local entities are not strictly necessary.
With an XML/OS there may well be cause to change the binding semantics.

James Tauber wrote:
> 
> Paul Prescod:
> >This is a serious problem inherited from SGML. It is my opinion that
> >instead of namespace declarations being done through attributes they
> >should have been just another form of PI-based declaration. Then in
> >version 2.0 of XML, all declarations should have been made "localizable."

I don't understand the connection here. Where is the dependancy?

> >In particular, there should be a way to declare internal text entities,
> >unparsed entities and attribute defaults locally.
> 
> Well this is actually where my thinking has been going. I've been thinking
> about whether localizable declarations might be achievable using an
> *attribute* mechanism following on from how namespaces ended up.
> 
> ...
> 
> The xmldecl attribute takes the value of the URI of an external parameter
> entity.
> 
> To avoid name clashes, it might be an idea to have
> xmldecl:foo="local-decl.pen" and then qualify entities with the prefix foo
> (ie &foo:SomeEntity;)

While I agree that this will ultimately prove necessary, I suggest that
qualified entity names render local entity bindings redundant. It is
equivalent to mapping each of the definition names into a distinct namespace
at the point of definition and using lexically bound prefixes to map the
reference name accordingly.

Opps! Yea, I forgot, there is no way to bind a prefix in the DTD (ie "the
point of definition"). Got to wait for schemas for this.


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)