Namespaces and validation -- missing information?

I'm trying to figure out what the intent is for how Namespaces and validating
parsers interact, since it will affect my arguments in the Document Object Model
discussion. But I just can't seem to put it all together.


The REC says namespaces change the DTD grammer to use qualified (prefixed)
names. In other words, the DTD is now able to actively say that an element or
attribute with a specific prefix is expected at a given point in the document.

Architecturally, I presume you don't mean that to be taken literally, but
instead that the prefix should be read as "an element with any prefix that maps
into the same namespace as this prefix." That's a serious change to parser
validation engines, but if it's what you intend, fine.


Unfortunately, I can't see anything that says how to map prefixes to namespaces
inside the DTD.  Best guess I've got is that it should be a declared #IMPLIED
attribute on that element, with an appropriate default value ...?

It would help a great deal if the Namespace spec had a FAQ that provided a
_complete_, theoretically validatable, illustration of a namespace-using
document and its DTD, demonstrating all the features of how namespaces are
declared used, and reconciled on both sides. This might also be important input
to the schema working group, since they're going to have to figure out how to
work with the new behaviors you've created.


Many thanks for any advice you can offer. If there's a document that does answer
these questions already available on the web, feel free to point me to it.

______________________________________
Joe Kesselman  / IBM Research
Unless stated otherwise, all opinions are solely those of the author.

Received on Friday, 19 March 1999 16:23:58 UTC