xml-dev - Re: More on Namespaces (also long, but also optimistic)

["David G. Durand" <dgd@cs.bu.edu>]

At 6:53 PM -0400 8/16/98, james anderson wrote:
>Peter Murray-Rust wrote:
>> ><?namespace prefix='VanityHG' ns="http://VanityHouseAndGarden.org.uk" ?>
>>
>> I'm lost with these PIs. I thought they had disappeared....
>>
>
>Ah, but since the standard no asserts an interpretation for them which
>precludes validation, a processor is free to use them in a way which
>enables validation.

I am assuming you meant to type "now asserts an interpretation..."

But this isn't true. It provides a facility which can be used in a
validation compatible way (just as the original facility), and which can
also be used in a validation-incompatible way, just like the earlier one.
In both cases you need to know what's in the instance to create a DTD which
will make an instance valid  -- that's an ineveitable side-effect of mixing
"namespaces" with out a namespace-aware validator. Interestingly, this
gives you no worse validation than architectural forms do, although AF
validation was one argument againt AFs and _for_ namespaces...  Anyway
that's water under the bridge.

The thing that the current proposal gives you if you are _not_ validating
-- and I take the NS people at their word when they claim that they think
they can live without validation -- is the ability to limit the scope of a
namespace prefix declaration, and thus, limit the region of the document
within which that prefix cannot be used to declare another namespace.

Since a prefix is essentially a variable, bound to a URI, and used to
differentiate like-named elements, this kind of scoping mechanism is not
unsensible.

The contextual dependency implied by the element-nesting of declarations
may create entity re-use problems, but with care those problems can be
avoided. Many people argue that there are other problems with entity reuse
(David Megginson and Eliot Kimber are two worthy adherents to this belief).

>Take them, for the sake of discussion, to establish a prefix/URI binding for
>which the scope is the remainder of the logical entity in which they appear.

To the extent that you are proposing a separate namespace standard, I hope
that you are unsuccessful. It's a hard thing sometimes to buckle under to
the fact that much of the time _any_ standard that is accepted is in fact
better than a custom solution that is not accepted.

Namespaces per se are unnecessary, since Architectural Forms can do the
same thing, but since we have a decent namespace standard, we should use it
where it makes sense.

  -- David

_________________________________________
David Durand              dgd@cs.bu.edu  \  david@dynamicDiagrams.com
Boston University Computer Science        \  Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/   \  Dynamic Diagrams
--------------------------------------------\  http://www.dynamicDiagrams.com/
MAPA: mapping for the WWW                    \__________________________



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo@ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)