Lists Home |
Date Index |
Mike Champion wrote:
> Let me take this opportunity to make my standard namespace rant: The change
> suggested in the proposed v 1.1 only smooths over the tiniest tip of the namespaces
> iceberg. The part that takes its toll is not the common syntax, but the different
> namespace data models -- DOM, XPath/XSLT 1, XPath/XSLT 2, Canonical XML, and
> probably more. Are prefixes significant, i.e., should the InfoSet retain any
> memory of the prefix used to map to a namespace, or is this just syntax sugar? Are
> namespace declarations attributes? Is namespace information associated with each
> node, or does the scope of the namespace implicitly define some region in the
> document? And what is supposed to happen when nodes (namespace declarations or data
> nodes) move around?
one can sort this out by answering a fundamental question: are names
constants or variables? despite the variation among and discrepancies
between the respective models, the standards describe operations which
align, in that they treat names as constants. given that, the models can
be unified to describe names as constants and variations can be handled
when parsing and/or serializing.
> There's no way to sort this out without breaking lots of code and making people
> unhappy, but unless XML disentangles itself from this mess, its forward progress
> will be significantly hindered.
it may well be that a significant effort is necessary to fix this in the
java world. judicious class specialization could reduce it. better late