[
Lists Home |
Date Index |
Thread Index
]
On Friday 07 March 2003 08:16 pm, John Cowan wrote:
> Rich Salz scripsit:
> > Has anyone come up with an alternate proposal that allows you to mix
> > and match a document with elements that come from distributed creators?
The answer here is "yes". There have been a largish number of proposals
BUT what *precisely* do you want to do? What is the *precise* problem you're
trying to solve?
Is it
a) To mix and match elements from arbitrary markup vocabularies
b) To mix and match elements from arbitrary markup vocabularies with
no prior agreement.
c) To mix and match elements from arbitrary markup vocabularies with
no prior agreement, and to then *process* them.
or something else?
From what I have seen, mixing and matching arbitrary vocabularies without some
form of prior agreement, and then expecting to process them meaningfully, is
rare. In most cases, the set of tag vocabularies is tightly consrained, *or*
to most processors, well-defined subtrees are opaque to the outermost
receiving processor, which *may* have some means of dispatch to arbitrary
code (and the definition of that binding is at this point entirely
application-specific).
The point here is that namespaces solve few (if any) real problems, and
introduce more problems in return.
Even when namespaces were being standardized, a lot of us fought them as they
are currently defined, and even then, I *never* saw a problem where they
offered a clear benefit over things like simple prefixes, or attributes. I
still haven't, and close to 5 years have past.
|
