Lists Home |
Date Index |
On Sat, 2002-02-16 at 11:19, Paul Prescod wrote:
> > However much everyone argues, URIs remain the same unchanging black
> > hole. I see a lot of people heading off to their own corner to build
> > systems which use URIs (or ignore them in favor of QNames) however they
> > see fit - and less and less chance of making the systems share a common
> > understanding.
> I agree with that but I'm surprised to hear you say it. I don't really
> understand your philosophy of standardization.
We're stuck with namespaces tacked on to XML 1.0, along with a lack of
understanding on how best to work with namespaces and URIs.
URI hocus-pocus isn't exciting to a lot of people, but they chant the
incantations because they're told they must. Given this lack of
understanding, a lot of people are writing code that totally ignores the
hocus-pocus and concentrates on QNames. The result is XML 1.0
processing that looks like XML 1.0+Namespaces processing, and its
creators may even think it _is_ namespace-aware.
I'm happy to encourage anarchic models of development where individual
programmers decide how they will process XML, even allowing for the
choice to opt out of Namespaces entirely. I'm not happy when developers
simply can't figure out what their choices are and how to implement
Anarchies of empowered developers communicating using their own
vocabularies are one (good, IMHO) thing; the chaos of developers who
can't figure out the very basic rules by which they might build and
share those vocabularies is another.
My "philosophy of standardization" seems to involve standardizing as
little as possible, but making sure that the "little" is readily
comprehensible and easily applied. Namespaces in XML doesn't do very
much, but what it does is not obvious, nor are best practices
I hope that clarification helps the conversation.
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!