Lists Home |
Date Index |
On Friday 07 March 2003 06:46 pm, Daniel Veillard wrote:
> There is also (or were when I looked last year) elements allowing
> any kind of extensions (in iq, it was <x> IIRC), i.e. allowing
> to plug XML-RPC, XHTML or SMIL or anything else the sender may wish
> to propagate at that point. In a message bus kind of use that
> flexibility is critical.
Sure. You don't need namespaces to have such flexibility.
> tight in coupling, but very flexible and relatively safe (from an unicity
> point of view).
XPaths are just as safe, and even more powerful/flexible.
> > Not *the* available tool, *an* available tool... and, FWIW, we've had
> > prefixes since *way* before namespace declarations.
> But namespaces is the syntactic construct which is standardized for this.
They are not (yet) part of the XML specification itself.
The fact is that namespaces exist, and we're not likely to get rid of them
soon. I think people should examine their real needs before relying on
them... there are alternatives, and we have a chance to minimize.
In all of my years working with XML/SGML, I've seldom needed anything like
namespaces, and when I did, a prefix or attribute was enough. Namespaces do
complicate software... the DOM being a good case in point. I have *still* to
see a good reason why everyone should pay for that. I have heard many people
assert that you need namespaces in order to mix and match tag vocabularies,
but the fact is that you don't often do that in an unconstrained manner, and
even when you do, you don't need namespaces.