Lists Home |
Date Index |
Tim Bray wrote:
> At 04:11 PM 04/04/02 -0800, Ronald Bourret wrote:
> > Namespaces were (and still are) a good idea. Using URLs as example
> > URIs in the namespaces spec was a bad idea. It led people to think
> > about dereferencing namespace names which, if you think about it,
> > is about as odd as dereferencing Java package names. Sure, you can
> > come up with a mechanism to do it, but why would you want to?
> It would be nice if there a canonical place to find the
> javadoc output. In fact it would be *very* nice. -Tim
I agree, but I'm not sure that directly dereferencing the package name
is the best way to do it. This smacks of storing business logic in
database keys, which tends to be fragile.
(I suppose at some level, any lookup involving an identifier can be
thought of as "dereferencing". To me, it's just a question of whether we
build our identifiers around a certain dereferencing scheme (e.g. URLs)
or build a dereferencing scheme around our identifiers. The former can
exploit existing infrastructure but, carried to an extreme, is absurd.
Should we make all programming variables URLs? The latter places the
identifiers in the center of things, but means we need to build