Re: URIs quack like a duck

John Cowan wrote:

> Michael Mealling wrote:
>
> > > John's proposal was that if you wanted a URI to refer to the namespace
> > > you could use data:, prepended to the namespace name, which would be
> > > data:,http://www.dcarlisle.demon.co.uk here.
> >
> > I still don't understand the difference between these two proposals.
> > I guess it comes down to my own narrow view of the world.
>
> The difference is precisely in the use of relative URI references like "foo/bar".
> In my ("data scheme", "new synthesis") proposal, every reference to
> the namespace name "foo/bar" is a reference to the same namespace,
> independent of context, because the name is mapped to the URI
> "data:,foo/bar".  Accessing "foo/bar", on the other hand, produces
> an entity body that is emphatically dependent on context.

For the sake of argument only, let's combine that with the proposal that all forms
of URIs other than data: should be deprecated.  We're then in almost exactly the
same place we'd be in if we just say that URIs in xmlns attributes are uninterpreted
and just taken literally.  In other words, taking namespace names literally and
moving the burden of interpretation onto another attribute (or several) achieves
what I think folks are trying to achieve with "data:,".

By the way, I think that "data:" as a URI scheme [all too easy to confuse schemes
and schemas, alas] would be valuable no matter what happens with namespaces.   It
provides something not there now, namely, the effect of a literal in a programming
language.

Paul Abrahams

Received on Thursday, 1 June 2000 12:19:34 UTC