"absolute URI reference" considered awkward (and in one case, overly constraining)

Regarding the introduction of the term 'link'...
  "Architectural Principles of the World Wide Web"
  Fri, 30 Aug 2002 15:34:04 GMT

| When one resource refers to another via an absolute URI reference,
| a link is formed.

Hmm... that can be read as excluding lots of links like...
	<a href="chapter2">next chapter</a>
because chapter2 isn't absolute.

I think the simplest fix is to s/absolute//.

But there's some subtlety... the "endpoints" of the link
are absolute URI references, even though the syntax
of the reference is relative. I suppose we just explained
that a few paragraphs above in the bit about relative
URI references.

(why is that a Note, by the way?
There are some architectural principles about
relative URI references; see
Document sets and relative addressing in
 "URI Model Conseqences"
More on that separately, perhaps.

| When many resources are linked this way, the
| large-scale effect is a shared information space, addressable
| by absolute URI reference.

Boy, this is really awkward. URI elaborates
to Uniform Resource Identifier... one would think
that those are good for Identifying Resources
in the shared information space; but no,
we have to use absolute URI references; i.e.
absolute Uniform Resource Identifier references.

That comes right from the department of redundancy department.

The term 'URI reference' was introduced to handle
the case like
	<a href="chapter2">next chapter</a>
where it's clear enough that "chapter2" isn't,
all by itself, a resource identifier in the
universal context; only its full form is.

Stuart and company, are you *sure* you don't want to use
the term URI to include things like http://example/x#y?

Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Friday, 6 September 2002 12:32:17 UTC