Re: URI versus URI Reference

John Cowan wrote:

> "Paul W. Abrahams" wrote:
>
> > OK.  Can you come up with a replacement for the i.e. that describes, in a short simple
> > phrase, what ``absolute'' means?   Or (this is the possibility I dread) is it the case
> > that we don't really know what it means?
>
> It's almost true that a URI reference is absolute if and only if it begins with a
> scheme name followed by a colon.

If you believe the syntax in Appendix A of 2396, it's true, period.

> A URI, as distinct from a URI reference, *always* begins with a scheme name
> followed by a colon.  One must examine the RFCs to find out what scheme
> names exist.
>
> (There is a hack in RFC 2396 that allows deprecated relative
> URI references like "http:foo", meaning the same as "foo", provided that
> the base URI uses the "http:" scheme also.  If not, "http:foo" is undefined.)

It would be better to define ``absolute'' in semantic terms, i.e., not merely in terms of
its syntactic form.  My definition did that, albeit, as you point out, not quite
correctly.  If we just go by the syntax, then there's no question as to what an absolute
URI is.

Paul Abrahams

Received on Friday, 26 May 2000 14:25:37 UTC