RE: Fwd: I-D ACTION:draft-daigle-uri-std-00.txt

> I'd suggest that a clarification/revision of RFC2396 would be
> a good idea
> _prior to_ the W3C taking up any general consideration of how
> XML and URIs
> fit together.
>
> I'd like to think that a few of the past 1700 messages could have been
> avoided if RFC2396 was a tighter and more comprehensive document.

Not really - as you point out just below - many of the issues raised on
this list were/are not directly related to URI syntax but rather whether
XML (and in particular XML-NS) is part of the Web or not.

> >For many IETFers, I think an overview like this is quite
> appropriate and
> >helpful.

The problem is that it is proposed as a standard which would be wrong.

> Agreed, but I don't think it helps substantially with the
> kinds of issues
> that have been repeatedly raised on this list.

I agree

> Comparison of URIs is only the beginning, and I note that this draft
> doesn't refer to either Henrik Frystyk Nielsen's favorite RFC
> (2616) or Larry Masinter's favorite (2557).

RFC 2396 is actually fairly comprehensive in describing the syntax of URIs
and the BNF gives fairly clear rules on how to deal with relative URIs,
schemes, hostnames, port numbers, and hierarchical paths - in other words,
the common constructs defined by the RFC. There are plenty of URI parsers
that have been implemented to this spec. The only thing it doesn't do
directly is dealing with escaped characters but rather leaves that to
specific URI schemes which means that a URI parser doesn't have to
unescape.

It sounds more like you are after a description of the Web architecture -
or at least a terminology section describing resources etc. A syntax spec
is really not the right place to do that. In fact, there was an attempt of
writing a terminology document in the W3C:

    http://www.w3.org/1999/05/WCA-terms/

Although this draft is not actively being edited anymore, it may be a
better starting point - it was written out of necessity because we wanted
to characterize the Web.

Henrik

Received on Wednesday, 6 September 2000 11:43:06 UTC