Lists Home |
Date Index |
Simon St.Laurent wrote:
> At this point, I start to doubt the value of specifications when they
> use URIs rather than URLs. The Identification/Location mess has been
> swirling for years, despite the claims of those on the Identification
> that everything's just dandy.
Hmm, as far as the W3C goes, you're preaching to the choir. In fact, I
think the main reason that everyone talks about URIs is that RFC2396 is
the only IETF doc that's anywhere near up to date and being maintained,
and it defines URIs.
There is definitely some tension here, in that there are folks (mostly
in the general IETF direction) who claim to see a clear distinction
between names and addresses, and want to use URNs for names, and
specifically want to use URNs for namespace names because they can't be
dereferenced, and they think this is somehow better, and make statements
about good old-fashioned HTTP URLs being "single points of failure". I
have a hard time understanding this world-view, but that doesn't mean
For a lot of discussions that have swirled around URIness and URI karma
and so on, you could s/URI/URL/ without loss of information or
relevance. The trouble is, in a normative document, if you want to talk
about URLs, it's hard to find a good reference to point to for people
who want to look up the definitions.
> If the W3C wants to sort out Web architecture, a cold hard look at URIs
> would be the first target I'd recommend. That seems pretty unlikely
> overall, since URI supporters generally refuse to acknowledge that
> problems beyond misunderstandings exist at all.
Hmm, I think there are already some beginnings of that in the emerging
architecture document, and I suspect that it'll end up saying a bunch of
things that Simon agrees with. -Tim