Re: URI-protocol mapping (was Re: How to add new "protocols" ?)

> From connolly%w3.org@beach.w3.org Fri Feb 21 09:39:57 1997
> 
> touch@ISI.EDU wrote:
> > 
> > > OK, I'll bite: how is it that "location-dependent" vs.
> > > "location-independent" is a technical distinction?
> 
> > It's very technical. The host requirements RFC specifies locations
> > as either fully-qualified DNS names or IP addresses. And that's what
> > you have here. I.e., you have as much of a location as the internet
> > allows.
> 
> Ah! I wan't aware of that. I really appreciate you pointing
> that out.
> 
> OK, I'm happy with 'location-independent' as a technical
> term if 'location' is defined as 'FQDN or IP
> address'. I inferred the more geographic connotations.

> A quick scan of the URN requirements/framework draft[1]
> and the URN requirements RFC[2] doesn't
> show a similar definition of 'location'. And there's
> no reference to the host requirements RFC.

That is a clear inconsistency. Given that HTTP/1.1 does ref it.

> Hang on... I went to add it to my glossary of web
> architecture terms[3], but a brief scan of the
> host requirements RFC[4] shows:
> 
> |the DNS provides globally-unique,
> |              location-independent names.
> 
> If a FQDNs are locations, how does DNS provide
> location-independent names?

We're confusing the term location here, as is probably obvious.

The DNS provides globally-unique, 
	(geographically) location-independent names.

FQDNs are locations, just not geographic ones.

Joe


----------------------------------------------------------------------
Joe Touch - touch@isi.edu		    http://www.isi.edu/~touch/
ISI / Project Leader, ATOMIC-2, LSAM       http://www.isi.edu/atomic2/
USC / Research Assistant Prof.                http://www.isi.edu/lsam/

Received on Tuesday, 25 February 1997 13:31:13 UTC