Re: httpRange-14 Adjunct: 302 is Valid for Non-Information Resources

On Nov 30, 2007 4:15 PM, Richard Cyganiak wrote:

> I would prefer that the TAG considers the future of the web,
> instead of wasting time in discussion of hypothetically existing
> obsolete clients.

Yet doesn't the httpRange-14 issue as a whole hinge on the
interpretation of the following passage from RFC 2616?

   3.2.2 http URL

      The "http" scheme is used to locate network resources via the
      HTTP protocol.

Which argument over which passage is the greater waste of time? Which
argument's proposed resolution is the more detrimental to the future
of the web?

> The burden is on you to give a few examples of clients that
> don't support 303

Not at all. Compatibility is ipso facto desirable, and thus the burden
is on you to prove that it is no longer required. Unfortunately to do
so would require you proving a negative (that no such clients exist),
which you can't.

We could play this game across many more emails, but my case is
watertight. The tension isn't in my interpretation. Non-Information
Resources may be served using 302.

> FWIW, RoyF's proposed new definition of the 303 code from the
> HTTP issue tracker [1] doesn't include the backward compatibility
> clause. Do you think he was wrong in dropping it?

No.

> Evidence please.

Because it doesn't say that you can't use 302 for non-network
resources, so that's fine. Or does it? You tell me. But nobody dares
talk about that, do they, otherwise the case for 303 would crumble too
:-)

P.S. Note that purl.org support 302 but not 303; and that they're a
popular choice for publishing Semantic Web ontologies. Do you think
they should have to support 303? What if they never do?

-- 
Sean B. Palmer, http://inamidst.com/sbp/

Received on Friday, 30 November 2007 16:58:24 UTC