W3C home > Mailing lists > Public > www-tag@w3.org > December 2007

Re: httpRange-14 Two Years On

From: Sean B. Palmer <sean@miscoranda.com>
Date: Tue, 4 Dec 2007 17:43:02 +0000
Message-ID: <b6bb4d890712040943s433370c3s314f29908ce0b0e@mail.gmail.com>
To: www-tag@w3.org

On Dec 4, 2007 5:15 PM, Sean B. Palmer wrote:

> If it returns a 200 then it could be documentation about Amaya
> or it could be Moby Dick

Thinking about it more, I think the crux is here. TimBL says [1] that
in fact the 200 response's representation is part of the content of
the information resource, and then in [2] he says that the definition
of some information resource's content is a kind of social contract
based thing.

So you could serve a page about Moby Dick from
http://example.org/amaya and have it denote Amaya documentation, or
you could serve Amaya documentation from it and have it denote Moby
Dick, but both of these things would be breaking social contract
(which is fuzzy, but okay).

This still leaves the problem that when you return a 303 you don't
know if it's an information resource or not. As the TAG resolution
states, 'If an "http" resource responds to a GET request with a 303
(See Other) response, then the resource identified by that URI could
be any resource'. *Any* resource, including documentation about Amaya!

Perhaps there's a social contract for 303 responding URIs too? Say I
have a document that 303s to the Wikipedia entry on the Moon. Would
you guess the URI denotes:

a) the moon
b) Wikipedia's entry about the moon
c) some other documentation about the moon
d) or something else?

The thing is that Tim claims the content/social contract for
information resources as axiomatic for the web. What about social
contracts for 303s; are such a thing even possible?

I'm a little dubious of the former's axiomaticity, I mean based in
specifications at least (except for those famous two words in RFC
2616), but at least I can understand now why it's consistent with
common sense. It doesn't solve the whole problem though--unless
there's a great consensus for my question above--so why deploy a
partial solution?

Then again, if it's an axiom, perhaps it's not being deployed as a
partial solution at all; it's just this thing which makes sense.

[1] http://lists.w3.org/Archives/Public/www-tag/2007Nov/0034
[2] http://lists.w3.org/Archives/Public/www-tag/2007Nov/0028

-- 
Sean B. Palmer, http://inamidst.com/sbp/
Received on Tuesday, 4 December 2007 17:43:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:51 GMT