W3C home > Mailing lists > Public > semantic-web@w3.org > June 2008

Re: Naive question on redirection.

From: Richard Cyganiak <richard@cyganiak.de>
Date: Sat, 7 Jun 2008 15:24:50 +0100
Cc: "Sergio Fernández" <sergio.fernandez@fundacionctic.org>, "Alan Ruttenberg" <alanruttenberg@gmail.com>, "Phil Archer" <parcher@icra.org>, SWIG <semantic-web@w3.org>
Message-Id: <E1AEFA3D-DE67-4E28-ACFF-8E0F3EBA3BAA@cyganiak.de>
To: "Jonathan Rees" <jar@creativecommons.org>


On 7 Jun 2008, at 12:24, Jonathan Rees wrote:
>> That is not correct. That HTTP dialogue only says that
>> in http://www.example.org/home.asp you can find an alternative
>> representation of http://www.example.org/. But nothing more
>> about http://www.example.org/home.asp.
>
> I don't see how your interpretation "you can find an alternative  
> representation" is supported by the HTTP/1.1 spec. In fact section  
> 10.3.3 doesn't use the word "representation" once.
>
> I also don't see any reason why one would want to interpret it this  
> way. The need to avoid equating the source and target resources of a  
> 302 sounds legitimate, but I'm not sure why such an equation would  
> be forced if non-IRs were allowed at the other end of the 302.
>
> I certainly interpreted the spec a different way, and am depending  
> on the ability to do a 302 to a 303, so I confess I have a stake.

Okay, so you do a 302 to a 303 to a description of the original  
resource. Why does this require a particular interpretation of the  
redirect?

A 303 doesn't tell you the relationship between the original resource  
and the target, it only means that the server thinks that people who  
want the original resource might find the target interesting.

I think throwing in a 302 doesn't change anything, no matter if you  
interpret it as “ask over there for the representation” or “this guy  
over there is same as me”.

So I find your practice consistent with Sergio's interpretation of 302.

> In any case it is clear that the HTTP spec is not really serving us  
> very well as its language around resources, requests, servers, and  
> representations was not written with  non-IRs in mind.

Good observation. Lawyering over the language used in the HTTP spec  
won't be too helpful here.

Richard



>
> I see that this is an open TAG issue [*]. Since SWIG email gets a  
> bad email triage outcome for me (very sorry) and possibly other TAG  
> members, one might consider moving this discussion to www-tag so  
> that the TAG can keep track of it more easily.
>
> Jonathan
>
> [*] http://www.w3.org/2001/tag/group/track/issues/57 ISSUE-57
Received on Saturday, 7 June 2008 14:25:30 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:41:10 UTC