W3C home > Mailing lists > Public > www-tag@w3.org > August 2008

Re: Question about the On Linking Alternative Representations TAG Finding

From: Roy T. Fielding <fielding@gbiv.com>
Date: Thu, 7 Aug 2008 13:59:04 -0700
Message-Id: <8623C049-CB75-492F-B89F-9C097E057660@gbiv.com>
Cc: "Sebastien Lambla" <seb@serialseb.com>, "T.V Raman" <raman@google.com>, <john.kemp@nokia.com>, <www-tag@w3.org>, <kidehen@openlinksw.com>, <tthibodeau@openlinksw.com>
To: Richard Cyganiak <richard@cyganiak.de>

On Aug 7, 2008, at 7:34 AM, Richard Cyganiak wrote:
> Content negotiation as defined in the HTTP spec does *not* involve  
> redirects.

Wrong.  http://www.w3.org/Protocols/rfc2616/rfc2616-sec12.html#sec12.2

The 300 response isn't used much because it was spottily implemented
by browsers (last I checked).  302 is commonly used to direct initial
visitors to language-specific subtrees.  That is a form of conneg, as is
any form of handling the request based on the content of the request
headers instead of just the method and URI.  In spite of the one-time
redirect, it ends up being more efficient than varying generic resources
because language has locality of reference and all of the content
(including in-line images and stylesheets) are also language-dependent.

HTTP provides several different techniques for conneg because each
is *best* for at least some subset of resources.

....Roy
Received on Thursday, 7 August 2008 20:59:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:56:19 GMT