Re: Remaining issues related to redirects [#43 #295]

On 02/09/2011, at 6:34 PM, Julian Reschke wrote:

> On 2011-09-02 03:06, Mark Nottingham wrote:
>> 
>> On 01/09/2011, at 8:51 PM, Julian Reschke wrote:
>> 
>>> <http://trac.tools.ietf.org/wg/httpbis/trac/ticket/295>  - fragment handling
>> 
>> We discussed this a little while back, and the only pushback was from you as to whether it was appropriate for HTTPbis to go here.
>> 
>> We have feedback from Eric and Adam that they'd like to see us do this.
>> 
>> I put forth the idea of going to the TAG about it earlier, but when we asked them about #43, they didn't seem to have any strong advice (see<http://www.w3.org/mid/760bcb2a1003120853q6548f1a9u54f0aef723f4f45a@mail.gmail.com>).
>> 
>> Speaking personally -- I think it would improve interop if we did this, and I think we should also consider reopening #43 to define the case where both have a fragment (since there appears to be emerging interop).
>> 
>> Having HTTP define fragment combination (both in the single and dual cases) makes sense, because IME, client-side stacks handle HTTP redirects automatically, before the format-specific machinery ever gets to see the HTTP. Requiring -- or even allowing -- that this be format-specific opens up a huge can of worms in most implementations, and makes the Web a much more complex place.
>> ...
> 
> Are you saying that these stacks actually handle fragments? Example?

No, I'm saying that many handle redirects, and re-engineering them to delegate fragment combination to format-specific handlers will be awkward.

--
Mark Nottingham   http://www.mnot.net/

Received on Friday, 2 September 2011 08:53:18 UTC