Re: HTTP Status Code for TC returning EMPTY

On Wed, 20 May 2009, Michael Hausenblas wrote:

>
> Yves,
>
> As of my ACTION-81 I'd like to ask you (given your HTTPbis background),
> what's your take on the HTTP status code to be returned for the EMPTY case?

Well as the 'empty' case is an error in any case (and a client might 
detect an error before even sending something to the server), we have two 
options:

1/ display the error using a 416 (or a plain 400, but that would be a bit 
confusing).
2/ server the whole content. (then the client can cut the connection is 
this fallback is not what was wanted).

> Is this, in your opinion, a client error situation (4xx class) or should we
> handle it as Ok with no content (204)? Can you give us some evidence from
> your experience how the candidates (406/416 and 204 on the other hand) are
> used in the wild in order to asses if there are potential clashes?

Serving 204 in that case would give bad information for caches, as it 
states that the whole resource is empty (which is not the case here, we 
ask for an empty fragment, which is quite different).

  >
> Cheers,
>      Michael
>
> [1] http://www.w3.org/2008/WebVideo/Fragments/wiki/TestCasesDiscussion
>
>

-- 
Baroula que barouleras, au tiéu toujou t'entourneras.

         ~~Yves

Received on Monday, 25 May 2009 09:18:44 UTC