- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Mon, 8 Jun 2009 17:48:54 +0200
- To: Javier Godoy <rjgodoy@fibertel.com.ar>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
On Jun 8, 2009, at 4:58 PM, Javier Godoy wrote: > Julian Reschke wrote: > >> Adrien de Croy wrote: >>> Since 205 is for resetting a form, and since therefore it's >>> intended audience is browsers, and since there's no apparent >>> browser support for it, perhaps 205 should go the way of 305. >>> ... >> >> Perhaps. Otherwise we probably should think about why it's not >> implemented. Maybe it isn't sufficiently clear from the >> description what it's good for? >> > IMHO the description is clear, but it is not implemented because it > doesn't give feedback to the user, thus it wouldn't be used. I think it was assumed that the user agent will provide whatever feedback is appropriate, according to the nature of that user agent. > Suppose a website returns 205 after successful form submissions. If > user-agents implement 205 as specified in section 10.2.6 (i.e. > "SHOULD reset the document view"), the human users will find that > after clicking the submit button, the form is suddenly blanked, and > they may think their submission didn't complete. > > This feature is better realized by returning an entity with 200 > status, containing the same (reseted) form, and a message such as > "submission successful, please enter more data" (not to mention > using AJAX and similar approaches). On the other hand, 205 > responses may be useful if the server will interact with purpose- > specific clients, because these clients may implement the user > notification by their own. > > BTW, I found interesting that HTTP cares about presentation issues... It doesn't -- HTTP cares about not re-transmitting the bytes of a document that the user already has in front of them. I don't remember who requested the status code, but I do know it was for data-entry forms that are expected to be submitted repeatedly. I have no objection to deprecating 205 if nobody implements it, but that would be a different issue. ....Roy
Received on Monday, 8 June 2009 15:49:28 UTC