Raphaël Luta escribió:
>
> Christian Sell wrote:
> >
> > > - the semantic of the GET requests is they should *never* change the state
> > > of a server, this is not true for a login screen or for most form
> > handling
> > > screens for this matter.
> >
> > sorry for jumping in, but I'm getting curious - where is that "law" laid
> > down? How exactly is it justified?
> >
>
> Well Santiago already quoted the spec, so I'll just add that the "law" is
> actually only a best practice because it's possible to write idempotent
> GET operations that do modify the server state if you really want to.
>
> I perfectly know that most web developers don't care about this (and
> Turbine does a poor job of enforcing this), it's just that I'm a HTTP
> facist old timer (the GET idempotence requirement was already defined
> in HTTP/0.9: http://www.w3.org/Protocols/HTTP/AsImplemented.html)
>
I am of the same thinking. Less formally, if you have GET uris that
don't deliver "consistent" results, search engines are no longer useful,
and bookmarks, and even the concept of hyperlink. :)
I try to:
- not modify server state as a result of a GET operation.
- different GETs of the same uri should deliver the same result, except
for other server state changes (for instance if the database was updated
separately, or a mail arrived to the list in a list archive).
> --
> Raphaël Luta - [EMAIL PROTECTED]
> Vivendi Universal Networks - Services Manager / Paris
>
> --
> --------------------------------------------------------------
> To subscribe: [EMAIL PROTECTED]
> To unsubscribe: [EMAIL PROTECTED]
> Search: <http://www.mail-archive.com/[email protected]/>
> List Help?: [EMAIL PROTECTED]
--
--------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/[email protected]/>
List Help?: [EMAIL PROTECTED]
