Re: the meaning of stateless

On Mon, Apr 22, 2013 at 3:09 AM, Matt McClure <[email protected]>wrote:

> On Apr 21, 2013, at 7:52 PM, Brian Craft <[email protected]> wrote:
>
>
>
> Fielding's definition didn't do much for me, because the terms aren't
> adequately defined, and there aren't enough examples to clarify his intent.
> Also, some of the claims seem to be at odds with practice. Again, the "POST
> to query" thing involves storing state between requests, while Fielding
> says RESTful APIs are better because of "not having to store state between
> requests". Well, which is it?
>
>
> I think a key distinction is creating and destroying resources vs. other
> kinds of state. The former is a key feature of REST.
>
>
That's very much the impression I'm getting: that it's not generically
server-side state that is the key issue. It's something to do with the
difference between explicitly created state that can be directly referenced
with a URL vs. implicit state. If that is what it's about, then "stateless"
should be qualified, and some of the claims should be relaxed.