[PRL] REST?

[Richard Cobbe]

On Wed, Feb 16, 2005 at 07:40:42AM -0500, Matthias Felleisen wrote:
> On Feb 16, 2005, at 6:53 AM, Richard Cobbe wrote:
> 
> >On Wed, Feb 16, 2005 at 01:19:48AM -0500, Johan wrote:
> >>REpresentational State Transfer.
> >>
> >>As near as I can tell, the idea is that the URL represents the
> >>totality of the current client state.  [ie, very distictly NOT a
> >>session key].  The connection to continuations would appear to be
> >>that if the server maintains client state in a hashtable of
> >>continuations, each continuation's hash makes a fine rest URL.
> >
> >Does it?  I don't know much about designing web apps; that's more
> >Greg's area.  But it seems to me that a continuation hash is just a
> >fancy session key.  If you actually want to represent *all* of the
> >client's state in the URL, then you'd need to do something like
> >representing the continuation itself (rather than a continuation ID)
> >in the URL -- a bit like what Greg's been working on recently.
> 
> Yes, but I doubt that's enough. You need the state of variables, too. 

Yes, of course.

Greg, aren't you starting out by assuming purely-functional web apps?
Or am I misremembering?

Richard