[Rest] intiial rough draft

[Dale Worley]

On Tue, 2009-03-03 at 18:38 +0000, Theo Zourzouvillys wrote:
> On Tue, Mar 3, 2009 at 9:10 AM,  <Simo.Veikkolainen at nokia.com> wrote:
> > Thanks for putting the draft together. I would like to propose one more case to call forwarding: call forwarding on no answer. This would align the feature set with current GSM supplementary services.
> 
> After a small amount of thought, i came to the (perhaps incorrect?)
> conclusion that call forwarding on on answer was simply "all" (also
> known as 'unconditional') with a timeout?
> 
> Is there a reason they should be separate?  I have no objection to
> adding a new type in if there is!

People think of them as different features, although that does not
determine how we should think of it.

The critical difference seems to me to be that "call forwarding all"
does not cause an INVITE to the original destination.  Very strictly,
"call forwarding after timeout 0" will send an INVITE, followed
immediately by a CANCEL, followed by an INVITE to the alternative
destination.  So we would have to define "call forwarding after timeout"
to have exceptional behavior if the timeout = 0, namely that the INVITE
to the original destination will not be generated.

At this point, defining "call forwarding after timeout 0" seems
complicated enough that I think we should just define two
(non-parameterized) alternatives, "call forwarding on no answer" and
"call forwarding unconditional".

Dale