Re: [dispatch] draft-jesske-sipping-etsi-ngn-reason-04

["Dale Worley" <[e-mail]>]

On Wed, 2009-11-11 at 10:40 -0500, Paul Kyzivat wrote:
> 
> [e-mail] wrote:
> > Hi Paul,
> > Sorry. I was to fast with my writing an sending. 
> > What I meant to write was the 100. But I dont know any other 1xx beyond 18x and 199 that would fit. Or did I forgot one.
> > 
> > Or is it more the general avalibility for 1xx for future.
> 
> Enumerating which ones are meaningful isn't future proof.
> Maybe this is a special case since there is a requirement to explicitly 
> document which responses this header may appear in (why?),
> but IMO it would be better not to second guess, and just allow them all.

Yes, we should generically allow Reason in any 101-199 or final
response.  The proper use of "Reason: Q.850;cause=..." is "to describe
the cause for the generation of the response by a datum that is to be
interpreted based on the Q.850 standard".  The usual reason for this is
that the ultimate cause of the generation of the response was an SS7
protocol operation that contained the specified cause code, but even
that need not be true.  The goal is to specify how the receiver should
interpret the header; based on that, the restrictions on the sender are
implied.

Dale