[asterisk-dev] proper way to signal user information layer 1 (ISDN)

[Steven Critchfield]

On Fri, 2007-09-14 at 17:09 +0200, Klaus Darilion wrote:
> 
> critch schrieb:
> > On Fri, 2007-09-14 at 15:24 +0200, Klaus Darilion wrote:
> >> Hi!
> >>
> >> In case of digital ISDN calls I want to signal the ISDN user information 
> >> layer 1 values between the dialplan (and maybe some apps) and the ISDN 
> >> channel driver.
> > 
> > I'll show some ignorance and some assumptions here. 
> > 
> > What all user information are you interested in setting up or is
> > available to be changed?
> > 
> > If you are just looking for a way to set up the H223 information, then I
> > suggest you go back to the drawing board with respect to the latest
> > suggestions.
> > 
> > For the H223 information, if the channel driver knows it is handling
> > H223 data, it should either request translation to a format it can
> > handle, or be able to set up the H223 call itself. 
>  >
> > This eliminates the dialplan designer from needing to get involved with
> > the whole video problem.
> 
> This would be possible. Whereas for testing purposes it would be usefull 
> to set it from dialplan. For example the user information layer may be 
> different in different countries (eg Austria vs. Japan)

But is it different based on origination or destination. If it is based
on origination, then the user information differences could be dealt
with just like tonezone is already, as a configuration time variable
instead of a call time variable. 

> > I would hate to have to design tests into the dialplan to determine the
> > call type and branch from there. The idea of me having a phone on my
> > desk capable of handling the 3g calls or regular calls and having to
> > hand out either conditional extensions based on the call type coming in
> > or test the stream from the dialplan would be just plain wrong. Handing
> > out a single extension that could just bridge the call regardless of
> > call type and let the underlying magic of asterisk make sure the call is
> > compatible is the right way to do things.
> 
> Usually magic is good, but sometimes there is need to enforce special 
> signaling.

Which is why I was wondering what other user information or other
signalling bits we would be interested in having up at the dialplan
level.
-- 
Steven Critchfield <critch at basesys.com>