[OGo-Discuss] CTI Integration

[Adam Tauno Williams]

> Again, some people do this, but its obviously an even smaller  
> minority. And extremly uncommon (the far majority of users just have  
> a Windows PC).
> Anyways, I'm not complaining :-) If you want to add support for such  
> setups, feel free to do so! :-)

Failing to resolve by device name,  I think falling back to looking up
an extension via username is reasonable.  How to do that gets
complication,  querying the users object on the groupware server for an
extended-attribute/property makes sense.  Although it might make just as
much sense to go-around and look at the user's account in LDAP (I image
most shops with PBXs, VOIP, etc... are also using a directory).

> >> I'm not sure why you need a separate number for that? Sounds to me
> >> like it would be any of the regular numbers which is then matched
> >> against the phone prefix pattern of the PBX?
> > ah, you mean this snippet of code, I found in the STLIDialer.m:

Yes, that makes sense.  If the hasPrefix part is configurable;  the
user's object on the groupware server certainly contains their own
number.

> > #warning need prefix mapping table here ...
> >   if ([_number hasPrefix:@"+493916623"])
> >     _number = [_number substringFromIndex:10];
> >   else if ([_number hasPrefix:@"+49"]) {
> >     _number = [_number substringFromIndex:3];

Since many VOIP systems, including ours, are federated systems (many
cooperating systems) it would be very nice if hasPrefix was an array.

Of course, out VOIP system isn't Asterisk. :(  We have Nortel BCM
systems,  which run Embedded NT 4.0 [Ick!] currently.  But the next
upgrade turns that all into Linux boxes, which is pretty cool.

But the concept/configuration should be pretty common.