[Asterisk-Dev] Features requests on bugs.digium.com

[Tilghman Lesher]

On Friday 31 December 2004 10:57, Lee Howard wrote:
> I'll give an example, and I hope that it's not a trite one, but because
> I'm involved with faxing so much it's the one I'll choose to use.
>
> Some time ago a friend of mine filed a bug report that Zap's faxdetect
> wasn't detecting incoming faxes from certain kinds of fax machines.
> Personally I'm not a strong believer in automatic fax detection because
> inevitably it fails with some fax machine or another, especially the
> (nowadays few) analog kind that are "listening" for ringback rather
> than "beeping" some tones.  However, my friend does use faxdetect, and
> he uses it somewhat religiously for his own purposes.  And, inevitably
> he encountered a number of fax machines that weren't detected as such
> by faxdetect.  So what did he do?  He filed a bug report with Asterisk,
> of course.  And then he went through the process of interacting with
> the bug marshals, running tests, attaching traces, etc.  And in the end
> it was determined that these fax machines weren't making the "proper"
> and normal fax tones that are expected.  (If that wasn't already
> obvious enough.)  And because the tones weren't "proper" the issue was
> determined to be "not Asterisk's problem", and the report was closed at
> that.
>
> Now, I wonder what we expect my friend to do.  Stop using faxdetect?
> Harrass these fax machine manufacturers into producing a fix and
> replacing these people's fax machines?  Tell the fax machine user to
> get a different kind of fax machine?  Well, none of these really are
> viable solutions.  Understand that the opinion of the fax sender is
> always, "I can send faxes to everyone else without any problem.  It's
> not my fault you can't get them.  Maybe I should just avoid this
> headache by not doing business with you."  That the fax machine could
> be detected is clear, so it's not really a technical problem where this
> fax machine is getting mistaken as my Aunt Mabel's squeaky voice.  In
> this case it ended up appearing to be a choice made that Asterisk will
> not ever support faxdetect for these fax machines.  And of course, my
> friend doesn't want to lose his customer, and so he sets up a different
> faxing route for them, certainly without faxdetect, possibly without
> Asterisk.  And this "selective stagnation" is where Asterisk is being
> impacted.

One of the fax standards is T.30 (4/99), issued by the ITU:

	4.2 Calling Tone CNG

	Function
	1) To indicate a calling non-speech terminal.  This signal is MANDATORY
	for automatic calling terminals and manual terminals.  However, manual
	calling terminals conforming to the 1993 and previous versions of this
	Recommendations MAY not transmit this signal.

So what we have here is that the current fax standard REQUIRES this tone
be sent, although older fax machines might not send it.  Why does it require
tone?  Because it's the ONLY reliable technique for detecting whether or not
the sending station is, in fact, a fax machine.

Now it's very likely that the system you're replacing has a particular phone
number which is devoted to fax, and all calls sent to that number are
delivered to the fax machine.  That's fine, and Asterisk can do that, too.
However, it's also likely that if somebody had sent a fax to the main number
of that company, a person would have answered the phone and gotten dead
air, as the sending fax waited to hear the receiving fax machine's auto
negotiation tone.

> If, on the other hand, my friend had developed a patch to Asterisk
> which would have permitted the detection of these fax machines, then
> I'm most certain that it would have been readily accepted into CVS
> without any arguments about it supporting "broken" fax machines.  So it
> seems to be a double-standard: happy to allow code contributions, but
> not happy to admit contributions of mere observation.  Maybe I'm wrong
> in this assumption, but to make sense of this double-standard I had
> merely assumed that the handling of the bug tracker was done under
> Digium's eye, and thus it was Digium's business interests that were at
> play in the decision making on the bug tracker, that those of us who
> disagreed with the handling of the bug tracker should set up our own
> bug tracker, and that we should maintain our own fork, if we dare.
> But, maybe I'm just looking for conspiracy where there isn't one.
> Maybe there is no reason behind the madness.

If your friend can develop a patch to allow Asterisk to detect fax machines
which aren't sending the CNG tone (without a significant number of false
positives), I have very little doubt that it will get accepted.  But that's
not the case here:  we have a fax standard, grandfathered in old fax machines
that don't follow the current standard, and Asterisk, which must detect a fax
machine (without engaging the fax machine).  Good luck.

-- 
Tilghman