> Actually.. I
seem to have jumped to improper conclusion..
One thing you will
find abour spandsp is that some fax machines will just plain have a problem
sending to spandsp, period. Mr. Underwood has localized it to cetain HP fax
machines and I can confirm that HP all-in-one SOHO machines exhibit this
problem, but I have also noted it with Canons and Panasonics. In my context, I
have several hundred fax machines (~400) that can potentially send to me,
and a certain percentage exhibit this problem. Of these fax machines, I have
narrowed down a list of 12 that constantly crap out when using spandsp. My
workaround, while not very elegant, solves the problem for me: When a call
comes in matching callerID X, forward the call to a "real" fax
machine:
exten =>
5555,1,Gotoif($["${CAALLERIDNUM}" = "5555551212" ]?,2:3)
exten =>
5555,2,Dial(ZAP/g0/5552222,60)
-keep going down
the list until the list is exhausted, then kick in RxFAX-
In the end, for
me, it was a tradeoff between the flexibility of spandsp vs these niggly
little problems, and if you do the math, in my case it's only 3% of fax
machines exhibit the corruption problem, so it was worth it to do this
workaround and train the office gopher to check the "real" fax machine every
couple of days.
FWIW, I did a
back-of-napkin calculation using spandsp vs our old method of having a bank of
fax machines and came up with an astonishing 40,000 sheets of paper saved
annually, with a cost of over $10,000 cdn saved annually (this
included labor costs to distribute the paper fax + our "click rate" on our fax
/ copiers) and we get the added bonus of taking the generated PDF and
attaching it to our in house CRM; if you can believe it, we used to take the
fax (on paper) and scan it into our CRM as a PDF, and spandsp + tiff2ps +
mine-construct took care of this step for us. So for me, it was totally worth
it to go through teething problems with spandsp and I would not go back for a
second.