Path: utzoo!attcan!uunet!wuarchive!julius.cs.uiuc.edu!apple!bionet!hayes.ims.alaska.edu!accuvax.nwu.edu!nucsrl!telecom-request
From: chapman@alc.com (Brent Chapman)
Newsgroups: comp.dcom.telecom
Subject: Re: NJ Bell Lab in Morristown, NJ
Message-ID: <14456@accuvax.nwu.edu>
Date: 6 Nov 90 21:54:01 GMT
Sender: news@accuvax.nwu.edu
Organization: Ascent Logic Corporation; San Jose, CA
Lines: 59
Approved: Telecom@eecs.nwu.edu
X-Submissions-To: telecom@eecs.nwu.edu
X-Administrivia-To: telecom-request@eecs.nwu.edu
X-Telecom-Digest: Volume 10, Issue 798, Message 6 of 11


In <14343@accuvax.nwu.edu> matt_mcgehrin@pro-sherwood.cts.com (Matthew
McGehrin) writes:

>In-Reply-To: message from sba8_ltd@uhura.cc.rochester.edu

>I have to find the number, but I remeber calling a system in
>Morristown, NJ. It was a NJ Bell Lab. It was really interesting.
>Depending on which number you called you could make the system do the
>following:

   (examples deleted)

Are you sure this was NJ Bell, and not BellCore?

A few years ago, some of the researchers at BellCore (can't recall
their names at the moment) started experimenting with using a UNIX
system to drive a phone switch that was controlled by commands sent
over an RS-232 serial port.  I seem to recall that the work was
described in some detail in a paper given at one of the USENIX
conferences a few years ago.

The switch, if I recall correctly, had a trunk (DID?) interface on one
side and lots (40 or so?) of individual lines on the other.  Most of
the lines had just normal touch tone phone sets attached to them, but
many had things like voice syntesizers, recorders, radios, and so on
attached.  They created a scripting language for controlling the
switch, so that they could associate different actions with calls to
different numbers.  When a call came in on a certain number, the
computer would find and execute "program" for that number.  The
program could then do something like grab a free voice synthesizer,
connect that to the circuit the call was coming in on, and then have
the voice synthesizer say whatever they wanted it to.  Since the voice
synthesizer had a touchtone decoder, they could recognize TT input
from the caller, and the program couliWd respond accordingly.

The voice synthesizers they were using were fairly featureful, and
could do things like vary the tone of their output, to produce (for
instance) either a deep, masculine voice, a warm, feminine voice, or a
squeeky little child's voice.  One of the researchers used this
feature to create what they called the "touch tone shell".  Basically,
you'd call up the number for this service, and it would connect you to
a UNIX system, with the voice synthesizer "reading" the output which
would normally be on the screen.  If I recall correctly, stdin was
read back in the masculine voice, stdout in the feminine voice, and
stderr in the squeeky child's voice.  You could use two-digit codes on
the keypad to enter whatever characters you wanted as input, to
control the system.

I wish I could remember more about the system.  It was absolutely
fascinating to read about, and wish for.  Hopefully someone who knows
more about it will read this, and can provide pointers to the original
articles and papers describing the system.


Brent Chapman                    Ascent Logic Corporation
Computer Operations Manager	 180 Rose Orchard Way, Suite 200
chapman@alc.com                  San Jose, CA  95134
                                 Phone:  408/943-0630