Xref: utzoo comp.sys.mac:43017 alt.fax:430
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!wuarchive!brutus.cs.uiuc.edu!apple!rutgers!att!cbnewsh!aaron
From: aaron@cbnewsh.ATT.COM (aaron.michael.chesir)
Newsgroups: comp.sys.mac,alt.fax
Subject: Re: Multiplexing FAX & voice
Summary: The 2 "conclusions" are erroneous
Message-ID: <6110@cbnewsh.ATT.COM>
Date: 27 Nov 89 21:34:42 GMT
References: <0ZQIlkm00UhBM0wn5n@andrew.cmu.edu>
Organization: AT&T Bell Laboratories
Lines: 46

In article <0ZQIlkm00UhBM0wn5n@andrew.cmu.edu>, jhm+@andrew.cmu.edu (Jim Morris) writes:
> A while ago we had a long discussion here about using a single phone
> line for incoming FAXes and phone calls from people. The conclusion
> seemed to be
> 
> 1. Outside the U.S. the sending FAX announces itself first, so a device
> that detects the signal was feasible.
> 
> 2. U.S. FAX machines expect the receiver to announce its intentions, so
> the best solution is a FAX machine that starts off as a phone answering
> machine, asks the caller for a message, and switches to FAX mode if it
> doesn't detect voice.

Neither of these are true. All Group 3 FAX machines (which are the current
business standard), whether made in the US, Pacific Rim, Asteroid Belt, or
elsewhere MUST conform to the CCITT T.30 FAX protocol specification. The
latest revision to it was made in 1988 and can be found in the CCITT "Blue"
Book. The revisions refer only to the enhancement of an Error Correction
Mode (ECM, for short). The 1984 version (the Red Book) is much more available
and is good enough for a beginner's understanding. Anyway, the point is this...

Whenever a Group 3 is called, it responds with a V.21 HDLC packet or packets
that identify its CAPABILITIES (what speed it can operate at, ECM capability,
resolution capability, etc.). The last of these initial HDLC frames is 
called the DIS frame and is the frame with the above capabilities indicated
by the settings of the various information bits. The calling station then 
responds with packet(s) indicating its INTENTIONS. The intentions are encoded
in either an NSS or DCS packet (assuming that the calling station is also the
sending station) or in the NSC or DTC packet (assuming that the calling
station is polling the called station for FAX pages).

While T.30 specifies a protocol, it also allows "like" machines (machines
from the same manufacturer) to use any protocol they wish, AFTER the two
machines finish the above handshake. Basically, one of the initial packets
sent by the called machine is the NSF packet, and identifies the manufacturer
within a particular country, and capabilities if communicating with a "like"
machine. One of the return packets (generated by the calling machine) is the
NSS or NSC packet, which lets the called party know that it is a "like" machine
and also identifies its INTENTIONS.

This is all I can say while standing on one foot....


Aaron Michael Chesir
cbnewsh!aaron OR ..att!twitch!aaron
1-201-949-1530