Telex and TWX History [telecom]

[Bill Horne]

Donald E. Kimberlin

Telex certainly should be called the original form of E-Mail. Far from
"dead" on a global basis, UN reports published in the "Brittanica Book
of the Year" indicate there are about three million Telex lines around
the globe. Contrary to the impression international telephone people
like to create, direct, immediate access via Telex still exists to
more of the world's political entities than does telephone. This has
been the case for many years. (Totalitarian governments must like
Telex; they have been known to shut down telephone service, but not
Telex. The suspected reason: It can be monitored with hard copy
easily, and often [has] been, too. Of course, they themselves use it
for military messages.)

http://www.baudot.net/docs/kimberlin--telex-twx-history.pdf

--
Bill Horne
(Remove QRM from my email address to write to me directly)

[tlvp]

On Sun, 25 Dec 2016 12:15:53 -0500, Bill Horne wrote:

>
> ... Telex certainly should be called the original form of E-Mail. ...
>

Indeed, one of the prime attractions the early commercial email ventures
(Compuserve, AT&T Mail, and MCI Mail) held for me, back when my university
first made email available to its faculty, was the bidirectional eMail to
Telex gateway each maintained, for it was only Telex that eastern european
universities relied on for electronic transmission of text messages then,
rather than email.

Makes me wonder, now that both AT&T Mail and MCI Mail are no more: are any
bidirectional eMail to Telex gateways still being maintained? If so, where?

TiA, with Season's Best Greetings, and a Gracious New Year to all, -- tlvp
--
Avant de repondre, jeter la poubelle, SVP.

[Bill Horne]

On Sun, Dec 25, 2016 at 01:50:20PM -0500, tlvp wrote:
> On Sun, 25 Dec 2016 12:15:53 -0500, Bill Horne wrote:
>
> >
> > ... Telex certainly should be called the original form of E-Mail. ...
> >
>
> Indeed, one of the prime attractions the early commercial email ventures
> (Compuserve, AT&T Mail, and MCI Mail) held for me, back when my university
> first made email available to its faculty, was the bidirectional eMail to
> Telex gateway each maintained, for it was only Telex that eastern european
> universities relied on for electronic transmission of text messages then,
> rather than email.
>
> Makes me wonder, now that both AT&T Mail and MCI Mail are no more: are any
> bidirectional eMail to Telex gateways still being maintained? If so, where?

In 1986, I was using Western Union's email service, which was IIRC
called "Easylink". It had such a gateway, and I used it to overcome a
serious problem while I was applying for a license to operate my ham
radio station in Ireland. I was able to send and receive Telex
messages to/from the Ireland radio governing authority, twice in a
single day, which would have, at that time, been unthinkable by any
other means.

Bill

--

[Neal McLain]

On Sunday, December 25, 2016 at 11:17:01 AM UTC-6, Bill Horne wrote:

> Donald E. Kimberlin
>
> Telex certainly should be called the original form of
> E-Mail. Far from "dead" on a global basis, UN reports
> published in the "Brittanica Book of the Year" indicate
> there are about three million Telex lines around
> the globe.

[snip}

The quoted article states:
> Because it was using the Public Switched Telephone
> Network (DDD in American parlance, TWX was given
> reserved area codes... 510, 610, 710, 810 and 910.
> Some few remote locations on TWX are still on those
> area codes

Not any more. Those codes were recovered from TWX network and reassigned for POTS
voice service during 1992-94. During the same period previously-unused codes 210,
310, and 410 were also assigned.

As we all know the Great Area Code Format Change occurred on 1/1/1995. Prior to
that date the original 144 area codes were confined to two formats: N1X and N0X.
Starting 1/1/1995 new codes could be assigned with the middle digit in the range
0-8. But in the years just before the magic date there was a big demand for new
area codes so the N10 codes were assigned:

210 (previously unassigned) - assigned to San Antonio, TX area, split from 512.

310 (previously unassigned) - assigned to Los Angeles area, split from 213.

410 (previously unassigned) - assigned to Eastern Maryland, split from 301.

510 (recovered from TWX) - assigned to Oakland and East Bay area, CA, split from 415.

610 (recovered from Canada TWX) - assigned to Philadelphia area, split from 215.

710 - (recovered from TWX) - assigned to Government Emergency Telecommunications
Service in September 1994, covers entire United States.
http://tinyurl.com/710-GETS

810 - (recovered from TWX) - assigned to east central Michigan, split from 313.

910 - (recovered from TWX) - assigned to southeaster North Carolina, split from 919.

Source: Wikipedia
http://tinyurl.com/NANP-N10

Neal McLain

[Bill Horne]

On Mon, Dec 26, 2016 at 04:18:26PM -0600, Neal McLain wrote:
> On Sunday, December 25, 2016 at 11:17:01 AM UTC-6, Bill Horne wrote:

> The quoted article states:
> > Because it was using the Public Switched Telephone
> > Network (DDD in American parlance, TWX was given
> > reserved area codes... 510, 610, 710, 810 and 910.
> > Some few remote locations on TWX are still on those
> > area codes
>
> Not any more. Those codes were recovered from TWX network and
> reassigned for POTS voice service during 1992-94. During the same
> period previously-unused codes 210, 310, and 410 were also assigned.

When I was in "Toll Test" at the Back Bay CO in Boston, I would
sometimes have to test the dedicated lines that carried TWX dial tone
from the TWX CO at "Boston 2", which was known as the "WADS" office,
and had been leased to Wester Union after AT&T divested TWX.

We had a Model 35 ASR Teletype there in the office: it had been used
on TWX when AT&T had the service, but all the "official" machines,
which were used for time sheets and other daily reports, had been set
up with regular dial tones and were no longer part of the TWX network.

Because the "TWX" machine in our office wasn't on the TWX network
anymore, I had to figure out how to dial into the TWX network to send
test message to Western Union customers whom were complaining about
garbled characters and other "data" problems. It wouldn't work if I
dialed a "710" or other TWX number directly: my teletype would only
print garbled nonsense.

I found out that there had been two versions of TWX, and the original
system was, like TELEX, set up using BAUDOT machines running about 60
wpm. Of course, when the "100 speed" ASCII machines were introduced,
they had to interoperate with the older BAUDOT machines, so the calls
to 100-speed units from the older 60-speed machines were served only
via WADS offices, which were equpped with speed and code converters
that would translate between BAUDOT and ASCII, and do a speed
conversion in real time.

In order to tell 60-speed and 100-speed machines apart, any call form
the "regular" DDD network to a 100-speed machine (the ones with the
newer x10 area codes) was automatically converted from Baudot to ASCII
code, and the speed adjusted. That was the source of my problem: I was
using a "100 speed" machine to call from a "60 speed" phone number, so
the WADS office was trying to convert Baudot to ASCII, etc.

It turned out that the x10 TWX area codes were actually reversals of
the associated Terminating Toll Center (TTC) codes for the WADS
offices that handled each region. So, a 710-123-4567 number, dialed
from an ordinary DDD number, would be from an older 60-speed machine
to a newer 100-speed machine, and the code converters would be
used. However, if I used a tandem trunk line and dialed the call as
017-123-4567, then it would be completed properly, i.e., as a
100-speed to 100-speed call that didn't need any conversions.

I don't know if /all/ the 100-speed machines had to be served directly
from a WADS office, or if it was just the place where speed and code
conversions happened. I invite anyone with first-hand knowledge to
fill in the details.

[r.e.d.]

For readers who may not have seen it, I re-publish here my post from some
years ago:
+--------------------------------------------------------------+
A bit off topic, but this thread reminded me of one of my favorite
published papers (because of its sheer readability) and I could not
resist bringing it to the attention of others, old and dated though it
may be. Scrounge through the stacks of your local engineering
library:

Test yourself: how much do you know about international communications?

[International numbering systems]
Robrock, A.
Italtel, Milan;
This paper appears in: Communications Magazine, IEEE
Publication Date: Dec 1989
Volume: 27, Issue: 12
On page(s): 38-40
ISSN: 0163-6804
References Cited: 0
CODEN: ICOMD9
INSPEC Accession Number: 3582708
Digital Object Identifier: 10.1109/35.41420
Current Version Published: 2002-08-06

Abstract

We like to think of international telephone communications as
`transparent', the successful outcome of 100 years of technical
progress and standards setting, but the author shows us that it is
not. The user still has to be something of an expert to understand how
to make international calls, and there are chaotically differing
numbering systems for telephony, telex, and electronic mail. We should
be reminded that usability of services, not just their usefulness, is
a critical component of communications. Simplicity, consistency, and
rationality of service features and the `human interface' that allows
users to invoke them should be a high priority for communications
engineers as they work toward the integrated services networks of the
future

[HAncock4]

On Monday, December 26, 2016 at 10:06:03 AM UTC-5, Bill Horne wrote:

> In 1986, I was using Western Union's email service, which was IIRC
> called "Easylink". It had such a gateway, and I used it to overcome a
> serious problem while I was applying for a license to operate my ham
> radio station in Ireland. I was able to send and receive Telex
> messages to/from the Ireland radio governing authority, twice in a
> single day, which would have, at that time, been unthinkable by any
> other means.

Thanks for posting that interesting historical article about Telex.
I too wonder what remains in use today, 30 years later.

Here are some additional historical tidbits:

. On 2/26/1987 and 5/24/1988 the NYT reported that Western Union was
losing large amounts of money and was close to bankruptcy. Their
EasyLink service was cited as one of the reasons for the losses.

. In the 1930s, AT&T began TWX, a switched teletypewriter communication
service. Western Union always felt that should've been a WU service as
TWX definitely took telegram business away from WU. In the late 1950s
WU established Telex service in the U.S. and Canada.

. In 1968, WU announced it would acquire TWX from AT&T. In 1968, Telex
was a 50 baud/66.7 wpm/Baudot network with 26,000 subscribers while
TWX was a 110 baud/100 wpm/ASCII network with 44,000 subscribers.

. In 1971 WU acquired TWX. The facilities were to be leased from
AT&T for five years, then transferred over to WU. WU would build a
modern computerized switching network for it.

. In 1977, the physical transfer of leased facilities began. There was
new urgency, as AT&T was steeply raising its lease rates and that was
hurting WU. (I recall reading that pressure from MCI for equal treatment
didn't help matters).


In my own opinion, the significant reductions in dialed direct toll
rates, especially during overnight hours, that occured in the 1970s,
reduced the value of a dedicated separate Teletype network. While I
don't know how voice and TWX/Telex rates compared in the late 1970s*,
one could pre-punch a tape and send a reasonable message in one or
two minutes at a reasonable cost over the voice LD network, and short
haul calls could be as cheap as 5c/minute. Also, private line computer
networks were growing and could of course be used for messaging.

*In 1950, voice long distance was much more expensive and telegraph was
significantly cheaper, so far more long distance communciations was by
telegraph rather than voice.

I can't help but suspect the number of TWX/Telex subscribers fell off
significantly between 1970 and 1980 and even faster after 1980 when used
for domestic service. The big drop in overseas toll rates took a little
longer, as described in this thread.

[Bill Horne]

On Wed, Dec 28, 2016 at 08:43:01PM -0800, hanc...@bbs.cpcn.com wrote:

> On Tuesday, December 27, 2016 at 10:40:35 PM UTC-5, Bill Horne wrote:
>
> > In order to tell 60-speed and 100-speed machines apart, any call form
> > the "regular" DDD network to a 100-speed machine (the ones with the
> > newer x10 area codes) was automatically converted from Baudot to ASCII
> > code, and the speed adjusted. That was the source of my problem: I was
> > using a "100 speed" machine to call from a "60 speed" phone number, so
> > the WADS office was trying to convert Baudot to ASCII, etc.
>

> Would anyone know how the speed was adjusted? I would guess that
> in a long message the 100-speed machines would overrun the 60-speed
> machines unless there was some sort of intermediate storage buffer?

I'm glad you asked: I spent many hours figuring out exactly how this
was accomplished, and I'll pass this info along. We went over this in
the Digest four or five years ago, but I figure the old stories are
the best. ;-)

TWX 100-Speed machines had a "Restrain" feature, which was a circuit
that would pause the tape reader (and IIRC, also locked the keyboard)
when the 100-speed machine was sending to a 60-speed machine and the
message was too long for the buffer at the WADS office, which is where
ASCII was converted to Baudot, and 110 baud was converted to ~45 baud.

The operator at the 100-speed machine would see the "Restrain" light
come on, and that was their queue to wait until the 60-speed machine
at the other end had printed enough characters so that the buffer in
the WADS office could accept more from the 100-speed machine. Of
course, if the 100-speed machine was sending from its tape reader, it
would stop until the Restrain light went out.

I was curious how this feature worked, and I arranged for a
keyboard-to-keyboard call between our Model 35 and another machine in
our building, both of which were using regular dial tone and were not
part of the TWX network. When the operator of the other machine sent a
"X-OFF" command, which stopped the reader at my machine until he
sent a "X-ON" command, the "Restrain" light did not come on. In
other words, his "X-On/Off" keys would properly control my tape
reader, but did not cause my machine to go into "Restrain" mode.

This seemed like a small matter, but I wanted to know, and so I drew
up a list of all the possible signals another machine or the WADS
speed/code converter could send. If the "DC1" and "DC3" ASCII codes
weren't being used to control the tape reader on the remote machine,
then I figured that there had to be some different signal which was
unique to TWX.

In one of my rare occasions of looking for the simplest solution, I
dialed my TWX machine from the testboard, and fed it the "Mark" modem
tone from the test oscillator. The TWX sat quietly and did not drop
the call, for as long as the tone was on the line. If I interrupted
the tone for more than a fraction of a second, the TWX machine dropped
the call and turned itself off after a few seconds, so it was clear
that dropping the carrier signal wasn't part of the "Restrain"
mechanism. I had a moment of inspiration, and added another oscillator
to the circuit after I established a call to my Model 35 machine: one
was already set to the "Mark" tone, of course, and the other I set to
the "Space" tone.

Sure enough, the "Restrain" light came on when both Mark and Space
tones were present at the same time. I think Bell Labs chose to use
that dual-tone signal for that purpose because there was no ASCII code
to signal "Lock the other machine's keyboard". Anyone with first hand
knowledge is invited to tell us the rest of the story.

Bill

--
Bill Horne
(Remove QRM from my email to write to me directly)

[HAncock4]

On Tuesday, December 27, 2016 at 10:40:35 PM UTC-5, Bill Horne wrote:

> In order to tell 60-speed and 100-speed machines apart, any call form
> the "regular" DDD network to a 100-speed machine (the ones with the
> newer x10 area codes) was automatically converted from Baudot to ASCII
> code, and the speed adjusted. That was the source of my problem: I was
> using a "100 speed" machine to call from a "60 speed" phone number, so
> the WADS office was trying to convert Baudot to ASCII, etc.

[HAncock4]

On Friday, December 30, 2016 at 8:22:16 PM UTC-5, Bill Horne wrote:

> The operator at the 100-speed machine would see the "Restrain" light
> come on, and that was their queue to wait until the 60-speed machine
> at the other end had printed enough characters so that the buffer in
> the WADS office could accept more from the 100-speed machine. Of
> course, if the 100-speed machine was sending from its tape reader, it
> would stop until the Restrain light went out.

Thanks for the explanation.

Now I know what the orange lamp marked "REST" was used for on the
Teletype 33 with the built in dial. As we used our terminal for
computer time sharing, speed control wasn't an issue.

> I was curious how this feature worked, and I arranged for a
> keyboard-to-keyboard call between our Model 35 and another machine in
> our building, both of which were using regular dial tone and were not
> part of the TWX network. When the operator of the other machine sent a
> "X-OFF" command, which stopped the reader at my machine until he
> sent a "X-ON" command, the "Restrain" light did not come on. In
> other words, his "X-On/Off" keys would properly control my tape
> reader, but did not cause my machine to go into "Restrain" mode.

An early time sharing system we used (supported by an IBM 1130) was
very slow. When we prepared paper tape input, it was necessary to
punch an X-OFF at the end of each line. As the computer read our
tape, it would send out an X-ON when it was ready to receive the
next line.

The 1130 was not a good machine for this service, but I think they
got a good deal on it. It was replaced by an HP-2000, which was
designed for timesharing. It's BASIC was far superior, and it
was much faster. On that machine, when we wanted to enter paper
tape input, we gave it a command, and just let the tape fly on in.

The 1130 was a (relatively) low-priced machine for engineering and
science users. Despite being very slow I/O, it sold well. Presumably,
it was faster than someone sitting down with a desk calculator.
I believe it had somewhat of an open architecture, allowing it
to be utilized for a variety of applications, making it flexible.

bitsavers has a lot of material on the 1130:
http://bitsavers.informatik.uni-stuttgart.de/pdf/ibm/1130/

There's also material on the HP timesharing systems:
http://bitsavers.informatik.uni-stuttgart.de/pdf/hp/2000TSB/

[snip]

One other note. Originally, our Teletype 33 terminal for timesharing
had the built in dial and modem. Using it was simple--one pushed a
button, got a dial tone, and dialed the number. One other button
disconnected and shut the machine off. The machines were designed
for automatic answer so that they could be called and receive a
message. These machines were rented directly from Bell, I think
at $100/month, including free maintenance. (Let's just say users
frequently eating their lunch while using the machine did not
help its reliability).

These machines were replaced by plain Teletypes with a separate
modem. There, one had to put the handset in the acoustic
coupler, and could not receive calls automatically.

--

Happy New Year!

[IrishGuy]

There may well be some kind of legacy, privately managed, emergency
intergovernmental wire/cable services that are running over satellite or
leased lines or something. But, in terms of the publicly accessible networks
Telex would appear to have disappeared more than 14 years ago.

I had a look at the history of it here in Ireland and standardised Telex
services began in 1955 and finally ended with the equipment going into various
dumpsters in 2002. Even by then it didn't have many active users.

If you want to access Telex now, the main telco 'Eir' doesn't provide anything
like that. You can however connect to some kind of virtual Telex service via
SwissText in Switzerland but it seems to actually go by IP.

There doesn't seem to be much evidence of any live Telex services elsewhere in
Europe either.

I wonder are those 'live services' real or just some kind of virtualisation of
the old protocols or perhaps something else entirely being described as
"Telex" ?

[jhha...@earthlink.net]

Concerning Teletypes and modems.

Dial TWX used the 101 family of modems, which were installed in the cabinet of
the machines. The modem required most of the elements of a telephone, so
Teletype Corp. had to build that into the call control unit. There was a 99
wire cable between the machine as made by Teletype and the modem as made by
Western Electric.

The other service that came out about the same time was DataPhone, which used
the same modem but presumed that the operators would talk on a DDD call before
switching to data transmission. Since TWX was priced cheaper than DataPhone
the TWX set had no way to talk on the circuit.

Some of the Bell operating companies realized that they could save some money
by supplying a private-line Teletype and a 103 series modem, which had the
modem and telephone all built into one package.

Those modems used a separate tone pair for each direction of transmission. A
strapping option controlled which tone pair would be used for originate or
receive. Then each tone pair could use the higher or lower frequency tone for
the mark signal. Thus there were eight different mutually incompatible
services that could use the same modems. TWX and DataPhone were two of them,
and there were plans for a WADS and WADS-prime and maybe some additional
non-interoperable services. FCC rejected the proposed WADS tariff.

Then when DataPhone was being used mostly for terminal-to-computer operation
there was no need for the terminal to accept incoming calls so the modem could
be simplified by making it originate-only. And then the acoustic coupled
modem came along as a way to avoid first renting a modem from Bell, and later
to avoid paying for the Data Access Arrangement that Bell required for a
while if you wanted to use a customer-provided modem.


***** Moderator's Note *****

IMNSHO, "Dataphone" was just a way to continue to offer TWX service
without advertising it as TWX, after AT&T was forced to sell the TWX
network to Western Union. I never saw anyone use the phone on the side
of the Teletypes that we rented under "Dataphone" service.

To be fair, most Dataphone Teletypes were used for computer access, so
there was little need to talk on them anyway. I don't know if there
was an option to strap them for full-duplex, though, and if there
wasn't, users had to put up with transmission errors that they
wouldn't see at the time they were typed.

Bill Horne
Moderator

[Jim Haynes]

There is a strictly hobby version of Telex developing now, called
i-Telex.

http://www.i-telex.net/