[GreenKeys] Re: Model 29 History

[Eric Scace K3NA]

Don and Ben --

   This message provoked a lot of old brain cells!  Here is some ancillary information.

   Frieden explored applications for their machines in at least one other area beside 'data processing' (as described by Ben).  I
had first hand experience with Frieden equipment in the newspaper typesetting world.

   In one implementation, Frieden provided an ASR-style machine with a keyboard, printer, reperf, and TD.  This machine was used for
manually composing an article; more on this process in a moment.  A second, similar machine was used to produce a finished version
of a column of newsprint by typing the article using a one-time carbon film ribbon on high gloss, extremely white paper.  This paper
sheet was photographed and the negative was pasted up in the preparation of a photo-typeset page for the newspaper printers.

   The second machine differed from the first in that it had TWO tape readers.  The reason for this lay in a clever solution of the
problem of justifying a line of proportional-spaced type.  ("Justification" is the process of causing a printed line to exactly fit
with width of a column; it's achieved by making the spaces between words slightly fatter or narrower, by clever use of hyphenation,
and very occasionally by inserting minimal spaces between letters of a word.)  TTY machine lovers will delight in learning how it
worked.

   As the keyboard operator typed a line of text, the composing machine would punch a paper tape with each character typed.  I
believe this was a 6-level tape, with the 6th bit used to indicate upper vs. lower case... but it might have been a 7-bit code.
When the operator reached the approximate end of the line and hit the Return key, the composing machine would add a special code
indicating the total width of the line to the tape.  The width code was preceded by a single character on the tape that was not
assigned to any printing character.  I think the width code was 2 characters long; it was just a binary representation of line
length.

   The typescript on the page in front of the operator was just for the operator's reference to see if he made any typing errors,
and to see when he was approaching the end of the line.  It was usually typed on junk paper with a cloth ribbon.

   After the article had been punched into tape, the tape was moved to the second machine.  The head of the tape was threaded
through the first tape reader, and then immediately to the second tape reader, which stood a few inches behind the first.  The first
tape reader would read through the first line of text, hunting for the justification code at the end of the time.  Meanwhile the
printer just sat idle.  As soon as the first tape reader found and read the justification code, the printer now knew how large a
space to insert between the words of text.  The second tape reader would then re-read the first line of text, and the printer would
type it, inserting a space of appropriate width between the words.  [There were some special codes used for spaces that should not
be adjusted; e.g., indentation at the start of a paragraph or in a bulleted list.]

   Meanwhile the first reader was ticking away, looking for the justification code at the end of the 2nd line of text.  And off the
system would go, working its way line by line through the entire article.

   Now, if the operator made a typing error, there was just one recourse -- to cancel the entire line and retype it.  The "line
cancelled" code was simply a variant of the justification code added at the end of the line on the paper tape.  When the first tape
reader on the final printing machine read a "line cancelled" code, it just munched away to the next line... and the second tape head
would munch through the cancelled line without allowing the printer to print.

   The character codes were just ITA2 plus a case bit.  This meant that a clever operator had a second option available to him upon
making a typing error.  He could read the paper tape holes and decide if he could over-punch the paper tape in a way that would
correct the error -- without screwing up the justification arithmetic!  If so, he just backed up the tape and repunched it.

   I worked at a small newspaper that was replacing its Linotype (hot metal) typesetting equipment with Frieden machines.  (I was
one of the very last people trained to operate hot metal Linotype machines.  If you think TTY machines are a thrilling Rube Goldberg
device, you'd love working on a Linotype!  It was bigger than an organ console, with hot lead, motors, bands and belts, and all
sorts of marvelous mechanical inventions.)  Of course the management decided to give the "new-fangled stuff" to the 18-year-old kid
on the staff.  (That would be me.)  By the time I got involved with this, I had been operating RTTY for several years as a Navy MARS
operator and knew ITA2 codes by heart.  The other typesetters were amazed that one could fidget with a paper tape and correct these
kinds of errors -- something clearly not in the operator's manual.

   The Frieden printers used type baskets in order to keep the printed line precisely aligned on a page.  They were not terribly
fast -- perhaps typing around 80-90 WPM but with frequent pauses to let the mechanics settle down after case shifts, returns, etc.
And they were temperamental.  As I recall these machines lasted only a couple of years at that paper.

=========

   Touch Tone dials:  In fact, DTMF keypads were initially designed with 16 buttons.  Ben is correct that only the digits 0-9
appears on customer telephones.  But the operator switchboards included all 16 buttons.  Special network functions were assigned at
some point to some of those extra buttons, but I can't recall when this was done (before my time).

   By the way, the # key is properly called "octothorpe" (the typesetter's name; i.e., eight points) or "square" (ITU's definition,
when used in this context).  If you look carefully at a European or Japanese version of the 12-button keypad, you'll see the square
symbol has vertical lines and horizontal lines crossing perpendicularly, like a square with little extensions.

=========

   Anyone going to Dayton?  I normally hang out at the contest, 160m, and DX events... but I would enjoy meeting Greenkeyers!

   -- Eric K3NA

-----Original Message-----
From: greenkeys-admin@mailman.qth.net
[mailto:greenkeys-admin@mailman.qth.net]On Behalf Of Don Robert House
Sent: 2004 May 7 00:01
To: K9kom@aol.com
Cc: Ed Hickey; Bob Liddy; Bob Cnota; Ken Clinkman; Phil Schelinski;
Steve Kissinger; Warren Brader; David Weil; Rush Glick;
greenkeys@mailman.qth.net
Subject: [GreenKeys] Re: Model 29 History


Thanks a million for this history Ben,

You will have to come over and see the last machines of the TELETYPE
marke in my garage.  I think I am the only one left with a Model 45
printer!

Give me a call when you get a minute.  Sometime soon Jim Haynes is
supposed to be in our area, it would be nice to get together.
---------------------------------------------------
Don R. House
4716 Patty Lane, Ringwood, IL 60072
Tel: 815-653-0683
FAX: 815-653-0684
*****************************************
North American Data Communications Dept. of CMA
URL: http://www.nadcomm.org
Computer Museum of America (CMA)
URL: http://www.computer-museum.org


Don

>The AOL server lost your inquiry regarding the Model 29, but I
>recall you were looking for some history, so here goes.
>
>I started at Teletype in 1960, and that was before ASCII.  As you
>know, the biggest problem with Baudot code was the FIGS-LTRS shift,
>which slowed things down and used a 3-row keyboard.  The military
>and others had been asking for a 4-row keyboard for some time, and
>the Model 29 was the answer at the moment.
>
>The Model 29 was styled just like the Model 28, but used one inch
>wide 8-level paper tape.  The big technical achievement was a front
>plate which would go immediately from a letter to a number and back.
>There were some shift links on the front plate which would break
>with extended use, but nevertheless the Model 29 was good enough for
>service.  The same typing unit went on to be called the Model 35
>with a different type box and some other minor modifications.
>
>There were two markets for the Model 29.  The first was the
>military, and their units had Fieldata code.  I never got into that
>code, and I don't think the military bought all that many.  The
>other market was for as a console for main frame computers, and we
>sold a bunch to the likes of GE, RCA, Univac, Honeywell, etc.   The
>reason they bought the Teletype machine was because IBM was driving
>its computers with heavy duty IBM typewriters and IBM punched card
>equipment, and their competitors would rather die than ship a
>machine with an IBM typewriter as the console.  Although Friden had
>a similar heavy duty typewriter, Teletype was much better known and
>their equipment was more rugged.
>
>The Model 29 headed for data processing ran on the Extended Binary
>Coded Decimal Interchange Code, or EBCDIC.  This was a pure IBM
>code, and "extended" meant that it included characters for both
>upper and lower cases, although the lower case letters were never
>implemented by Teletype.   The first on-line multipoint Teletype
>circuit run by a computer was the Westinghouse "whistle" system,
>with an RCA computer in Pittsburgh driving the whole show.  The
>console printers there were Model 29's; I don't know whether the
>machines on the circuits across the country were 28's or 29's.
>
>I mentioned Friden, and they were pushing a concept called
>"Integrated Data Processing" (IDP) in which paper program tapes were
>used to type repetitive letters and forms, with an output tape
>capturing the data for subsequent processing.  The Model 29 IDP set
>was specifically designed to go after this market.  We were afraid
>that Friden was going to get all of what looked like a juicy new
>market opportunity.
>
>The Model 29 IDP set looked like a conventional ASR of the day,
>except that there were two tape readers.  I worked on a system for
>the Western Electric works in Kansas City, and its application was
>for Purchasing and Receiving.  When a purchase order form was
>processed, a loop of tape, containing a program, was in one of the
>readers.  When the operator stepped on a foot pedal, it started up
>and printed in fixed stuff like the "bill to" and "ship to"
>addresses.  The current date could, if you wanted to, be printed
>from another loop of tape in the other reader made up on a daily
>basis.
>
>When it was time for the "items" in the purchase order, the reader
>would stop so that the operator could type them in by hand.  When
>finished, the operator would step on the pedal and the program tape
>would complete the form.  Key factor:  while all this was going on,
>the reperforator on the keyboard was punching out selected data from
>the PO, and this tape went with receiving's copy of the PO in a
>special pouch which included a plastic lined pocket for the tape so
>the oil would not seep through.
>
>Later, with the ordered stuff arrived in Receiving, the clerk would
>get out the original form and put the data tape in the second
>reader.  With a program loop tape in the first reader, the process
>was repeated, with the receiving clerk adding whatever information
>was needed when the program tape would stop for same.  When the
>job was done, the new tape punched by the reperf at the Receiving
>Department went to Accounting, where it went into some sort of
>computer to pay the bill for the shipment.
>
>The big problem here was legal and political.  The 1956
>(1953?) Consent Decree said that the Bell System must confine its
>stuff to "common carrier communications, and services incidental
>thereto."   This meant, in plain English, that the Bell System was
>to keep the hell out of data processing.  It was the direct result
>of the BSDP (Bell System Data Processing) project at Bell Labs, in
>which AT&T was fixing to build and market a main frame number
>cruncher targeted directly at IBM and Univac.  The Consent
>Decree changed the course of the BSDP project into a different
>direction, and it finally became the main gizmo driving the 1ESS.
>
>But when you read the Consent Decree, how can you justify the Bell
>System building Teletype machines as computer consoles?   The trick
>was twofold.  First, they were fundamentally start-stop telegraphic
>communications machines, vs. the Friden Flexowriter which was a
>parallel transmission office machine.  Second, the Teletype Model 29
>was not allowed to do arithmetic functions.   That is, it could not
>total up a column and print the results automatically.   Doing so
>would constitute data processing.  We lost lots of business to
>Friden, who could not only do arithmetic functions, but also
>type lower case.
>
>This concern not to be perceived as doing data processing was so
>strong that for several years the Touch Tone dials had only ten
>digits, without the * and # keys which are standard today.  The fear
>was that someone at the Justice Department would view a phone with
>the 12 key pad including the * and # as a data terminal, being
>marketed in competition with some early low grade key pad terminals
>made by IBM.   Western Electric made 16-key dials for Autovon, but
>the POTS user only got ten keys.
>
>The ASCII was a neat end run around the data processing issue, since
>the code was clearly designed for "information interchange", which
>suggests communications.  Teletype recoded the Model 29 to run on
>ASCII, and called in Henry Dreyfuss to restyle the case, and the
>Model 35 was born.  The Model 29 became instantly obsolete.
>
>So the Model 29 only ran for a couple of years, 1961-1963.  That PO
>system in Kansas City, which was provided by Southwestern Bell, ran
>at least until 1970, because I got occasional phone calls from the
>technician about it, even after I left Teletype in 1969.
>
>A final technical note.  The one inch tape included even parity, and
>they never made a printing reperforator nor a tape printer for the
>Model 29.   The case never said "Model 29" on it, and the only way
>to quickly tell if a "Model 28" is just that or a Model 29 is to
>look at the keyboard.  The 4-row keyboard was a Model 29.=
>
>Ben Stephens=
>
>
>


--
_______________________________________________
GreenKeys mailing list
GreenKeys@mailman.qth.net
http://mailman.qth.net/mailman/listinfo/greenkeys