[Telecom] TTY 33 and 35 case and cover composition?

On the very earliest computers it was standard procedure to run a job through twice and compare the answers to detect for such errors.

But I believe fairly early on they inclued internal parity checks for memory moves and arithemetic to check for that sort of thing. The first computers so equipped would simply stop and light a red light. Later computers would have internal retry capability and better circuitry.

At a PPOE it did. We had a dedicated box (made by Datatronix, SN00010 or something like that, with all the bugs it implies) that was attached to the TWX line and interfaced with a Displaywriter (with the comm package). It had enough memory so that it could receive and store TWX, for access the next morning when the WP gear was fired up.

I would have sworn we purchased TWX service from RCA rather than WU, but maybe I misremember that (this would have been around 1975). Perhaps RCA was later and involved some sort of gateway accessible via dialup. Telex/TWX was really the only halfway reliable way to communicate with offices in the third world.

Dave

I mention it because it goes to the heart of the the question: WU's policy vis-a-vis ASCII wasn't a factor. What put WU out of contention in the data market was its unwillingness to allow anything but TWX or TELEX machines to attach to its switches. As I mentioned in a previous post, I had an Anderson-Jacobson 841, which used EBCD code at 137.5 baud, but Tymnet allowed me to connect it to other nodes without trouble, doing code and speed conversion transparently. The only problem Tymnet couldn't solve was the need to constantly feed paper into the Selectric, since it had a friction-feed platen that wouldn't pull from a roll.

In any case, TWX machines were always half-duplex, and the bulletin boards that started the computer-to-computer ball rolling were all running full-duplex. I know, because I used to take the TWX machine (a Model 35, BTW) at Back Bay in Boston, and connect to Ward Christensen's bbs in Chicago: I think his was the only one that could work with a half-duplex machine. Ward was so amazed when I finally got a 300 baud modem and my own computer, he broke into my login session with the words "Hey, speedy!".

You're right: it was the combination of everything that made everything else possible. People bought computers, and in the process acquired CRT's and all the circuitry needed to drive modems and supervise online data transfers.

The bbs was the talisman that launched the Internet age: many of them grew into ISPs and Internet portals as we know them today. I had accounts at The Well, The World (and even traded emails with he-who-greps), and Northeastern University's Lynx system, which lasted until just a few years ago after a thirty-year plus lifespan.

I only used FidoNet a couple of times, when my cousin was using it to get emails back to me from his Army base overseas, but FidoNet might be the "special sauce" that let people realize inter-computer message networks _were_ viable. Of course, FidoNet was self-defeating, since as soon as it created the "critical mass" of customers willing to pay for bandwidth, the Internet was used to serve the demand and FidoNet wasn't needed anymore.

It's been fun walking down memory lane. I think I'll leave Western Union and its demise to the historians.

Bill

The first IBM PC I purchased in early 1980s was over $3,000. Two 360k floppy drives, 32k of memory, and an EGA color display. Top of the line.

Steve N2UBP

In defense of programmers, I will testify to the tough choices sometimes forced on them by legacy applications and by printed forms that never have enough space.

For the benefit of the non-BLUEBLOODS in the audience, a quick explanation: due to reasons known only to the ghods-of-Amronk, it is possible to print a numeric field with three kinds of sign.

The first, of course, is the ordinary "+" or "-", e.g., -900.00 for an amount of negative Nine Hundred.

The second is to "overpunch" one of the digits, so that its value combines both a digit and a sign, e.g., "900.0A" might mean +900.00, but the precious printing position taken up by the sign has been saved, so that values above 999.99 can now be printed.

The third method is to overpunch one of the digits so that it shows non-standard values _only_ if the sign is negative, e.g., "900.0a" would signify a negative sign, but a positive value would show as "900.00". Again, since the position taken by the sign has been freed, values less than -999.99 can be shown.

(These are only examples: I don't remember the actual overpunch digit inidcators).

Overpunching is, to be sure, a compromise: it typically happens when the totals shown on reports grow too large for the existing forms, and programmers are told to add more digits to a printed output than they have available space for. Given a choice between using an unsigned field which can only take positive values, or overpunching a sign in the least-significant-digit, experienced programmers will overpunch a negative sign. After all, most totals don't come out with negative values, but those few times that they do, the users will toss it off as a printing error, and (more importantly), the programmer won't get one of the dreaded 3 AM abend calls we all grew to know and hate.

Bill

It seems to have been government policy to keep WU with exactly one foot in the grave at all times. Oslin also argues government favoritism toward ITT.

Or didn't dare to rock the boat. W.U. could have handled voice over its microwave network, could have done essentially what MCI did. But MCI had to go through a bruising court fight to get Bell to connect to their customers.

As Oslin poiints out, charging telegrams to telephone bills was something that originated back when AT&T had acquired a controlling interest in W.U.

We can only speculate what would have happened if AT&T had not been forced to divest its W.U. stock shortly afterward. We might have had an integrated and rational voice and record communications business in this country.

If memory serves correct, pardon the pun, 32K was only available on the shortly lived cassette model. 64K was standard, and I believe required, for the diskette based versions with the top of the line being 128K on the mother board. The EGA display was much later, it was announced along with the AT.

My Osbourne Executive (circa 1982) cost me more than that! (I still have it at the bottom of a cupboard - it still boots!).

The amount of computing equipment I could get today for the same value of cash is staggering.........

Certainly seems that way. ITT came under scandal in the late 1960s.

In the 1960s, federal and state regulators were generally very happy with the Bell System/AT&T as it was. Their overriding goal was universal service, meaning cheap entry-level phone service for all. Official policy was to cross subsidize, so that high volume, very profitable toll lines paid the same rate as low volume, high expense toll lines, and premium offerings cross-subsidized bare bones service. This plan worked because over the years the percentage of Americans who had telephone service continually grew upward.

Generally, the regulators were opposed to outside "competition" because they wanted no cream-skimming; cream-skimming would wreck the cross subsidy model. Later, MCI came along and used the court system to overturn official policy and let MCI skim the cream; to get the good without the obligations for the bad.

It seems, from Oslin's book, that the Federal Govt expected WU to meet all sorts of "public obligations" which were very expensive. But WU didn't have the [highly profitable other businesses] in which to do it.

Oslin also admits WU made business decisions along the way, which is why I suspect WU missed the boat in the 1960s.

I can't help but wonder if AT&T would've given WU as hard a time as they gave [MCI]. MCI was obviously a major threat to AT&T, [and] AT&T saw them for what they really were - ... cream-skimmers, playing by free market when to their advantage, playing [as] a regulated [public utility], when to their advantage. WU, in contrast, was a mature company seeking a much more limited area (MCI claimed to want to offer ONLY private line point- to-point service in ONE corridor, but then promptly changed its plans to be switched in multiple corridors. WU would've been strictly private line).

The WU Tech Rev has an article on WU's private line voice service. How extensive it was I have no idea. In the 1960s very large companies were making use of AT&T "SCAN" networks which were private lines switched by the commercial network. (In a business, a centrex telephone set would have two exchanges, one the internal, one the external. To call another telephone, one would dial 8, the internal exchange, and the 4 digits; this would route the call to the location and telephone.) I suspect these SCAN networks, which allowed all phones in a business to participate, were more desirable than a WU special phone line.

Also, I strongly suspect Bell's local terminal connections for WU were not voice grade, but lower grade narrow bandwidth lines suitable only for low speed teletype (per Stone's book). That makes it less of a threat.

Given AT&T's record of public service, I'd say we would've had a superior and more cost effective record communications network, and improvements in voice communications, too. In the old days, piggybacking teleprinter signals over a voice line was very efficient due to different bandwidths, so a better use of the old network would've been easy. Data communications evolved over the 4k bandwidth of voice lines, but a separate pulse network could've evolved designed best for data. I suspect data communications, even high speed high accuracy digital transmission lines would've been developed much sooner as an integrated outgrowth of record communication needs. Cheap cost effective (albeit slower) facsimile may have been available sooner than it was.

Of couse, the flip side is this is that AT&T would've been an even bigger company than it was. Over the years, a great many politicians, regulators, and activists deeply resented and attacked AT&T _solely_ on account of its bigness. As a defense to the general public, AT&T could say "we don't do everything, there is, after all, Western Union" and take some of the heat off.

Actually, one major complaint was AT&T's ownership and sole-source through Western Electric, something AT&T would not give up. But I think in reality AT&T's real strength came out of its national network and ownership of all the phone companies, not Western Electric, and it could've safely divested WE (not Bell Labs) and still gotten high quality gear. I'll note, though, one thing WE did do for AT&T which was to provide a layoff base; during lean times WE would lay off workers when regular Bell companies were very hesitant to do so. Anyway, a divested WE could've thrived not only making the best telephone gear, but also using its electronic skills to make other high quality products for industry.

I'm sure they would have, and WU was already very dependent on AT&T for local loops for Telex and Desk-Fax.

I doubt that. Bell ran wires all over town to provide telephone service, and would use the same wires for W.U. loops. Now they might have artificially reduced the bandwidth; but then for Desk-Fax you just about need full voice bandwidth. (Used to be you could rent private lines for burglar alarm service very cheaply. Some hobbyists around Chicago discovered that and used burglar alarm lines to set up a private voice and TTY communication system. When Bell found out what was going on they shunted 2uF capacitors across all the burglar alarm lines, making them useless for voice transmission but still good enough for burglar alarms.)

I reread the WUTR article and found that:

WU would be using ASCII for its new AUTODIN system for the US Dept of Defense and Advanced Record System for the General Records Administration.

Teletype Corporation expected half machines to be ASCII in a few years.

ASCII was developed as a compromise code between messaging and data processing needs.

WU right then anticpated considerably more computer use.

I recommend readers interested in this subject check out the history. Those who are technically minded can judge where WU stood at that period of time compared to industry in general and the Bell System. Some of the questions we've been asking in the recent discussion are answered.

The articles begin at:

For WU data processing see:

Not typesetting; that had its own 6-bit 2-shift code, rather unimaginatively named TeleTypeSetter aka TTS. And its format controls aren't TAB and BS etc., but rather things like 'quad left' -- and officially the end-line code was 'elevate', because that's what actually happened on a Linotype, although by the '70s when I was involved most people called it 'return'.