This is the talk page for discussing improvements to the Teletype Model 33 article. This is not a forum for general discussion of the article's subject.
This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing articles
I recall using an ASR33 in high school. It was in a classroom and had a handset cradle 110 baud modem. I do not recall the connection between the ASR33 and the modem. I do remember dialing the number, hearing the beep, and putting the handset in the cradle. We then connected to a Honeywell minicomputer running GCOS. I would guess the connection to the ASR33 was RS-232.
Karstdiver00:00, 29 July 2007 (UTC)[reply]
I've got a set of ASR manuals somewhere, can anybody comment on whether Teletype is sufficiently defunct that it would be permissible to include a scan to illustrate the internal complexity? MarkMLl22:42, 10 September 2007 (UTC)[reply]
I'm no expert, but I date back to the says when both 20ma current loops and RS232 were common. Many DEC devices in the PDP11 era supported both. Anyway, if I remember correctly, the intended advantage of the current-loop interface is that a current loop is _much_ more resistant to electrical noise than is a simple voltage interface. This was more important ion the old days when a computer would span several equipment bays, and where a remote device might be separated by a goodly distance from the host, and almost certainly on a different electrical feed. I don't remember the details, but the operational distance limits on a 20 ma current loop (e.g. on an electrically noise factory floor) were much greater than RS232.
These details don't belong on this page, of course, but I do question the implication that computers had 20 ma interfaces merely because that was what an ASR33 used.
The sentence "In the photograph, the two holes appearing on the upper right side of the teletype are for an optional acoustic coupler to an internal modem." is not valid because (A) there are three pictures, rather than one implied by the phrase "the photograph"; and (B) none of the three photos depict an ASR33 with an installed acoustic coupler. In many places, this is called "ironic".
I am not sure where the confusion came from but Teletype Corporation documentation including all manuals, marketing material, user guides call this a Teletype Model 33 ASR and not an ASR 33. Why is the name inverted on Wiki? Wa3frp (talk) 13:01, 29 July 2011 (UTC)[reply]
Because it was inverted in real life long before the invention of Wikis, HTTP and HTML, and even TCP/IP? :-) For better or worse, "ASR-33", "KSR-33", "ASR-35", "KSR-35", etc. were used in the computer industry, the official Teletype name for the machine nonwithstanding. I guess a lot of computer people thought "Model 33 ASR" was a mouthful and just used "ASR-33", whether or not they even read the manuals from Teletype, just as, for example, people referred to the "360/30" or "360/40" or "360/50" or... rather than the "System/360 Model 30", "System/360 Model 40", "System/360 Model 50", etc..
It's certainly OK to fix the terminology in Wikipedia pages, and perhaps even appropriate to rename this page (leaving behind, of course, the redirects, so somebody can look for "ASR-33" and find the "Teletype Model 33 ASR" page). Guy Harris (talk) 23:42, 31 July 2011 (UTC)[reply]
Thanks for your time and your interest in commenting. I'm going to wait to see if anyone else has a more detailed explanation. I'll hold off rewiring this article until then. My thought is the Teletype product was repackaged and renamed by a third party as a part of a larger system. I don't agree with the example explanation that the name was a handful since your example does not easily fit. We didn't end up calling the IBM System/360 Model 30 a "30/360", we called it a "360/30". Many people in my era of teleprinter use, 1960s, 1970s and 1980s did call the Teletype Model 33 ASR a "33 ASR" but not an "ASR 33" or an "ASR 33 Teletype".Wa3frp (talk) 01:55, 1 August 2011 (UTC)[reply]
The example was somewhat of a guess; I definitely remember people talking about "ASR-33"s from long before the Web, much less the Wikipedia, existed. Googling for it finds a bunch of hits, many of which come from people whose involvement in the computer field antedates the WWW (as mine does), much less Wikipedia; {ASR,KSR}-xx might be a slang term, but plenty of slang terms are pretty common. (As for S/360 Model xxx, perhaps a better example would be calling the IBM System/360 Operating System "OS/360".). Guy Harris (talk) 02:28, 1 August 2011 (UTC)[reply]
Better to use the original name from Teletype Corporation in an encyclopedia. I think that you agree. I'll tackle the task shortly. Waiting to see if anyone else comments.Wa3frp (talk) 03:49, 1 August 2011 (UTC)[reply]
Wikipedia has a preference for secondary sources over primary sources. See Reliable_sources. A book or magazine by an independent publisher has more weight than promotional material created by a company. Your proposal to use "the original name from Teletype Corporation in an encyclopedia" does not follow policy or style guidelines. In general things should use the most common name, not the formal name from the company's spec sheet. Biographies on Wikipedia use the person's common name for the article title and typically use the formal name once in the lead paragraph.
Many product articles use the Company Name followed by Model number, so this article could be named Teletype ASR-33. However current title was commonly used in the 1970s. You changed ASR-33 Teletype to Teletype Model 33 ASR in Micro Instrumentation and Telemetry Systems. [1] MITS sold the product as Teletype ASR-33 as seen in the advertisement in the Altair Basic section. The Computer Notes newsletter shown in marketing section has this paragraph. "The MITS-MOBILE is a camper van equipped with an Altair BASIC language system. Included is an Altair 8800, Comter 256 computer terminal, ACR-33 teletype, Altair Line Printer, Altair Floppy Disk and BASIC language." (Typo in newsletter, ACR should be ASR.)
The phrase "Teletype Model 33 ASR" is an uncommon in references and the word "Model" adds nothing to the sentence. It reads overly formal. There in no reason the link to a stub Teletype article. It is very tangential to MITS article. -- SWTPC6800 (talk) 18:34, 1 August 2011 (UTC)[reply]
Sometimes Wikipedia's preference for secondary sources over primary sources ventures into the silly. If Teletype sold the Model 33 as the "Teletype Model 33 ASR", that would be the official name of the product, regardless of the fact that other publications may have called it the "ASR-33"; this is different from, for example, Teletype claiming in their promotional literature that the Model 77 prints 1200 characters per second and an independent product reviewer finding that it only prints 700 characters, which would be a legitimate place to prefer the secondary source.
Whether, in the computing world, it was more commonly known as the "ASR-33" than the "Model 33 ASR" or the "33 ASR", which may have been the case (perhaps in the telecommunications field, with Teletypes used for Telex/TWX service or other human-to-human messaging, the official names were used), is another matter. If so, perhaps calling it an "ASR-33" in articles that mention the Model 33's use with computers makes sense.
As for the link to the stub page, the fact that it's a stub page now doesn't mean it'll remain one in the future - it could probably use a lot more historical material - but, as the page for the ASR-33 links to the page for the Teletype Corporation, perhaps just a link to the ASR-33 page suffices. (At least some other pages refer to the Lear SieglerADM-3A, in a similar fashion.) Guy Harris (talk) 19:03, 1 August 2011 (UTC)[reply]
It appears that Teletype referred to their terminal by a variety of names in their 1974 advertisements; "ASR terminal", "model 33" and even "33". The June ad has this sentence: "Some improvements make the 33 more dependable and versatile." The word "model" is always lower case. A survey of books and magazines from the 1970s and 1980s show that "ASR 33" is a very common name. Many books used Teletype in lowercase and as a generic term. Magazines wanted to keep advertisers happy so they were more careful with trademarks. The ads have several reminders that Teletype is a registered trademark. -- SWTPC6800 (talk) 20:24, 12 August 2011 (UTC)[reply]
Teletype is no longer a registered trademark. The US Patent and Trademark Office records state "This registration was not renewed and therefore expired." This status change happened on December 15, 2000. Serial Number: 73157988, Registration Number: 1119161. The last renewal was on May 29, 1979. The trademark was transferred to AT&T on July 12, 1985 and they never renewed it. The Teletype T logo has also expired. -- SWTPC6800 (talk) 21:08, 12 August 2011 (UTC)[reply]
Keyboard send receive is a single paragraph stub definition that essentially duplicates the definition given here. The article should be merged to give it context; the reader gets no benefit by clicking on the link to read essentially the same information as is given here. --Wtshymanski (talk) 15:03, 18 March 2012 (UTC)[reply]
Disagree - not all KSRs were Model 33s; perhaps there needs to be a page about Teletypes (the teleprinters made by Teletype Corporation) that mentions that many of them came in three flavors, describing the three flavors, and perhaps enumerates the models. "Keyboard send receive" could redirect there. Guy Harris (talk) 17:49, 18 March 2012 (UTC)[reply]
There is a range of potential options here and I agree that this stub should be merged elsewhere. The issues is where? I suggest that some new articles are created:
The following discussion is an archived discussion of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.
ASR-33 Teletype → Teletype Model 33 ASR –
The existing article is misnamed. Specifically, the manufacturer, Teletype Corporation, called this teleprinter a Model 33 ASR. The earliest known source for this Teletype Corporation equipment naming discrepancy comes from Digital Equipment Corporation documentation where the September 1963 PDP-4 Brochure calls the Teletype Model 28 KSR. This naming discrepancy continued from the Teletype Model 28 to other Teletype equipment in later DEC documentation. Biglulu (talk) 08:05, 23 March 2012 (UTC)[reply]
It seems to me that Teletype Model 33 would be a more appropriate name, since the ASR was just one variant, the one including a paper tape reader and punch. The tape-less KSR was also common. --agr (talk) 11:47, 23 March 2012 (UTC)[reply]
Yes, the existing article is misnamed! Either Teletype Model 33 or the more specific Teletype Model 33 ASR is more encyclopedic that the current article. It is time to undo the naming error created by DEC in 1963.13:14, 23 March 2012 (UTC)
The above discussion is preserved as an archive of the proposal. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.
I am deleting the claim that the Model 33's 127 (Delete) key assignment became a de-facto standard. First of all the key is labeled rub out and the code is 177. More importantly this code usage is build into ASCII by design. Having NULL be all 1's allows any paper tape equipment to erase characters by punching all holes. That's why the NULL character is not with the rest of the control characters in ASCII. Furthermore this practice goes back to Baudot code where the LTRS key punched all five holes, allowing it to be used to delete characters, since repeated LTRS characters were harmless. An operator who caught an error early while punching a tape could backspace the tape and then over punch with the same number of LTRS in 5-level or rub out in ASCII. Note that the Model 33 ASR punch unit has a B.SP button for that purpose (it simply connected to a lever that moved the tape back one unit). --agr (talk) 14:31, 9 April 2012 (UTC)[reply]
"Another 100,000 were made in the next 18 months, so that No. 600,000 was painted red-white-and-blue and shown around the country during the last part of, and the year after, the Bi-Centennial year."
If this is referring to the US's bi-centennial in 1976, that might be OK if Wikipedia is for Americans only. I am pretty sure some people in other countries would be interested in this article, and not everyone in the world is familiar with this date, which comes across as sort of Ameri-centric if not entirely obscure. America is not the only country enjoying electronic innovation, but the reference to its birthday celebration almost makes it sound as if it is the only country that matters w/r/t technology.
Unless someone can demonstrate some very straightforward technical or political connection between the 600,000 teletype machines produced up to that time and the American Bi-Centennial, I move to remove the phrase and replace it with something more generic, like "1976."
The connection was between the 600,000th machine in particular, not to the machines built prior to that, that machine being 1) the Nth machine built, with N an even multiple of 10^5, and 2) manufactured in the bi-centennial year, that being why it was painted the US's national colors and shown around the US. I've rewritten that sentence (which was copied-and-pasted from the reference) to note that it was 1976 and say United States Bicentennial and link to said page. Guy Harris (talk) 02:01, 20 June 2013 (UTC)[reply]
Thank you for your modification. For one thing, it flows nicelier, and it removes what sounded Ameritistical. It is much clearer. Sorry to make a fuss, but one thing I cannot stand is anything that sounds like breast-beating (not that you actually were, just that it sounded that way). Now, anyone from anywhere can enjoy the article and understand the significance of parading that machine around the country.
(You know, there are probably a lot of young American people today who even might not know what that date signified.) Thanks again, and hope you don't think I'm overly huffy.
Bozonius (talk) 07:03, 20 June 2013 (UTC)[reply]
I find the use of the word "level" puzzling. It reminds me of rock strata in geological formations, or floors in a building. I assume it refers to the unit which counts the number of possible positions of holes in each row (representing a keystroke, or character) on the paper tape, for which I'd consider "bit", "hole position", or simply "hole" more appropriate. 172.56.27.140 (talk) 13:43, 14 May 2014 (UTC)[reply]
Pilot program for the use of computers in education.
I just a wrote an e-mail to the Superintendent of the Palo Alto School District for Wikipedia citations for the first pilot program of the use of computers in education. We deaf kids at Greendell Elementary were connected to Stanford University via acoustic modems (!) to KSR-33 Teletypes which served up rather too rapid spelling and arithmetic lessons, circa very late '60s/early '70s. (I physically broke one of the machines by pressing the "Press to Start" button into the chassis when the system was slow one rainy day. I haven't played the innocent as well since.) Anyway, my father tells me that this was the first usage of computers in education, so I am seeking Wikipedia-worthy citations. kencf0618 (talk) 17:57, 30 August 2015 (UTC)[reply]
The Technical Information section states that "The Model 33 used the seven-bit upper-case only ASCII code". This must definitely be wrong. Either it used ASCII (aka IA5 (ITU T.50), aka ISO/IEC 646) which is 7 bit and has both upper and lower case, or it used some other encoding which was upper-case only. As I don't know which is actually the case, I cannot correct the information. --Lasse Hillerøe Petersen (talk) 04:34, 15 June 2019 (UTC)[reply]
The Model 33 used ASCII as its character encoding, but 1) the keyboard was incapable of generating the ASCII codes for lower-case letters (and possibly {, |, }, and ~) and 2) the printer was unable to print those characters - I'm not sure what it did if it received them. I couldn't find a user's manual that indicated what happened; this technical manual describes the internal workings, and seemed to support point 1), but it looked as if it'd be difficult, at best, to determine from the electromechanical description what would happen if a Model 33 received an 'a' character). Guy Harris (talk) 13:41, 15 June 2019 (UTC)[reply]
I rephrased it to say that it used ASCII, without speaking of ASCII as upper-case-only, but only a subset, lacking upper-case letters and some symbols. Guy Harris (talk) 20:14, 24 January 2022 (UTC)[reply]
I think your rephrasing was a mistake, based on Lasse's mistaken assumption, and has made the relevant section less accurate. The previous phrasing referring to "the seven-bit upper-case only ASCII code" was highly specific and technically correct. I would encourage you to revert or revise your edit. See my other reply to Lasse. I am however less confident about the correctness of the next part, which says said code was "also known as CCITT International Telegraphic Alphabet No. 5", which I think is false precision. My sources suggest that ITA-5/IA5 corresponded to a later version of ASCII that actually did include lowercase characters. —ReadOnlyAccount (talk) 08:12, 1 January 2024 (UTC)[reply]
It is not clear what "the seven-bit upper-case only ASCII code" means. Does it mean "the seven-bit ASCII code, which only incudes upper-case letters"? If so, then it does not refer to X3.4-1967, which includes lower-case letters; it could refer to X3.4-1963 (I don't know what X3.4-1965 looks like, so I don't know whether it could refer to that or not). Or does it mean "an upper-case-only version of the ASCII code", in which case it could refer to X3.4-1967 as long as you ignore the lower-case letters (and anything else where there's a difference between -1963 and -1967).
This 1963 technical manual for the Models 32 and 33 has a table of character codes, for use on answer-back drums, on page 2-5, which appears to be X3.4-1963, with an up-arrow for 0x5E, a left-arrow for 0x5F, and ESC for 0x7E. This is not too surprising, given that neither X3.4-1965 nor X3.4-1967 existed in 1963. :-)
This 1974 technical manual for the Model 33 has a table of character codes, for use on answer-back drums, on page 7 of section 574-100-201TC "Teletypewriter Set - Installation". It shows all of the -1967 version of ASCII, and, for both the -1965 and -1963 versions, shows characters for those codes the assignment for which changed (e.g., it claims that 0x5C is \ in the -1963 and -1967 versions but is ~ in the -1965 version. The keyboard table (Figure 3 - 33 Application of ASCII) on page 3 of section 574-121-100TC "Keyboard - General Description and Principles of Operation" shows 0x5c as \; it says of 0x7D that "This code can be generated on model 33 nonparity keyboards by depressing the ALT MODE key.", and it says of 0x1B that "The ESC control function may be generated by depressing the ESC key or by simultaneously depressing the K, SHIFT, and CTRL keys." Pages 5 and 6 of X3.4-1963 show 0x7D as an "Unassigned control"; they show 0x1B as "S3" (subscript 3) and 0x7E as ESC.
So perhaps the original Model 33 could be described as implemening the upper-case-only X3.4-1963 (again, not surprisingly), and later Models 33 implement an upper-case-only subset of X3.4-1967. Guy Harris (talk) 10:24, 1 January 2024 (UTC)[reply]
(Point of order: Instead of breaking up your above reply with further replies, I'll quote you and reply. That's repetitive, but overall hopefully neater and more respectful.)
It is not clear what "the seven-bit upper-case only ASCII code" means. Does it mean "the seven-bit ASCII code, which only incudes upper-case letters"?
Have a look at this section (of the ASCII article) I previously linked in my reply to Lasse. I know it's a lengthy table, but it's a nice columnated overview of all the characters in ASCII 1963, 1965, and 1967, respectively (i.e X3...etc.). I thought it somewhat obvious "the seven-bit upper-case only ASCII code" referred to ASCII 1963, but I take your point that a reader without preconceived notions might parse that description some other way. I parse it to mean "[of several variants,] the ASCII code that's upper-case only", i.e. ASCII 1963.
I think it might actually help to drop the mention of the seven-bit nature of ASCII from that sentence, since non-extended ASCII is seven-bit per definition, and bringing that up again there is only likely to confuse readers into thinking there was some link between 7-bit and uppercase-only, which there isn't.
So perhaps the original Model 33 could be described as implemen[t]ing the upper-case-only X3.4-1963 (...), and later Models 33 implement an upper-case-only subset of X3.4-1967.
I think your conclusion is correct. I hadn't actually looked at those different manual revisions. Good find. I suppose if you want to be a stickler for the best kind of correct(ness) and can think of a decent way to phrase that, you could make the article section in question reflect that (plus, of course, ix-nay the 5-ITAy shenanigans). That said, I also think there's such a thing as over-egging the cake, so caveat redactor, maybe. —ReadOnlyAccount (talk) 13:48, 1 January 2024 (UTC)[reply]
PS: For the record, page 7 of section 574-100-201TC "Teletypewriter Set - Installation" appears to be PDF page 75 of the 1974 technical manual for the Model 33. It's essentially the same as the aforementioned section of the ASCII article, even though the chronological column order is reversed and some of the content is considerably harder to read. The documented character changes from the mostly unassigned sticks 6/7 of ASCII 1963 to ASCII 1965 and finally 1967 are also discussed in the ASCII article and mentioned in some of my and your earlier comments here. The control codes shown in the PDF are also shown ahead of that section in the ASCII article – just scroll up. Finally, personally I find the 8th level info on PDF page 75 to be potentially confusing. The key appears to be that it's a parity of removed tines, i.e. not of binary value but of numbers of ones and zeroes, if that makes sense.
ASCII 1963. Note the absence of lowercase characters. Your assumption that "this must definitely be wrong" is itself wrong. The flaw in your argument lies in your thinking that ASCII was synonymous with ISO 646. That is broadly, but not strictly true. ISO 646 codified ASCII, but the ASCII standard itself had had a bit of history before it got homologated as ISO 646. There was a bunch of R&D that went into ASCII, and there were several revisions. As the ASCII article also shows, the 1963 version of ASCII did not yet include lowercase characters, but the subsequent 1965 and 1967 versions did.
Sadly, the father of ASCII is no longer with us, but if you want to know this stuff in excruciating detail, you could always readthumb through Charles E. MacKenzie's Coded Character Sets, History and Development, where Fig. 14.11 (PDF page 268) shows you that ASCII 1963 left sticks 6 and 7 mostly unassigned – but you'll be relieved to hear that chapter 14.10 ASCII, 1967 (ibid.) informs the reader that soon after ASCII 1963 and still in that same year, "[a]t the first meeting of ISO/TC97/SC2 in 1963 October 29–31, a resolution was passed that the lower-case alphabet should be assigned to columns 6 and 7."
By the way, the fact that old Teletypes couldn't do lowercase was also the reason some Unix-variants included uppercase-only operation support until well into the 2010s, perhaps so that gramps emeritus could still ssh into his alma mater from his retirement home using his trusty old ASR Model 33 Teletype. Would Bayesian spam filters still have delivered all-uppercase emails in the 2010s though? Press X for doubt. —ReadOnlyAccount (talk) 08:12, 1 January 2024 (UTC)[reply]
Half-duplex-only vs. half-duplex and full-duplex options
At one point the manual claimed that the Model 33 had a mechanical connection between the keyboard and print mechanism and only supports "half-duplex" (local echo) operation. Both a 1971 manual From Teletype, and an article in Kilobaud Microcomputing about the Model 33, say otherwise.
The 1971 manual from Teletype indicates that the connection is electrical, not mechanical, and that the user can configure the Model 33 either for half-duplex operation or "full-duplex" (remote echo) operation; most of DEC's operating systems expected the latter (allowing them to, for example, echo DEL by printing the characters being deleted, with sequences of deleter characters being printed between "/" and "/", so that typing "F" "R" "O" "G" DEL DEL DEL DEL "P" "R" "I" "N" "C" "E" produced "FROG/GORF/PRINCE"; the BSD tty line discipline also supports this, but prints them between "\" and "/", as do UN*Xes that followed its conventions). The article from Kilobaud also says that either is available, and indicates how to configure the Model 33 for the desired mode.
...After taking certain preliminary steps such as signaling the distant station, etc., an operator types the message on the keyboard of his Printer. The Printer at the distant station prints the message on page-width copy paper. A local copy is made by the Printer at the sending station for record and monitoring purposes.
(emphasis mine). This seems to suggest that remote-echo mode was not an option at that time.
Perhaps the Model 33 originally didn't allow remote-echo mode; that mode wouldn't be useful if a message was being sent from mode teleprinter to another, as the remote teleprinters wouldn't echo the characters back, and the user typing the message wouldn't be able to see what they've typed. Later popularit of the Model 33 as a computer terminal might have caused Teletype Corporation to make it an option. Guy Harris (talk) 07:06, 4 February 2022 (UTC)[reply]
Forgive the pun about who got switched on first, but I think it might be interesting to know whether or not the Model 33 actually preceded the Model 32. I haven't seen good sources on that.
In terms of etymology, it would also be interesting to know whether the name of the Model 32 was in any way related to the fact that this Baudot-compatible model basically used a base32-like encoding (if we disregard the figure and letter shift shenanigans). —ReadOnlyAccount (talk) 08:27, 1 January 2024 (UTC)[reply]
April 1974 ad
June 1974 ad
