Hi Tom: I can’t resist reminding you that your very firm “determination" about Colossus depends on exactly what is meant by “programmability” (and indeed “program”), which I claim are two rather slippery concepts. Recall that a little while ago I provided you an excessively lengthy formal definition of ”Program" in a 1973 Computing Surveys paper that I co-authored with the late Jim Horning, which I believe neatly muddies the water of your claim about Colossus! :-) However, I’m with you on the undesirability of having a “first computer” argument in this list - and suggest we similarly avoid having one about “programmability" and “program”. Cheers Brian
On 23 Sep 2020, at 19:51, thomas.haigh@gmail.com wrote:
⚠ External sender. Take care when opening links or attachments. Do not provide your login details. I don’t believe the ABC was sold or leased to anyone, so it wouldn’t be a good candidate if we think of these as “sales records.” But adding an entry for 1, which is admittedly a power of ten, would mean figuring out how to define “sales” and (far more challenging) how to define “computer.”
Professional historians have a de-facto agreement not to declare anything “the first computer” as the answer depends so much on definitions as to require the insertion of multiple adjustives between “first” and “computer.” So I expect to go through life without ever giving a specific answer to the question “What was the first computer,” which is sad as it’s the question we get asked a lot.
Early discussion of computer history centered on patent-driven disputes regarding the “first computer,” and even after the ENIAC patent was invalidated the disputes continued for some time in the pages of Annals. Even today, partisan armies continue the fight in the comments sections of Internet sites and the reviews of books on Amazon. The first professional historians to enter the field agreed a de-facto true, in which every early machine received some kind of consolation prize. In _ENIAC in Action_ we quote Mike Williams on this: “If you add enough adjectives to a description you can always claim your own favorite. For example ENIAC is often claimed to be the ‘first electronic, general purpose, large scale, digital computer’ and you certainly have to add all those adjectives before you have a correct statement.”
In those terms I am quite bold in saying ENIAC needs only two adjectives. Since determining (in work with Mark Priestley) that Colossus was not programmable (see http://www.tomandmaria.com/Tom/Writing/ColossusProgrammabilityWeb.pdf) I am now comfortable saying that ENIAC was the “first programmable electronic computer.” I’d also call it the “first general purpose electronic computer.” If you happen to think that those are the most important adjectives then you have an answer.
But may reasonable people (among them large numbers of Iowans, Germans, and Britons) stress other combinations of characteristics. There are a lot of questions to answer. Is it the method of programming that counts (the so-called “stored program”)? Would a program stored in ROM count, or only in RAM? Does the program have to automate loops and branches (i.e. control operations) as well as sequencing mathematical operations? Does a computer have to be electronic? What counts as electronic anyway, as all machines have some mechanical components? Does it have to work fully or be used for its intended purpose? What if nobody else saw it working? Does the computer have to be built at all, or just designed? What makes something a computer, versus an automatically controlled machine of another kind? You might think that that last question is easy: a computer carries out a computation, unlike a loom, music box, or automaton, but just this week I was corresponding with a computer scientist who believes that a mechanical musical box is carrying out a computation.
Thus any claim to define the first computer, or even the first computer sold or installed, requires many caveats and definitions. Too many for me to be able to propose it in good faith for recognition as a world record. In fact the Guinness editor I am working with noted a proliferation of “first whatever” computer-related records in their existing database, each first hedged with so many words as to be of little use to a general reader.
Note to list: this message is to explain to Brian and others in his position why historians have agreed not to argue about the “first computer,” not to start a flurry of messages pursuing such arguments. I think we’ve got through 15 years or so of this list without having the “first computer” argument and I’d like to keep that streak going.
Tom
From: Members <members-bounces@lists.sigcis.org> On Behalf Of Brian Berg Sent: Tuesday, September 22, 2020 9:04 PM To: SIGCIS Listserver <members@sigcis.org> Subject: Re: [SIGCIS-Members] Computer sales records
Thomas,
Your table "begs" for an entry with an installed base of 1. Earlier today I posted re: the ABC computer which was "built at Iowa State University (then called Iowa State College) from 1937 to 1942" - do you want to have that perhaps be the very first entry?
Brian Berg
On Tue, Sep 22, 2020 at 5:47 PM Allan Olley <allan.olley@utoronto.ca> wrote:
Tom, At the risk of getting lost in the weeds of "what is a computer", there is probably a case to be made that the Card Programmed Calculator as the first computer to have more than 100 sales/installations/rentals (and more than 10). A quick check of IBM archives tells me it had 20 installations in 1949 when it was inaugrated and 700 by the mid-50s. However there were at least two versions of the CPC and it is arguably a calculator and not a computer. It was programmable enough that things such as an arrangement to make it handle floating point were widely shared, on the other hand I get the sense some plugging was involved and it and limited programming flexibility. Someone who actually knows the nitty gritty of control for it and cares about a strict definition of what is or isn't a computer would probably be better qualified then me to judge the case. The sales numbers are clear the "computer" part very unclear.
-- Yours Truly, Allan Olley, PhD
http://individual.utoronto.ca/fofound/
On Tue, 22 Sep 2020, thomas.haigh@gmail.com wrote:
Hello SIGCIS,
I’ve been emailing back and forth with Allan and David, and we appear to have converged on a sense that a 400 lb roll might plausibly been around 4 feet in diameter and that photos of the SSEC show space for rolls of at least 3 feet in diameter, maybe as much as 4. So I am inclined to credit the testimony of Grosch as being at most a small exaggeration. If you look at one of the classic SSEC images at http://www.columbia.edu/cu/computinghistory/ssec.html, the tape rolls are the three large round things at the top of the machine on the back wall. Each punch was directly above 10 tape readers, and it appears that the tape emerging from the punch could be threaded through all 10 readers to create a buffer memory with 10 chances to read each piece of data.
So with that and the hard drive platter taken care of, here is my next batch of candidate records. I am trying a “power of 10” approach to computer sales (or in the early days, leasing) records with the first machine to hit each milestone. This has the advantage that I don’t need to try to figure out exactly with of the many estimates of CBM 64 and iPhone 6 sales are accurate, just satisfy myself that they were the first to exceed the threshold.
Installed base in excess of…
first achieved by…
during sales years.
10
IBM 701
1952-1954
100
Probably IBM 650
1954-1962
1,000
IBM 650
1954-1962
10,000
IBM 1401
1959-1971
100,000
TRS-80 Model 1
1977-1980
1,000,000
CBM VIC 20
1981-1985
10,000,000
CBM 64
1982-1994
100,000,000
Apple iPhone 6
2014-2018
Points of possible uncertainty:
1: was the IBM 650 the first to 100 as well as to 1,000? Seems likely.
2: At the time that Commodore claimed the VIC 20 as the first million selling computer, claims were also made for the ZX81 as the first computer to sell a million. If anyone wants to make a case for Sinclair I will listen.
3: In case any Apple fans are about to speak up, re the idea that the TRS-80 was first to sell 100K, let me explain my logic preemptively. I am going from a combination of this advertisement from Computer World 18 Oct, 1979 celebrating 100,000 sold:https://books.google.com.au/books?id=UaKuzwnEiRMC&printsec=frontcover&source=gbs_ge_ summary_r&cad=0#v=onepage&q&f=false and these plausible seeming estimates of Apple II sales by model(https://retrocomputing.stackexchange.com/questions/459/over-its-lifetime-how... pple-ii-computers-were-sold), which suggest only 65K sales of the original Apple II. As Radio Shack must have written the copy somewhat prior to the magazine date, say in mid-September it is had to imagine that the Apple II Plus, introduced in June 1979, sold 35K in its first three months to make up the difference (actually more than 35K sales would be needed, to cover however many of the original model were sold between Sept 1979 and its withdrawal in 1981). Apple II sales did not peak until well into the 1980s, so although the Apple IIe eventually sold millions I believe the VIC 20 got to the one million milestone first.
Best wishes,
Tom
_______________________________________________
This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. Opinions expressed here are those of the member posting and are not reviewed, edited, or endorsed by SIGCIS. The list archives are at http://lists.sigcis.org/pipermail/members-sigcis.org/ and you can change your subscription options at http://lists.sigcis.org/listinfo.cgi/members-sigcis.org
This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. Opinions expressed here are those of the member posting and are not reviewed, edited, or endorsed by SIGCIS. The list archives are at http://lists.sigcis.org/pipermail/members-sigcis.org/ and you can change your subscription options at http://lists.sigcis.org/listinfo.cgi/members-sigcis.org
— School of Computing, Newcastle University, 1 Science Square, Newcastle upon Tyne, NE4 5TG EMAIL = Brian.Randell@ncl.ac.uk PHONE = +44 191 208 7923 URL = http://www.ncl.ac.uk/computing/people/profile/brianrandell.html