Testwiki:Featured article candidates/Manchester Small-Scale Experimental Machine

The following is an archived discussion of a featured article nomination. Please do not modify it. Subsequent comments should be made on the article's talk page or in Wikipedia talk:Featured article candidates. No further edits should be made to this page.

The article was promoted by Karanacs 22:02, 24 March 2009 [1].

---

Manchester Small-Scale Experimental Machine

Nominator(s): Malleus Fatuorum

It is with some trepidation that I present this account of the development of the world's first stored-program computer at the University of Manchester. The machine was designed as a test-bed for an early form of computer memory and was only in existence for a few months before being further developed to become a practical computer, so some technical details are inevitably sketchy. I believe nevertheless that this article gives a comprehensive account of the SSEM's construction and the background to its development in 1948. Malleus Fatuorum 15:13, 28 February 2009 (UTC)

• • I'd like to withdraw this nomination, as I won't be around to deal with any issues that may arise. --Malleus Fatuorum 01:22, 10 March 2009 (UTC)

• As an explanation, as Malleus will not be around to address any concerns raised during the nomination, the Greater Manchester WikiProject would like to take over the nomination. On behalf of the project, I believe we can deal with issues that may be raised; I know that I at least have access to some of the sources used in the article. Thank for your time, Nev1 (talk) 01:40, 10 March 2009 (UTC)

• Ref formatting, dabs, and external links found up to speed.--TRUCO 16:27, 28 February 2009 (UTC)

• Support, well written article on an interesting and important subject. But it would be nice to know how long this computer operated for, what happened to it and whether it had the the reliability problems of so many pre silicon computers. Also there's a reference to tape, if the sources say whether this was paper or magnetic tape it would be nice to link it appropriately (I suspect from the date it would be the former). WereSpielChequers 17:12, 28 February 2009 (UTC)
  • Thanks for your support. The machine only existed in its completed form for a few weeks. Once it had proven the practicality of the Williams tube and the stored-program approach it quickly evolved into a prototype for the Manchester Mark 1, with bits being continually added and/or redesigned. I'll try to make that clearer in the article. With a few hunded valves there's every reason to believe that it suffered from the same problems as other valved machines, but there's no information on its reliability that I've been able to find. It has to be remembered that the machine was in constant development and it was never intended as a practical computer anyway; work on it never really finished, it just evolved into the Mark 1.
  • The reference to "tape" is in relation to the Turing machine. The SSEM had no tape; I/O was done by manual switches, setting the value of each word in turn. I've added a phrase to the brief description of the Turing machine to hopefully clarify that as well. --Malleus Fatuorum 17:41, 28 February 2009 (UTC)

Comments -

• You've mixed using the Template:Citation with the templates that start with Cite such as Template:Cite journal or Template:Cite news. They shouldn't be mixed per WP:CITE#Citation templates.
  • Fixed. I copied one citation over from another article, but forgot to change the citation template. --Malleus Fatuorum 21:02, 1 March 2009 (UTC)

Otherwise, sources look okay, links checked out with the link checker tool. Ealdgyth - Talk 20:40, 1 March 2009 (UTC)

Malleus, it's good, but you need to watch your sentence construction and related punctuation. Many people would kill to have such an easy-to-fix issue, rather than deeper, more problematic issues; but fix it you must. It does need massaging throughout.

• "This lead to the setting up of"—That's on the periodic table, is it?
• " Tommy Flowers and his team from the General Post Office's (GPO) Dollis Hill Research Laboratory were approached; but eventually turned it down due to other commitments, although they did build some mercury delay lines for ACE.[9] The semicolon (unless a boundary between items in a list) is normally followed by a grammatical sentence ("he" is missing). Not thrilled with "but" then "although".
• "Performing" rather than "doing", formally? (DSIR). I guess we need this density of initialism: "The Department of Scientific and Industrial Research (DSIR), who ran the NPL, pressed the TRE by assigning the highest priority to ACE in respect of all the work that TPE was doing for DSIR." Maybe.
• "and that others would work with"

"Early electronic computers were generally programmed by being rewired, or via plugs and patch panels. There was no separate program stored in memory, as in a modern computer; it could take several days to reprogram ENIAC, for instance." Try:

"Early electronic computers were generally programmed through rewiring, or via plugs and patch panels; there was no separate program stored in memory, as in a modern computer, and it could take as long as several days, for example, to reprogram ENIAC."

"EDVAC was under development at the University of Pennsylvania's Moore School of Electrical Engineering, Flowers and Wilkes had visited the Moore School of Electrical Engineering and attended a presentation on EDVAC; EDSAC was being developed at the University of Cambridge Mathematical Laboratory; and Professor Max Newman had moved to the University of Manchester and hoped to set up a calculating machine laboratory based on the use of the Selectron tube memory that was under development by RCA." Try this; there's a distressing comma splice, inter alia—

"EDVAC was under development at the University of Pennsylvania's Moore School of Electrical Engineering, which Flowers and Wilkes had visited and where they had attended a presentation on EDVAC [earlier in 1946?]. EDSAC was under development at the University of Cambridge Mathematical Laboratory; and Professor Max Newman had moved to the University of Manchester and hoped to set up a calculating machine laboratory based on the use of the Selectron tube memory that was being developed by RCA." [unsure about whether their visit to Moore was a single event ("to attend a")]

Is the use of so many "develops" OK? Maybe it's reasonable as a parallelism in this bit of the text, where I can't think of an alternative. But I see lots of this word everywhere. It's not always possible to substitute, but the usual ones are "devise", "create", "research", "discover", "construct", and "reveal".

• "and Williams was leaving in six weeks to take up a professorship"—"six weeks later". Tony (talk) 12:43, 12 March 2009 (UTC)

Reply

• Thanks for the helpful and encouraging comments. I've addressed the specific issues you raised, and I'll go through the whole article again later. In my defence, I didn't initially want so much background material, so when it was added I didn't really pay it much attention. That'll teach me. --Malleus Fatuorum 15:43, 12 March 2009 (UTC)

Comments by Fowler&fowler

• What is a "stored-program" computer? Can "stored-program" be wikilinked or briefly explained?
  • I've added a link to Von Neumann architecture. --Malleus Fatuorum 20:33, 17 March 2009 (UTC)
• "The machine was designed as a test-bed for the Williams tube, an early form of computer memory, not as a practical computer"
  • The unpractical part comes too late and confuses the reader somewhat. How about, "The machine was, however, not designed to be a practical computer (explain more though what practical means), but rather as a test bed for ...." (I've wikilinked test bed, but you might want to clean up that page a little as well, so that a reader understands the concept in your context.)
    • I've made some changes along these lines, although I hope you'll forgive me if I decline your suggestion to clean up the testbed article. --Malleus Fatuorum 22:02, 17 March 2009 (UTC)
• "It was in existence in its completed form for a just few weeks in the summer of 1948, as its success inspired its further development to become the Manchester Mark 1, ..."
  • This is slightly confusing as well. Here too the effect seems to precede the cause. How about, "Since its success quickly led to the development of the world's first commercially available general purpose computer, it remained operational for just a few months in the summer of 1948." (I think it is understood that it was the finished form. Either don't mention the names of the successors or mention them in the next sentence.)
    • Hopefully I've succeeded in addressing this now. --Malleus Fatuorum 22:26, 17 March 2009 (UTC)
• Last paragraph: "The SSEM had a 32-bit word length and a main store of 32 words. Three bits were used to hold the instruction code, giving a maximum of eight instructions, of which only seven were defined. Three programs were written for the SSEM, the first of which, to calculate the highest factor of 2¹⁸, consisted of 17 instructions and took 52 minutes to reach the answer after performing 3.5 million operations."
  • This is too dense for a general reader. It needs to be expanded in accessible language to at least twice its length. The general reader needs to understand (a) What do these statements mean and (b) why they are significant. For example, what is involved in computing the highest factor of $2^{18}$? A reader might ask: "Since no factor can be higher than $2^{17}$, why couldn't the computer simply check that 2 divides $2^{18}$ and declare the quotient to be the highest factor?" This sort of question should be anticipated in the lead.
    • Well, it is of course self-evident that $2^{18}$ is divisible by 2, as you suggest, and that its highest factor must be $2^{17}$, but that's why the problem was chosen, because the expected answer was known in advance, and so it could easily be seen whether or not the computer actually worked. Why didn't the program simply calculate the value of $2^{17}$ and produce that as the result? Two reasons; first there was no floating point or multiplier unit, not even a binary adder, only a subtracter, so the calculation could only have been done by repeated addition simulated by subtraction. Secondly, and perhaps more importantly, the program had to run for sufficient time to exercise the Williams tubes and assess their reliability, so even if the machine had had some kind of multiplication unit the program would not have proven the machine's reliability, its primary purpose. I'm not sure how much of that needs to be said in the lead, but I'll try to clarify that the purpose of the program wasn't really to produce the answer it did but to exercise the computer. I'm reluctant to get into why the computer didn't need to have an adder in the lead though. As for the word-length and so, I think in these days of 16 and 32-bit Windows PCs most general readers have at least some understanding of what that kind of thing means, wouldn't you agree? Perhaps the size of the instruction code could be dropped, as that's maybe not quite so accessible. --Malleus Fatuorum 20:33, 17 March 2009 (UTC)

I will read the remainder of the article later today. It is the kind of article I would like to support, but my task will be made easier if the lead is more accessible to a general reader. Will be back later with more comments. Fowler&fowler«Talk» 19:50, 17 March 2009 (UTC)

Thanks for your helpful comments. I've replied in detail above, and hopefully I've succeeded in addressing your concerns. --Malleus Fatuorum 22:26, 17 March 2009 (UTC)

Response by Fowler&fowler (Oppose for now, but looking to support)

Thanks, the lead is much better! Here are some issues I noticed in the next section. I am offering examples from only the first paragraph,

• In 1936, mathematician Alan Turing defined a theoretical "universal computing machine", a computer which held its program on tape, along with the data being worked on.
  • Do you mean Turing defined a UCM to be a computer which held ...? If not, then why mention it? Also, it might be helpful, if "theoretical" were replaced by "theoretical construct" or "theoretical concept." Also, "tape" will likely not be clear to a general reader. Please explain a little more and please provide a link or two Turing machine gallery or Turing machine#Information description. The "main" link, History_of_computing_hardware, unfortunately, is not very useful.
  • Here, for example, is a sentence from Britannica's page on Turing machines: "The Turing machine is not a machine in the ordinary sense but rather an idealized mathematical model (linked) that reduces the logical structure of any computing device to its essentials." So, other tertiary sources do take the trouble of making these distinctions.
    • I've reorganised and made some changes that will hopefully make it clearer that the Turing machine was a thought experiment, not a computer. --Malleus Fatuorum 20:22, 18 March 2009 (UTC)
• Turing proved that such a machine was capable of solving any conceivable mathematical problem for which an algorithm could be written.
  • This too needs to be clarified. A general reader will click on algorithm and read "finite sequence of instructions." From this they will conclude that if a solution can be found and written down in a finite sequence of instructions, then the UCM will find it. They will then ask, "If I already know the solution, why do I need the UCM?" So, something needs to be explained here (perhaps a simple explanation of computable functions).
    • I don't understand this point. An algorithm isn't a solution, it's a method for achieving a solution. --Malleus Fatuorum 20:22, 18 March 2009 (UTC)
      • I mean they will ask, "How will you know that an algorithm can be written if you haven't actually found one? And, if you have found one, you already have a solution." Fowler&fowler«Talk» 21:20, 18 March 2009 (UTC)
        • I'm afraid that I still don't understand. You don't "find" algorithms you define them, as a sequence of steps to be followed. Is the question "How do I know that an algorithm can or can't be written to solve a particular problem?" If it is, then the answer is that ante hoc (dog Latin I know, but it hopefully makes the point) I don't, but I'm failing to see why this is relevant to a general overview of the Turing machine. --Malleus Fatuorum 00:03, 20 March 2009 (UTC)

Are you essentially saying that a Turing machine can implement any algorithm? (I was under the impression you were saying that there is a theoretical notion of computability (a kind of existence theorem) and if a function is computable in this sense, then a Turing machine can evaluate it using an algorithm.) Fowler&fowler«Talk» 00:19, 20 March 2009 (UTC)

No, that's not quite what I'm saying. I could, for instance, write an algorithm describing how you arrive at work after your alarm clock goes off in the morning. Indeed in my programming days that was one of the common aptitude tests for computer programmers—did you remember to open your eyes before getting out of bed? But clearly such an algorithm isn't executable. I'm simply saying that if an algorithm to solve a mathematical problem can be written, then that algorithm can be executed by a Turing machine. --Malleus Fatuorum 00:39, 20 March 2009 (UTC)

That I understand and I believe we've made some progress! Thanks for your patience. So, wouldn't it be better to say (what you just said), i.e. "Turing showed that if an algorithm can be written to solve a mathematical problem, then a Turing machine will execute that algorithm." than, "Turing proved that such a machine was capable of solving any conceivable mathematical problem for which an algorithm could be written."? Fowler&fowler«Talk» 01:26, 20 March 2009 (UTC)

OK, you've worn me down. I'll rephrase as you suggest. :-) --Malleus Fatuorum 02:13, 20 March 2009 (UTC)

• "During the 1940s, Turing and others such as Konrad Zuse developed the idea of using the computer's own memory to hold both the program and data, instead of tape, ..."
  • This needs to be explained more. At the point the reader thinks that everything is theoretical. So, they will likely be confused by "memory" as a theoretical construct. Something needs to be clarified.
    • I've reworded to make it clearer that by the 1940s there were real, physical computers, as opposed to the conceptual Turing machine discussed earlier. --Malleus Fatuorum 20:22, 18 March 2009 (UTC)
• "... but it was mathematician John von Neumann who became widely credited with defining that computer architecture, still used in almost all computers.
  • Same here. The expression "computer architecture" suggests something in real life which occupies space and volume. Three things here would not be clear to a general reader: (i) why was Turing and Zuse's contribution an example of computer architecture? (ii) why is "computer architecture" a theoretical concept? and (iii) why is von Neumann being mentioned if nothing explicit is being said about his contribution? For example, the text does not even refer to the accompanying figure.
    • I don't agree that the term "architecture" suggests something in "real life", or at least not anything that would not be understandable to the general reader. What's the architecture of the house that you live in, for instance? Do you live in the architecture or in the house? --Malleus Fatuorum 20:22, 18 March 2009 (UTC)
      • This is resolved now since the reader knows we are talking about the architecture of real machines. Fowler&fowler«Talk» 21:20, 18 March 2009 (UTC)
• "Early electronic computers were generally programmed through rewiring, or via plugs and patch panels; there was no separate program stored in memory, as in a modern computer, and it could take as long as several days, for example, to reprogram ENIAC."
  • This sentence should occur much earlier if it relates to the current paragraph, or it should be moved to the next paragraph. Also, something should be said about ENIAC, such as, "... which had become operational in 1946."
    • I've move that up as well as making the addition you suggest, and integrated into it into the transition from theoretical machines to practical ones earlier in the paragraph. --Malleus Fatuorum 20:22, 18 March 2009 (UTC)

I have quickly looked through the other sections and I see other explanatory gaps. Please try to go through the other sections and read them from the perspective of a general non-expert reader. Please understand that I think this is a very worthwhile topic and that your syntax is fine; however, the explanatory narrative needs to be improved. If more time is needed than is available on FAC review, the FAC could be restarted or the decision be held off for another week (I believe the problems could be resolved in less than a week). Regards, Fowler&fowler«Talk» 17:00, 18 March 2009 (UTC)

• Not being a "general non-expert reader" I am not able to put myself in that position. The whole article seems blindingly obvious to me, painfully so in some places. --Malleus Fatuorum 20:22, 18 March 2009 (UTC)

• Thanks once again for you comments. It is indeed regrettable that the supporting articles on historical computer hardware—and indeed on computers and computing in general—are so poor overall, but I'm keen to avoid over-compensating for that by including in this article explanations that really ought to be elsewhere. Nevertheless I shall take what you say on board before I reply to the detailed points you raise above. --Malleus Fatuorum 17:51, 18 March 2009 (UTC)

The first paragraph reads much better! I've responded to two of your comments above. I will add the remainder of my comments on the talk page of the article, but will leave a link (or note) here each time I do. Thanks for making the changes. I know it is a chore, but it will be very helpful to the reader. Fowler&fowler«Talk» 21:20, 18 March 2009 (UTC)

I've left my first set of detailed comments at F&f's critique - Background. Will keep adding more as and when I find time. The background section needs a lot more explanation. Fowler&fowler«Talk» 23:58, 18 March 2009 (UTC)

• Thanks again. I think those issues have been dealt with. Hopefully you feel that the Background section now provides at least satisfactory explanations. --Malleus Fatuorum 16:00, 19 March 2009 (UTC)

Great work. The Background section is much improved. Fowler&fowler«Talk» 00:19, 20 March 2009 (UTC)

Support by Fowler&fowler

I have added my last set of comments to the article's talk page. (Williams tube). I am now ready to support the article. Congratulations on writing a fine article! Fowler&fowler«Talk» 19:07, 20 March 2009 (UTC)

Reply: Thanks for your support, and for your very detailed critique of the article, which has undoubtedly improved it quite significantly. I think I've addressed all of your remaining concerns on the article's talk page, apart from the one about "computer speed", which I'll try to elaborate on shortly. --Malleus Fatuorum 00:42, 21 March 2009 (UTC)

• Support The first picture should be on the right and second on the left (to match the alternation throughout). Its a shame we don't have any images of the creators. I would oppose this over the grounds that the first programmer was Ada Lovelace and this page being an FA would take away from her awesomeness among Wiki Early Comp Sci nerdom. However, I don't want to deal with that ANI drama. :P Ottava Rima (talk) 00:20, 20 March 2009 (UTC)

• Fair point, I've swirched them around. We can perhaps discuss Ada Lovelace elsewhere. :-) --Malleus Fatuorum 00:29, 20 March 2009 (UTC)

Oppose less than 200 references. —Preceding unsigned comment added by 141.161.92.138 (talk) 00:55, 20 March 2009 (UTC)

Note: above IP blocked for disruptive editing. –Juliancolton ^{Tropical Cyclone} 01:09, 20 March 2009 (UTC)

• Support with comments. This looks to be ready, but I would appreciate attention to a couple matters:
  • "Following on from his work, programmable computers such as the Z3 and Colossus ..." Would "Continuing his work" change the meaning? If not, much preferred.
  • "The construction of a von Neumann computer depended on the availability of a suitable memory device in which to store the program." Isn't standard computer terminology that you store data on a device, not in it?
  • "This was achieved by comparing each received pulse with the previous pulse ..." Avoid using the ambiguous "this" in reference to a previous concept. This what?
  • Check those math-thingies at the end of Design and development... I think if you're not using the math tags, you are supposed to use − for the negative sign as well?

--Laser brain (talk) 17:10, 20 March 2009 (UTC)

• Reply:
  • I've slightly reworded that "Following on from his work ..." sentence. It's a little contentious who knew about what, when, and who was influenced by whose ideas, so I've tried to make a neutral statement. I'm not certain that Konrad Zuse, for instance, would have seen his work as a continuation of Turing's. The main point I was trying to make there was the transition from theoretical machine to practical ones, that were actually built and worked.
  • Agree about storing "on" rather than "in". I tried to update the language to use terms that would be more accessible to a modern reader, but I obviously missed a few.
  • Changed "This" to "The filtering".
  • I think all of the minus signs have been caught now.
  • Thanks for your comments and of course for your support. --Malleus Fatuorum 21:12, 20 March 2009 (UTC)

Images checked and seem to be ok 5 images, all self-made, don't seem to be any problems. Tom B (talk) 22:57, 21 March 2009 (UTC)

The above discussion is preserved as an archive. Please do not modify it. No further edits should be made to this page.