webber@porthos.rutgers.edu (Bob Webber) (06/28/88)
In article <1496@its63b.ed.ac.uk>, bct@its63b.ed.ac.uk (B Tompsett) writes: > In article <Jun.20.04.49.43.1988.3576@porthos.rutgers.edu> webber@porthos.rutgers.edu (Bob Webber) writes: > >[....] Also, it was only plugboard-programmed until > >1948 at which time it became the first stored-program computer (although > >the store was read-only). Prior to 1948, it was a parallel computer [....] > Correction: the Manchester Mark I was the first stored-program computer. It > first ran on 21st June 1948. This week marks the 40th aniversary ... PERHAPS. So far, I have found three references relevant to this question. The primary reference is the chronology of the stored-program concept presented in Metropolis and Worlton's A Trilogy on Errors in the History of Computing [Proc. USA-Japan Conf 1st,, Tokyo, Oct 3-7,1972.] Here, the chronology is laid out as: Manual Program Control: Bell Labs's Complex Calculator (1940) Automatic Program Control: Zuse Z3 (1941) and Harvard Mark I (1944) Internal Program Control: ENIAC (1946) Storage Program Control: ENIAC (modified-1948) Read-Write Storage Program Control Concept: EDVAC 1945 Read-Write Storage Program Control Implemented: BINAC and EDSAC (1949) Two questions are raised here: 1) Why no mention of the Manchester Mark I and 2) just when in 1948 did the ENIAC get converted. On the second point, I have so far seen no references to an exact day in 1948 when the ``first ENIAC'' program was run. Note that the ENIAC was already a working computer when it was converted and it is claimed that its design specifically invited this conversion (i.e., it was an ``easy'' upgrade). So it may be that the ENIAC ran a stored program before or after tha Manchester Mark I did in 1948. However, in the listings on the other points in history, this article does not seem one to cut such a fine distinction, i.e., I would have expected to simply see both machines listed for 1948. Their primary source of information on the conversion of the ENIAC was an interview with R. F. Clippinger that was part of the Smithsonian AFIPS History Project with the reference ``to be published,'' which I have not seen. One hypothesis is that the authors who were trying to cut thru all the errors surrounding the history of the stored program idea were themselves unaware of the Manchester Mark I. However, their bibliography shows that they had reference to the Earl of Halsbury's Ten Years of Computer Development (which appeared in the first volume of Computer Journal in 1959 on pages 153 thru 159) as well as personal correspondence with the Earl. Also, this paper has an extensive section on the awareness of the work of Babbage among the early computer designers with many references to the history of computing in Britain. There is also specific mention of Williams and Kilburn at Manchester as being a group working independent of the ENIAC-EDVAC stimulus. Part of the mystery seems to clear when we refer directly to the Earl's article. He is the only one of these three authors to actually define what he means by ``modern computer.'' He regards it as involving: 1) all-electronic data processing; 2) stored program; and 3) automatic peripheral equipment. This causes him to cite Wilke's EDSAC at Cambridge May 1949 as the ancestor of all such equipment. He mentions that the ENIAC had insufficient memory to tackle many problems as the reason for discounting it. He indicates that the Manchester Mark I is being discounted because of lack of ``automatic peripheral equipment.'' He then goes on to say that if you discount the 3rd criteria, then you would view the Manchester Mark I as the first and he cites a ``prototype'' of the Manchester Mark I in the ``Spring'' of 1948 that could ``only subtract and had only manual input.'' The third reference is S. H. Lavington's Computer Development at Manchester University which appeared in A History of Computing in the Twentieth Century (edited by N. Metropolis, J. Howlett, and Gian-Carlo Rota, Academic Press, 1980, 433-445). This article distinguishes to two machines, the University Mark I and the Ferranti Mark I, indicating that the first was a ``prototype'' (the Ferranti Mark I was referred to by Turing as the Mark II). The order code for the Ferranti MARK I was based on the order code for the University Mark I ``as it existed in October 1949.'' Reference is also made to ``a series of University prototypes completed during the period June 1948 and October 1949.'' Reference is also made to a ``52-minute program on 21 June 1948'' that was run. So as to the claims of the Manchester Mark I versus the Clippinger-modified ENIAC, there are a number of points unclear: 1) precise date for the ENIAC modification 2) precise capabilities of the June 48 Manchester Mark I 3) just what the heck is a ``computer'' anyway? ------ BOB (webber@athos.rutgers.edu ; rutgers!athos.rutgers.edu!webber)
webber@aramis.rutgers.edu (Bob Webber) (06/30/88)
In article <3255@thorin.cs.unc.edu>, omondi@unc.cs.unc.edu (Amos Omondi) writes: > In article <Jun.28.09.58.44.1988.6473@porthos.rutgers.edu>, webber@porthos.rutgers.edu (Bob Webber) writes: < < ... < < On the second point, I have so far seen no references to an exact day < < in 1948 when the ``first ENIAC'' program was run. Note that the ENIAC < ^^^ < This is probably understandable since someone appears to have remembered < these modifications only years after ... Hmmm. I am not sure quite what you mean to be implying there. The modifications were quite ``real.'' For example, there was a report authored by ``Anonymous'' entitled ``Description and Use of the ENIAC Converter Code'' reviewed in 1950 (vol 4, pp. 150-151 of Mathematical Tables and Aids to Computation) -- the report itself, consisting of 23 mimeographed pages, was dated Nov 1949 according to the review. Also of interest was the second meeting of the ACM, which was held in Dec 11 & 12 of 1947. At this meeting, two papers were presented whose titles suggest a relevance to this discussion (alas, I have not seen these papers anywhere): ``General Principles of Coding, with application to the ENIAC'' by J. von Neumann and ``Adaptation of the ENIAC to von Neumann's coding technique'' by R. F. Clippinger. The meeting was held at BRL. [I wonder if they drifted north to the Moore School Library.] < < was already a working computer when it was converted and it is claimed < < that its design specifically invited this conversion (i.e., it was an < < ``easy'' upgrade). So it may be that the ENIAC ran a stored program < < before or after tha Manchester Mark I did in 1948. However, in the < < How is it that neither Burks, Eckert, nor Mauchly, who surely ought to < know a thing or two about the ENIAC do not appear to have said anything < about these modifications or this 1948 date ? Well, for one thing, they seem to not have been directly involved in the modification which was made after the ENIAC had moved to BRL. However, they do figure in the question of who gets credit for what. Basically, I think the easiest way to understand the ENIAC is to view it not as a computer but rather as a computational laboratory. There are many interesting ways it could be ``put together'' to solve a given problem. The system was quite flexible was clearly intended to be modified. However, the designers of this marvelous machine seem to have abandoned this approach to computing before the ENIAC was even finished -- focussing instead on what is now commonly referred to as the ``von Neumann'' architecture. --- BOB (webber@athos.rutgers.edu ; rutgers!athos.rutgers.edu!webber)