[SIGCIS-Members] "Stored program" -- anyone know origins of the PHRASE

Allan Olley allan.olley at utoronto.ca
Mon Apr 2 22:00:22 PDT 2012


Tom,
 	I can't help with the actual phrase but I did track down an early 
example of the narrative of the Draft Report as disseminator of the 
particular features entailed by the phrase (I hope this is of interest)...
 	Reading your e-mail, I played a hunch and took out my copy of 
Wilkes, Wheeler and Gill's _Preparation of Programs for an Electronic 
Digital Computer_ (1951), I did not find a use of the term "stored 
program" (have not read the whole book but it is NOT in the index) but I 
did find a fairly standard ennunciation of the concept, an emphasis on it 
as key to the EDSAC design and an attribution back to the Draft Report as 
the source.

"A store or memory is needed in automatic computing machines for the 
purpose of holding numbers, and in the EDSAC the same store is used to 
hold the orders; this is made possible by the device of expressing the 
orders in a numerical code. Several machines working on the same 
principles as the EDSAC are now in operationin the United States and in 
England. These principles derive from a report drafted by J. von Neumann 
in 1946 in connection with a new machine (the EDVAC) then projected by the 
Moore School of Electrical Engineering (University of Pennsylvania) where 
the ENIAC had been built." p. 3

Note the machines being contrasted against the EDSAC type are the 
original wire programming of the ENIAC and the tape controlled machines 
such as Aiken's Mark I/IBM ASCC, Bell relay and the IBM SSEC (their 
categorization of the various machines not mine).

Later the book discusses modification of code "The facility of being 
able to modify the orders in the program by performing arithmetical 
operations on the numbers representing them is of great importance, and 
it is perhaps the feature most characteristic of program design for 
machines like the EDSAC." p. 9.

Now this does not explicitly refer back to the discussion of page 3 but 
I'm not sure what "machines like the EDSAC" could mean other than other 
machines that were in Wilkes et al. understanding derived from the Draft 
Report.

So the story that the Draft Report disseminated design of storing numbers 
and instructions in the same storage while taking advantage of 
self-modification, and the idea that this was its key contribution goes 
back at least to this 1951 book. Actually farther back then I would have 
guessed (I was not clear that these elements of the Draft Report had been 
singled out so early as the key novelty). So, if the phrase stored program 
with the modern meaning had been known to Wilkes et al. at the time of 
writing they would presumably have used it.

An interesting contrast is the 1948 papers A Discussion on Computing 
Machines D. R. Hartree, M. H. A. Newman, M. V. Wilkes, F. C. Williams, J. 
H. Wilkinson and A. D. Booth, Proceedings of the Royal Society of London. 
Series A, Mathematical and Physical Sciences Vol. 195, No. 1042 (Dec. 22, 
1948), pp. 265-287. JSTOR: http://www.jstor.org/stable/98328

This is a series of short articles (proceedings of a workshop) where the 
authors explain their machines (in case of Williams he just describes his 
eponymous tube and not a fully machine) or more general aspects of 
electronic computing.

Here in his 1948 description of the EDSAC, Wilkes does explain that the 
EDSAC will store orders and numbers in the same store, he does not 
note that this as special in any way (to be fair the article is short  and 
he has to outline the whole machine). Similarly Wilkson describing the 
ACE decides to emphasize ways of optimizing the speed of delay lines by 
spacing instructions through the system (clear but implicit that 
instructions will be in the same store) and use of short tanks (delay 
lines that only store a single number for quick retrevial).

Hartree and Newman discuss the general principles of the new machines and 
both emphasize the importance of "judgement" (Hartree after Babbage) or 
"conditional change of control" (Newman), what would now be 
called conditional branching, as a key control concept. They do not as far 
as I can tell mention storing instructions in the same device as data or 
self-modification of instructions. I mention all this because to me an 
interesting aspect of the story you are investigating is what various 
people have taken the key feature (essence?) of the computer to be (but 
you already knew that from reading my paper).

-- 

Yours Truly,
Allan Olley, PhD

http://individual.utoronto.ca/fofound/

On Mon, 2 Apr 2012, Thomas Haigh wrote:

> Hello everyone,
>
> I have a query related to a project I am working on concerning the
> conversion of ENIAC to stored program control in 1948, initially to run the
> first computerized Monte Carlo calculations. All this took place prior to
> the first operation of the Manchester Baby. That makes the question of what
> one means by "stored program" a very interesting one.
>
> This question was much discussed in the early days of the history of
> computing (1970s, early 1980s). I am starting to dig back into primary
> sources for early use of the phrases "stored program" and "stored program
> concept" to get a better idea of how these terms were used in the
> 1940s/early 1950s and what people thought they meant at the time.
>
> To clarify, almost everyone who has written about this cites the 1945 "First
> Draft of a Report on the EDVAC" as the initial dissemination of the stored
> program concept although there has been considerable debate as to the source
> of the ideas contained therein. However that document does not contain the
> phase "stored program." Or indeed use the word "program" in the body of the
> text. Or, remarkably, "EDVAC." "Stored" shows up a few times, though less
> frequently than "remembered." So, ignoring for the moment the relationship
> of the report to later definitions of the concept, we can agree that it was
> not the source of the phrase. The most obvious summary of the idea using the
> report's own terminology would be "remembered instruction device" rather
> than "stored program computer."
>
> I had thought about the 1946 Moore School lectures as a possible vector for
> the phase "stored program" as well as the concept. The phrase shows up many
> times in the Moore School lectures book but so far I have spotted it only in
> the 1980s editorial material rather than in the original lecture summaries.
>
> By 1954 "stored program computer" is showing up without explanation or
> citation required in the description of the IBM 650 published in the
> inaugural issue of Journal of the ACM. It is not particularly common in the
> ACM DL material for the rest of the decade ("automatic computer" and
> "digital computer" are more prevalent) but continues to pop up occasionally.
> The best the OED can do is 1957, which is even later.
>
> So, any thoughts on who came up with this phrase and when? I'm planning to
> dig deeper in search of early usage, for example into the 1950 "High-Speed
> Computing Devices" ERA book and some of the other CBI reprints from the
> 1940s, but it occurred to me that someone on the list might already know the
> answer to the question.
>
> Tom
>
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. The list archives are at http://sigcis.org/pipermail/members/ and you can change your subscription options at http://sigcis.org/mailman/listinfo/members
>



More information about the Members mailing list