Talk:Stored-program computer

This  level-5 vital article is rated C-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Computing High‑importance 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 High This article has been rated as High-importance on the project's importance scale. This article is supported by Early computers task force (assessed as Top-importance).

• As it is most often used adjectivally, is ‘stored program’ or ‘stored-program’ to be preferred? MOS:HYPHEN would seem to favour the latter.
• Was the virtual machine described in Turing’s 1936 ‘Computable numbers’ paper^[1], a stored-program machine?
• Turing’s 1946 Automatic Computing Engine was undoubtedly a stored–program computer design in the modern sense.
• Was ENIAC’s 1948 demountable ‘function table’, a stored-program feature, as some have claimed?
• Despite being essentially a test-bed for the Williams tube, the 1948 SSEM was undoubtedly the first true stored-program computer.
• Was the University of Cambridge Computer Laboratory's 1949 EDSAC the first practical stored-program machine to become operational, as is claimed?
• EDSAC pre-dated the Manchester Mark 1 by some five months.

--TedColes (talk) 12:45, 10 November 2010 (UTC)Reply

You bring up some good points. Honestly, I don't know enough about individual early machines to reply to all of them.

• To your first point, (Was the virtual machine described in Turing’s 1936 ‘Computable numbers’ paper[1], a stored-program machine?), no, a Turning machine can be implemented with a paper tape system or even toilet paper. No electronic storage is necessary.
• Turing’s 1946 Automatic Computing Engine was undoubtedly a stored–program computer design in the modern sense. - *Nod*. This says it was the 3rd stored program computer in Britain. Raul654 (talk) 15:44, 10 November 2010 (UTC)Reply

I have not been clear enough.

• The Universal Turing Machine as described by Turing in 1936 was only virtual and did not have any storage medium specified. It was called a 'tape' only to illustrate that one cell could be accessed at a time. My point is that the idea of instructions and data sharing the same storage medium and instructions being modifiable, both exist in this design. So it does not fit the definition of being electronic, but does have a strong case for being a very influential source of the idea.
• Turing's 1945/6 paper describing the Automatic Computing Engine design was also a virtual machine, as the Pilot ACE to which the above quotation refers was a substantially cut-down version of what he had proposed. It did indeed come after the SSEM and Manchester Mark 1 and the EDSAC.

This article ought to be well informed about early developments if it is to achieve credibility and if the redirect from 'Stored program' to 'Von Neuman machine' is to be abolished. --TedColes (talk) 17:21, 10 November 2010 (UTC)Reply

(Disclaimer: Manchester is one my Alma mata) I think some people are likely to regard labelling EDSAC as "the first practical stored-program machine to become operational" as controversial. Definitely, EDSAC was an important and innovative stored-program computer, however "practical" has a little slippery definitionally, I think. I think rewriting as an sourced opinion would work better. At the moment, it could be read as an attempt to belittle the innovation work at Manchester (though I don't think that's the intention). In my opinion both designs streams are important to future stored-program computers, with Baby leading towards distributed system-on-chip architectures whereas EDSAC leads towards CPUs RobertBurrellDonkin (talk) 20:01, 4 August 2013 (UTC)Reply

The case for this claim is made by Professor Martin Campbell-Kelly in "Martin, Campbell-Kelly (2012), "Ivory Towers and Tea Rooms", in Lavington, Simon (ed.), Alan Turing and his Contemporaries: Building the world's first computers, Swindon: British Computer Society, ISBN 978-1-90612-490-8" TedColes (talk) 22:08, 4 August 2013 (UTC)Reply

The Manchester Mark 1 was being used for research at Manchester University in April 1949, a month before EDSAC ran its first program.^[1] I've no idea where this 'EDSAC predated the Manchester Mark 1 by five months' idea could have come from. Yes, the Manchester Mark 1 didn't reach its final form until October 1949; on the other hand, EDSAC didn't reach its final form until perhaps 1955 when it finally got its full 1024 words of serial access mercury delay line memory - each word being 17 bits. For comparison, the Manchester Mark 1's October 1949 form had 256 words of 40 bits in random access CRT memory, and a further 4096 words in drum storage (also working memory). Michael F 1967 (talk) 04:03, 31 July 2026 (UTC)Reply

This text's reference 13 (to the claim about Zuse's two patents) isn't correct. It points to Nature's text where Williams & Kilburn describe Manchester "baby". That text certainly doesn't talk about Zuse's patents.

Apparently some people say it is "historically inappropriate, to refer to electronic stored-program digital computers as 'von Neumann machines'".

So what do those people say is the historically appropriate use of those two terms? Are these people saying there is some subtle distinction (or perhaps one is a subset category of the other) between stored-program computers and von Neumann mmachines? If so, what is that distinction?

Or are those people saying that the machines that people call "von Neumann machines" are identically the same as (synonymous with) "stored program computers", but it is anachronistic to apply von Neumann's name to machines that were developed before von Neumann ever thought of such machines? If so, please merge the appropriate parts of the von Neumann architecture article into this stored-program computer article. --DavidCary (talk) 18:46, 23 August 2013 (UTC)Reply

A paper documenting the history of the concept: ‘Stored Program Concept' Considered Harmful Diego (talk) 15:33, 27 December 2014 (UTC)Reply

The article says: "The third stored-program computer to be built, and the first one in continental Europe, was the MESM, completed in the Soviet Union in 1951." But EDVAC was completed in 1949, and it isn't listed. Bubba73 ^{You talkin' to me?} 02:48, 29 May 2015 (UTC)Reply

Well, actually it had problems until 1951. Bubba73 ^{You talkin' to me?} 04:46, 29 May 2015 (UTC)Reply

The article says that the ARC2 came online in May 1948. But did it run a stored program? The article about the development of the APE(X)C says that it wasn't completed until June 1952.

Early British Computers, by S. Lavington, page 63 says that the APE(R)C - different from the APE(X)C, I believe - "was operating with limited storage in July 1952." It doesn't mention the APE(X)C. Bubba73 ^{You talkin' to me?} 01:29, 24 July 2017 (UTC)Reply

"The Development of Computer Programming in Britan (1945-1955)" by Martin Cambell-Kelly, in Annals of the History of Computing", vol 4, #2, lists APEXC as 1953. But I need to read the whole article to see what it says about APEXC. Bubba73 ^{You talkin' to me?} 01:11, 11 August 2017 (UTC)Reply

It sounds like pretty much every computer today would fit into this category. Shouldn't there be a mention of that in the article? flarn2006 [u t c] time: 20:47, 9 January 2019 (UTC)Reply

Yes. Dicklyon (talk) 06:55, 10 January 2019 (UTC)Reply

As far as I know, every general-purpose one built since the late 1940s. Bubba73 ^{You talkin' to me?} 07:49, 10 January 2019 (UTC)Reply