Milestone-Proposal talk:The Manchester University 'Baby' computer; Small-Scale Experimental Machine (SSEM): Difference between revisions

From IEEE Milestones Wiki
Line 134: Line 134:
'''''Conclusion'''''
'''''Conclusion'''''
In conclusion, the proposed citation has been precisely worded to make sure that the statements made are literally true. Doing this within 70 words was quite a challenge, and I am happy to endorse the results achieved.
In conclusion, the proposed citation has been precisely worded to make sure that the statements made are literally true. Doing this within 70 words was quite a challenge, and I am happy to endorse the results achieved.
== Submission and expert reviews now complete. -- [[User:Rmuttram|Rmuttram]] ([[User talk:Rmuttram|talk]]) 22:35, 10 March 2021 (UTC) ==
The Proposers have had significant dialogue with the Advocate and through him to the Experts, in particular with Thomas Haigh. Through that process the Citation has been refined and I am happy to confirm that the Proposers approve it in its current form to go to the History Committee for consideration and hopefully recommendation to the IEEE Board.

Revision as of 22:35, 10 March 2021

Significant update following discussions between the team and the Advocate -- Rmuttram (talk) 18:28, 26 February 2021 (UTC)

Justifications for specific plaque wording and additional references added.

Photos embedded -- Rmuttram (talk) 00:01, 27 February 2021 (UTC)

Minor text adjustments to accommodate

Minor updates -- Rmuttram (talk) 21:35, 27 February 2021 (UTC)

Following team review after photograph insertion, a few more textual amendments to match.

Final update to citation following dialogue with Advate and Experts -- Rmuttram (talk) 11:01, 9 March 2021 (UTC)

Whilst the names of Wiliams and Kilburn are now only included as part of the name of the Williams-Kilburn Tube we have left the name inclusion justification untouched.

Minor edits following further dialogue with the Advocate -- Rmuttram (talk) 17:08, 9 March 2021 (UTC)

To meet the concerns of one of the Experts.

Expert Review #1 of 3, by Dr. Andrew Herbert -- Bberg (talk) 17:07, 10 March 2021 (UTC)

As Advocate for the Manchester "Baby" Milestone proposal, I solicited an expert review by Dr. Andrew Herbert, who is Chairman of Trustees for the UK National Museum of Computing, and Manager of the project at our museum to build a working reconstruction of EDSAC, another pioneering computer contemporaneous with the "Baby."

A PDF of Dr. Andrew Herbert's 10 March 2021 review of the proposal is at File:ManchesterBaby-ExpertReview(AndrewHerbert).pdf

The following is the text from this PDF:

Thank you for the invitation to review the proposal for the IEEE to erect a plaque in Manchester to celebrate the Small Scale Experimental Machine (SSEM / “Baby”) as the world’s first computer to execute a meaningful program stored in a memory. I am responding in a personal capacity, but also in the broader context of being Chairman of Trustees for the UK National Museum of Computing and manager of the project at our museum to build a working reconstruction of EDSAC, another pioneering computer contemporaneous with Baby.

The proposal is well researched and provides more than adequate documentary evidence to make incontrovertible the significant claim that Baby was the first computer to execute a program stored in memory.  The proposal also goes on to list other seminal developments from Manchester University that flowed from the development of Baby and its successors the Manchester Mark 1 and 1* developed commercially by Ferranti, also a world first.

I understand you have also asked my colleague and friend Mr Christopher Burton to review the proposal. He is an authority on Baby, having led the project to build the reconstruction now in the Manchester Science and Industry Museum, and he too confirms the claims are correct based on discussions with Williams, Kilburn and Toothill and supporting documentary evidence.

The matter of “first computer” is always a challenging one in which to make claims, as many groups in the USA, UK and elsewhere were on the threshold of building stored program digital computers at the same time, but the credit for arriving first at a working computer running programs from a read-write memory undoubtably goes to Manchester. There were mechanical, electromechanical and indeed electronic predecessors to Baby but none can claim to have had a read-write memory in the way we use the term today. The machines developed contemporaneously with Baby only came into successful operation significantly after Baby executed its first program. These machines may have been bigger, faster, and/or more practical for users but they commenced operation later. Similarly, while other machines were also subsequently developed commercially, the evidence is incontrovertible that Ferranti manufactured the first machines to be delivered, as a product, to customers and indeed customers in several countries.  Other manufacturer’s machines may have been delivered in greater volumes and more widely but this was to come later than Ferranti’s initial sales.

Reliable, fast, random access storage technology was the biggest technical challenge to beset the early pioneers. The development of Baby contributed an important new storage technology in the form of the Williams Kilburn Cathode Ray Tube store which enabled random access to a persistent memory and spawned subsequent developments elsewhere and commercially.  From the evidence presented it is unambiguously clear that the Williams Kilburn Cathode Ray Tube was a significant innovation at the time and entirely conceived of and developed by Williams and Kilburn.

As the proposal goes into at considerable length, the construction and success of Baby was the start of a series of further direct contributions to the development of key concepts in modern computers including index registers, virtual memory and early developments in programming languages and compilers.

For all these reasons I believe Baby is an achievement the IEEE should recognize and I look forward to seeing the plaque erected as proposed.

Dr Andrew Herbert OBE, FREng Chairman of Trustees, The National Museum of Computing Manager, The EDSAC Replica Project

Expert Review #2 of 3, by Christopher Philip Burton -- Bberg (talk) 17:11, 10 March 2021 (UTC)

As Advocate for the Manchester "Baby" Milestone proposal, I solicited an expert review by Christopher Philip Burton, who proposed and directed a project to build an authentic working replica of the Manchester “Baby” which successfully publicly demonstrated the running of the first program on its 50th anniversary in 1998, and which is now an operational display in the Museum of Science and Industry in Manchester.

A PDF of Christopher Philip Burton's 10 March 2021 review of the proposal is at File:ManchesterBaby-ExpertReview(ChristopherPhilipBurton).pdf

The following is the text from this PDF:

I have read the Milestone proposal carefully, and agree the correctness of the citation on the plaque, together with the supporting documentary material. I wholeheartedly recommend that the Committee should approve the project as described in the proposal. The date of the correct running of the first program Monday 21st June 1948 was indelibly fixed in the minds of the pioneers at the time, corroborated by entries in the notebook of G C Tootill. Independently from enquiries by others in the past, I was involved with an extensive discussion with T Kilburn and G C Tootill in March 1996 regarding the exact nature of the first program and the circumstances of its running on 21st June 1948. The results of that discussion were published by G C Tootill in 1998 in https://computerconservationsociety.org/resurrection/res20.htm#e

The date cited on the proposed plaque is evidently correct.

There is a further consideration to add to the supporting information of the proposal that resolves an apparent ambiguity. For many years, the reference in the paper in Nature [ref. 1(a) in the submission] to successful running of the factoring program for 52 minutes has been conflated with the events of 21st June 1948. This is because there was no explicit evidence to the contrary. Our 1996 investigation pointed out that on the day the first program ran, the two engineers T Kilburn and GC Tootill were trying to ensure that the hardware of the machine was correct so that their program would run correctly. As well as any wiring modifications they had been making up to that point, there were a number of adjustable controls that had to be set, such as the clock period and associated timing delays. On the Monday it was that kind of setting-up that they were doing, while testing how far the program would progress before failing. Clearly they would need to see the result of an attempted run as soon as possible after making an adjustment. So they initially used small trial numbers to factorise in the program, which could then run to completion quickly. Once those trials succeeded, for the rest of the morning and then in the afternoon of the 21st, and in the following days, they progressively used larger numbers up to the 52-minute run. That signified not only the correct functioning of the storage system, but also that it had adequate reliability for useful work. In our context, the relevant fact is that the program first ran correctly in electronic storage on 21st June 1948.

F C Williams and T Kilburn are unequivocally the inventors of the Williams-Kilburn CRT storage tube technology They and only they were responsible for the unique technique  that made digital storage applicable and successful in electronic computers.

Furthermore, that technique was the basis of and the reason for the construction of the “Baby” computer. Without that final experiment it would not have been justifiable to declare that the storage technology was appropriate and successful in a “universal” (in the Turing sense) computer. The proof was that they showed that an electronically stored program and its data worked correctly and reliably at electronic speed. As a result that computer became the prototype for a highly successful series of developments leading to commercial deliveries of computers. The storage technology and the first successful running of a program using it, are inextricably bound to their two names.

Thank you for the opportunity for me to support this excellent proposal.

Yours sincerely, Christopher Philip Burton FIEE CEng. Hon FBCS

I joined the Ferranti Computer Division in 1957 and worked on a succession of large-scale computer systems from vacuum-tube machines onwards to AI applications in the 1980s. I have been a member of The Computer Conservation Society since its foundation in 1989, leading a number of Working Parties conserving and restoring early computers. In 1994 I proposed and directed a project to build an authentic working replica of the Manchester “Baby” which successfully publicly demonstrated the running of the first program on the 50th anniversary in 1998, and is now an operational display in the Museum of Science and Industry in Manchester. I received a number of awards for this project. Currently I am a member of the team building a working replica of the Cambridge University EDSAC.

Expert Review #3 of 3, by Dr. Thomas Haigh -- Bberg (talk) 20:32, 10 March 2021 (UTC)

As Advocate for the Manchester "Baby" Milestone proposal, I solicited an expert review by Dr. Thomas Haigh, who is a Professor of History at the University of Wisconsin‐Milwaukee and a visiting Comenius Professor for the History of Computing at Universität Siegen. His PhD in the History and Sociology of Science is from the University of Pennsylvania. He is the lead author of ENIAC In Action (MIT, 2016) and A New History of Modern Computing (MIT, 2021). See also http://www.tomandmaria.com/Tom/

A PDF of Dr. Thomas Haigh's 10 March 2021 review of the proposal is at

File:ManchesterBaby-ExpertReview(ThomasHaigh).pdf


The following is the text from this PDF:

Evaluation of Citation

The final wording is the result of several rounds of back and forth exchange of proposed drafts with the proposers, mediated by Brian Berg. During this process the citation text became, from my viewpoint at least, far more precise and historically reliable.

The current version identifies several distinct contributions made by three related machines:

  • the 1948 “Baby” (known officially as the Small Scale Experimental Machine or SSEM), a minimal prototype computer which ran test programs to prove the viability of the
  • Manchester Mark 1, a full-scale computer completed in 1949 that was fully designed and approved only after the success of the “Baby” and in turn served as a prototype of the
  • Ferranti Mark 1, a commercial refinement of the Manchester Mark 1 of which I believe 9 copies were sold. The 1951 date refers to the delivery of the first of these, to Manchester University as a replacement for its home-built Mark 1.


The citation also foregrounds the importance of the Williams Tube memory. Creating a reliable, high capacity electronic memory was the crucial stumbling block faced by computer projects in the late-1940s and the reason that almost all of them took much longer than expected to complete. Some projects failed entirely because of the difficulty. The new memory tubes were widely licensed for other computer projects, for example by IBM, before being displaced by core memory in the mid-1950s.

The relevant facts are well documented, and widely available in scholarly historical work or even on Wikipedia. The main challenge was squeezing all this down into 70 words without simplifying to the extent that the text became untrue. The result is rather dense, so let me explain why all the words are needed.


Citation Sentence 1: At this site on 21 June 1948 the “Baby” became the first computer to execute a program stored in addressable read-write electronic memory.

The date in question and the program executed are well documented and uncontroversial.

Sometimes the claim would be made as “first stored program computer” but like “first computer” its validity would depend on how one defines “stored program.” That led to considerable historical wrangling around the term, since any automatic computer relies on a program stored in one form or another. As von Neumann wrote in the First Draft of a Report on the EDVAC: “instructions must be given in some form which the device can sense: Punched into a system of punchcards or on teletype tape, magnetically impressed on steel tape or wire, photographically impressed on motion picture film, wired into one or more fixed or exchangeable plugboards—this list being by no means necessarily complete.”

There’s a rather arbitrary historical process by which “stored program computer” emerged in the 1950s as a replacement for earlier terms such as “EDVAC-like machine” to describe computers patterned after the ideas described by von Neumann and his collaborators in Philadelphia and, later, Princeton. The problem in describing a first “stored program” computer was that the phrase could be interpreted in different ways. Machines such as the Manchester Baby and Cambridge EDSAC were clearly EDVAC-like and met any plausible definition of “stored program”, as were almost all the commercial computers of the 1950s, but several authors advanced definitions of “stored program” according to which other machines qualified as the first. In particular Allan Olley argued for IBM SSEC (Feb 1948) as the “first stored program computer” because it could potentially store and execute instructions held in a (non-addressable) electronic or relay memory, while authors such as Jean Bartik, Nick Metropolis, and Crispin Rope separately suggested that the reconfigured ENIAC (March 1948) deserved the title because it executed instructions fetched at electronic speed from an addressable read-only memory (switch panels). There’s also the question of Booth’s ARC computer, which is under documented to the extent that it is hard to be sure if or when it worked, but has sometimes been called a “stored program computer” built around a mechanical drum rather than an electronic memory. In response to issues such as these, Alan Bromley broke “stored program” into a series of staged innovations rather than a single “first.”

The adopted wording avoids all such potential challenges by being very specific about what “stored program” actually means here: executing a program stored in an addressable, writable and electronic storage medium.

For more on these points and citations to all the authors mentioned, see

  • Haigh, Thomas, Mark Priestley, and Crispin Rope. ENIAC In Action: Making and Remaking the Modern Computer. Cambridge, MA: MIT Press, 2016 (chapters 6, 7, 8 & 11)
  • Haigh, T., M. Priestley, and C. Rope. "Reconsidering the Stored-Program Concept." IEEE Annals of the History of Computing 36, no. 1 (Jan-Mar 2014): 4-17.
  • Priestley, Mark, and Thomas Haigh. "The Media of Programming." In Exploring the Early Digital, 135-158. Cham, Switzerland: Springer, 2019.


Citation Sentence 2: “Baby” validated the widely used Williams-Kilburn Tube random-access memories and led to the 1949 Manchester Mark I which pioneered index registers.

This highlights the crucial role of the “Baby” in validating the memory units. “Led to” reflects the fact that construction of the full-scale computer was not authorized until after memory technology was “validated” by the “Baby.” The proposers prefer this phrasing to anything involving the word “prototype,” and in my view this phrasing adequately captures the relationship.

Index registers were an important innovation present on the Mark 1 but not on the prototype “Baby.” They are worth mentioning even though space does not permit a definition. Anyone who has taken a computer architecture course will be familiar with the concept.


Citation Sentence 3: In February 1951, Ferranti Ltd's commercial Mark I became the first electronic computer marketed as a standard product ever delivered to a customer.

The wording here may seem clumsy, but the “first” in question requires all three elements to be there:

  1. Electronic computer (as opposed to mechanical computers such as https://en.wikipedia.org/wiki/Torpedo_Data_Computer) AND
  2. Marketed as a standard product AND
  3. Delivered to a customer.


This is because the Univac, while marketed as a standard product (and ordered by customers) as early as 1946, was not delivered for the first time to a customer until a few months after the Ferranti Mark 1. In contrast, a custom-built ERA computer and the BINAC were commercially produced and delivered prior to the Ferranti Mark 1 but were one-off machines that were never seriously marketed as standard products.

Conclusion In conclusion, the proposed citation has been precisely worded to make sure that the statements made are literally true. Doing this within 70 words was quite a challenge, and I am happy to endorse the results achieved.

Submission and expert reviews now complete. -- Rmuttram (talk) 22:35, 10 March 2021 (UTC)

The Proposers have had significant dialogue with the Advocate and through him to the Experts, in particular with Thomas Haigh. Through that process the Citation has been refined and I am happy to confirm that the Proposers approve it in its current form to go to the History Committee for consideration and hopefully recommendation to the IEEE Board.