An IEEE Milestone for the Manchester "Baby"

I am pleased to report that the IEEE History Committee has voted to move forward with approval of a proposal for the IEEE to add the early Manchester University computers to its list of historical milestones. The approved text is:

At this site on 21 June 1948 the “Baby” became the first computer to execute a program stored in addressable read-write electronic memory. “Baby” validated Williams-Kilburn Tube random-access memories, later widely used, and led to the 1949 Manchester Mark I which pioneered index registers. In February 1951, Ferranti Ltd's commercial derivative became the first electronic computer marketed as a standard product delivered to a customer.

(Official milestone name: "The Manchester University 'Baby' computer; Small-Scale Experimental Machine (SSEM)", details online here).

I was recruited as an expert reviewer of the proposal, but through months of back and forth with the proposers and members of the committee finished up providing input that led to substantial changes in the wording (including some negotiations during the committee meeting itself). This process was challenging for several reasons. Firstly, the proposers wanted to fit four distinct (if related) "firsts" onto one plaque: the prototype "Baby" computer, its Williams Tube memories, the full-scale Manchester Mark I's index registers, and the Ferranti Mark 1. Secondly, the word limit was fixed at exactly the length of the text above. No space to add one more word, however much it would help intelligibility. Thirdly, the IEEE is strongly discouraging the naming of people on its plaques (partly from a fear of leaving out important contributors). Fourthly, and most importantly, discussion of the early computers has traditionally been dominated by a highly partisan push to honor one or another machine as "the first computer," or at least as "the first stored-program computer." Professional historians tend to recoil in horror from these discussions, with good reason, but I thought that if plaques backed with the authority of the IEEE are being issued they should at least be scrupulously accurate. 

As someone who studied computer science in Manchester (loving both the town and the university) but has more recently written extensively about ENIAC, I tried to propose language that would be precise enough to avoid challenge from fans of computers build in other places (most obviously Cambridge and Philadelphia) and would at the same time be more specific about what the contributions actually were. The first three requirements limited comprehensibility, no space for example to explain what an index register is, but there will be a chance to place additional information next to the physical plaque and linked to the electronic version. The claims made on the plaque text are at least correct. For example, "the first electronic computer marketed as a standard product delivered to a customer" is not the most elegant phrase, but it at least manages to be not wrong by respecting the facts that the UNIVAC was marketed as a standard product years earlier, that non-electronic devices called computers had been produced in bulk during the Second World War, and that several one-off electronic computers had been previously built under contract and delivered to their sponsors including at least one (BINAC) built by a private company.

I feel particularly proud that I managed to steer the text away from a claim to have the first "stored program computer" and towards the more meaningfully operationalized claim that this was the first time a program stored entirely in an addressable read-write electronic memory had been executed. As I have discussed elsewhere, any machine that carries out a program of operations must in some way store or embody that program, so the phrase "stored program computer" is hopelessly vague and can be defined one way by boosters of ENIAC (which was reconfigured in 1948 to run modern-style code held in a read-only addressable memory a few months before the Baby ran its first program) and in another by proud Mancunians. Alas I have no power to wave a magic wand to stop others making claims for the "first computer" or "first stored-program computer."

Careful wording also let us gesture towards the role of the so-called "baby" as a prototype machine built to validate the new electronic memory devised at Manchester. While much writing on early computing suggests that the key requirement for success was a conceptual breakthrough, the reality is that the construction of a large, reliable, affordable, and high-speed electronic memory was the crucial challenge standing between the many projects set up in the second half of the 1940s to build programmable electronic computers and the creation of a usable machine. Almost all the projects fell years behind schedule, and some never worked properly.