A Shaggy Dog Story

Software Firsts

Back in July 1994, Frank King posted the following message to the Change-Ringers’ mailing list:

I suppose one can assume that many on this list share an interest in computers and ringing and some might have an interest in the historyof the intersection of these activities.

The following shaggy dog story appeared in response.

The story starts in the early 1980’s after a University of London Society ringing practice. I am in conversation in the pub with someone who was then a frequent visitor, and who shall remain nameless, for reasons which will become apparent.

”What I can’t understand” said he (after a few pints) “is why everyone wants to know whether you need singles in a peal of Stedman Triples, because years ago when I was doing my National Service (compulsory military service for US readers) I produced a proof that you couldn’t produce a peal with bobs only”

”That’s interesting” said I, “because I thought this was still an open conjecture for odd-bob peals. What a pity you never published it”.

”Ah I did, but I can’t remember when it was”.

Knowing that the speaker was prone to exaggeration, I set out to call his bluff. Making an inspired guess about his age, I searched through Ringing World indexes for the early 1950’s, and to my surprise found an article in the issue of February 20th 1953 under the pseudonym “Q.E.D.” entitled A Note Upon Some Aspects of Stedman Triples. It was indeed the proof.

An old volume of the Ringing World is a natural invitation to browse, and in the next issue I found a congratulatory letter from a certain Albert York-Bramble, a great generator of ringing pseudo-theory, then at his peak of gibberish production. Q.E.D. will surely have earned the thanks of all students of composition theory he gushed after the obligatory column of gibberish.

In the next issue there appeared a letter from one Brian D Price, opening baldly: Dear Sir, Q.E.D.’s article on page 125 is quite inadequate ... There followed a careful and systematic demolition of the “proof”.

Surprisingly, our frequent visitor did not seem pleased with the photocopies I presented to him at the next University practice, and became noticeably more reticent on matters of ringing theory thereafter!

The relevance of all this (for this is where the story really starts) is that the final paragraph of Brian Price’s letter read as follows:

Recently I persuaded the staff of the Computing Machine Laboratory at Manchester University to try composing such a [bobs-only] peal, on a 21-part plan I evolved, but I have not heard from them conclusively, and I hope to report later. They have kindly consented to run the problem at slack periods, and estimated that it would take 24 hours altogether. Considering that this 21-part plan is but a flea-bite of 1-part composition and that the machine can perform 600 12-figure multiplications in a second, one realises the immensity of the task; last August [i.e. in 1952] the machine actually succeeded in compiling a block of 40 sixes, which, however, did not join up to the start, and so the extent is in 21 unjoinable blocks. This was no mean feat, although useless for our purposes.

I think this must surely be the first attempt to compose a peal using a computer, and it is all the more remarkable when we consider the history of computing up to that time.

The locally-designed Manchester Mark I machine ran its first program on June 21st 1948 and is now acknowledged to be the world’s first stored-program computer to operate. This prototype was dismantled in 1950, and replaced by a commercial production version in February 1951. This machine (the Ferranti Mark I, of which 9 were eventually built) was the world’s first commercially available computer to be delivered, and was the one on which Brian’s program was run.

Thus, within four years of the first prototype computer running, and only two years after one became available for general research, ringers had discovered its potential. Peal composition is actually one of the earliest computer applications of any kind, and probably one of the very first non-numerical applications.

It is not surprising that Brian Price would be the man to achieve this “first”, as he had already developed an interest in ringing machines, and had published a number of articles on the subject in the Ringing World during 1950, of which I have incomplete photocopies.

The series begins on March 17th with an extensive description of John Carter’s machine, then in the Science Museum in London (now in Birmingham, as stated by Phil Barnes) and then goes on to describe G F Woodhouse’s Machines (he actually built 9!) mentioned by Ann Thompson in her recent mail message, Brian’s own machine (mentioned by Phil Gay) and a machine built in 1935 by the Compton Organ Company of Acton, which used electric organ technology to synthesise bell sounds and could ring (only) Oxford Treble Bob Major. When installed at Albury (Surrey) it could be heard 12 miles away, but caused complaints that the ringing was too perfect! There was another short article about Brian’s own machine on June 29th 1951.

Whilst on the subject of ringing machines, US readers especially may be interested to know that the Ringing World of December 26th 1958 has an article by E A (Teddy) Barnett describing a change-ringing machine designed by Professors R J Walker and J Kiefer of the Mathematics Department of Cornell University and built by Donald Stevens. Does this machine still exist? Apparently Cornell library holds many ringing books, including a 1733 edition of Campanologia Improved brought back from England by Andrew D White, Cornell’s first President.

But I digress. This might have been the end of the story, had not Michael Broe (last heard of in Chicago; has he been ringing there yet?) drawn my attention in 1987 to the book Alan Turing: the enigma of intelligence by Andrew Hodges. This is a biography of the British mathematician Alan M Turing whose 1936 paper on “Computable Numbers” introduced the abstract “Turing Machine” dear to computer scientists and deduced the theoretical limitations of computers 12 years before any such machine existed.

Turing became involved with the design of real computers through his work as a cryptanalyst during the war (read the book, it’s gripping stuff), and eventually ended up as deputy director of the Computing Laboratory at Manchester. Here he worked on a variety of topics, including biological morphogenesis, before dying in mysterious circumstances in 1954.

On page 445 (Unwin paperback edition, 1985) Hodges writes:

It was not all work on biology. He had, in particular, a ‘bell ringing’ program. Bell ringing? Working through every possible permutation? For whom did the bell toll? Ask not ...

Could it be that the Stedman-composing program had actually been written by the great Alan Turing, every computer scientist’s hero? I resolved to find this Brian Price and ask him myself.

This was easier said than done, for the 1953 letter gave a long-obsolete address in Wales. Starting from a chance remark by Giles Thompson who had corresponded with Brian at a boys’ school in Taunton in the 1960’s, I spent several weeks during the summer of 1987 on the telephone to various school secretaries, headmasters and former pupils, and eventually tracked him down, semi-retired and working part-time as a computer programmer in the Management Science Department of Imperial College.

Since I work at Imperial College myself, this was a rather remarkable end to my search. Over a lunch of institutional pudding-with-custard Brian explained that he had given up both ringing and teaching because of a hearing problem and stopped composing because nobody seemed interested in his productions (mainly odd-bob peals of Stedman Triples). I reassured him that there had been a resurgence of interest in Stedman, and that his 5,090 of Cambridge Major was still much rung as the only tenors-together alternative to Middleton’s.

The upshot of this was that Brian renewed his interest in composing peals of Stedman Triples and was persuaded to try the same technique on Cambridge Surprise Major, resulting in many novel tenors-together compositions. He also took up ringing again, first on handbells (eventually ringing one of his own new compositions) and then tower bells. He is a currently a member of the band at St Mary’s Willesden, although we don’t see him that often because he has taken to living in Thailand for the greater part of the year.

But I digress again. What of the 1952 Stedman-composing program? Brian certainly knew of the Turing biography, for he had actually met Hodges while the latter was researching his book. However he didn’t know who had written the program, because he had not corresponded directly with Turing at Manchester, but with “someone called Brooker”.

R A (Tony) Brooker is well known today as the co-inventor of the Brooker-Morris Compiler-Compiler (the great-grandmother of YACC for Unix hackers with no sense of history), and now a professor at Essex University. Amazingly he remembered the events of 1952 and assured me that he had written the program himself to Brian’s specification.

A disappointment, but clearly inviting another question. How did Hodges get the idea that Turing had written a bellringing program?

It just so happened that an ex-University of London ringer (Ian Roulstone) was then studying for his Ph.D. in Roger Penrose’s (of black hole fame) research group at Oxford, supervised by none other than Andrew Hodges. Ian was charged with finding the answer. Eventually it arrived.

It seems Hodges got a letter from someone called Brian Price ...

The story was printed more or less unchanged in the Ringing World of 20th February 1998. Several years further down the line it can be revealed that “Q.E.D.” was the late Eric Nixon, who was a student at Manchester in the 1950’s, becoming an lecturer in Computing at the University of London Institute of Computer Science in the 1960’s and 70’s and ending his career in the Computer Science Department of University College.

Eric sadly died in 2002, but I’m pleased to report that Brian Price is still with us in spite of a serious road accident a few years ago. He’s still ringing and composing ground-breaking stuff (all-the-work spliced these days), and was 85 in May.

Manchester University built a replica of the Mark I for the 50th anniversary of the first program, and it’s all described on their excellent website with pictures (and even a downloadable simulator!) to be found at: http://www.computer50.org/

HTML version by Roger Bailey, March 2003, updated May 2010. Arrived by search engine? Click here for a comprehensive list of other change-ringing links.