|L. J. Comrie|
© British Astronomical Association
Leslie John Comrie (1893-1950), born at Pukekohe (South of Auckland), New Zealand, and educated at Auckland University College and University College London. He became an Isaac Newton Student at Cambridge, where he received a Ph.D. in astronomy, his thesis being on the occultation of stars by planets. During World War I he saw action in France where he lost a leg in fighting (he was also deaf in one ear). While recovering Comrie attended some classes in practical computation given by Karl Pearson at University College, London and it was here that he learnt how to use a Brunsviga calculator, and went on to modify commercial calculators for specific projects.
He was the first director (1920-1922) of the Computing Section of the British Astronomical Association (BAA), resigning when he went to Swarthmore College and Northwestern University in the USA (1923-1925), where he pioneered the teaching of numerical analysis.
In 1925 he returned to England to join HM Nautical Almanac Office. From 1922 to 1939 the Office was on the first floor of King Charles' block in the Royal Naval College at Greenwich. Comrie was appointed Superintendent in 1930.
In 1938 he founded the Scientific Computer Service the world's first computing consultancy firm. During World War II he headed a team of 30 scientists to computerise war work, e.g., bombing tables for the USAF. Later he computerised British football pools.
Comrie was elected a Fellow of the Royal Society of London shortly before his death. A lunar crater (23.3N 112.7W) and a minor planet bear his name, as does the computer laboratory at his alma mater, the University of Auckland.
In April his article On the Construction of Tables by Interpolation described the use of punch card equipment for interpolating tables of data, and comparing this with the less efficient and more error-prone methods using mechanical calculators. Also in 1928, he was the first to use punch card equipment for scientific calculations, using Fourier synthesis to compute the principal terms in the motion of the Moon for 1935 to 2000. It is described as:
A Hollerith installation was used in H.M. Nautical Almanac Office in 1929 [Brown's Tables of the Moon; moon's position every 12 hours from 1935 to 2000] actually punching was started six months before the arrival of the sorter and tabulator, as it was necessary to punch 20,000,000 holes in half a million cards. ... The sorter and tabulator, which may only be hired, cost about £2 a day. The cost of doing by hand what was done on the machines has been estimated at £6000; this estimate probably errs on the low side. The cost with the machines was certainly less than £1500 .
In Sadler's Personal History  he noted that the first thing he remembered Comrie saying to him (during his interview) was that he did not suffer fools gladly! It is clear from Sadler that Comrie exacted the highest standards from his staff, and although he was generous with praise for jobs well done, with bonuses, he was intolerant of careless checking errors, unpunctuality and other things. At one time he caused a minor revolution by objecting to the women members of staff going to 'wash's in office time shortly before leaving work!
The staff of the Office consisted of the "established" staff, three Junior Assistants, and the post of Assistant Superintendent (filled by Sadler, see below). There were also a number (6 or 7) of un-established (temporary) staff, recruited locally by the Superintendent and paid out of the Lump Sum. Comrie appears to always be in a constant battle with the Admiralty about the complement and the number of established and un-established staff. In fact, you get the impression from Sadler's history that he was obsessed about staffing, and treated the Admiralty and "Civil Service" as fools.
Comrie enjoyed the status of consultant as many scientists and organisations, e.g. Hartree, Jeffreys, Watson-Watt, the Colonial and Military Survey, Armament Research, came to him for advice on their computing problems. He also got himself and his staff involved in the calculations, some of which were official and some were not. This included work for the BAA. Sadler is of the opinion that Comrie had:
no hesitation in "balancing" work done by "his" staff for the Office against work done by the Office staff for him. I have no doubt that Comrie considered he operated this agreement fairly and with mutual advantage, but it created difficulties of priority within the Office.
One particular job, known as "Winds" was an official job as it was classified, it required massive calculations, and involved an increasing number of the junior staff. The final straw came when Comrie again wrote to the Admiralty demanding more staff, saying that it was impossible to find the time to prepare copy for the Nautical Almanac Abridged for Seamen (ANA) for 1937 - and if the Civil Establishment Branch did not understand the seriousness of that, the First Sea Lord would!
On the 19th August 1936 a small investigating team from the Admiralty, descended on the office without warning. They of course found the ANA 1937 listings complete (since W.A. Scott was in charge of this). They suspended Comrie from duty forthwith; they impounded all his work on Winds and other material, and put Sadler in charge, with instructions not to communicate with Comrie. There was a formal enquiry, and as a result Comrie's appointment as Superintendent was terminated .
When Comrie joined HMNAO he stated that, "The Nautical Almanac was computed by retired Cornish clergymen with long white beards, using dog-eared 7-figure logarithm tables" [XXXX]. He built up the number of permanent staff, whereas Cowell had intended to revert to short-term contracts. But most importantly he revolutionised the computing methods employed. The bulk of the calculations had been done by hand using logarithms, by the time he left in 1936, the Office was using a range of calculating machines including Brunsvigas, Burroughs Machines and National Accounting Machines, and, on a more limited scale, Hollerith punched card machines .
Comrie was probably most widely known as a maker of mathematical tables which were renowned for their accuracy and typographical design. Wilkins  gives a summary of the factors that determine the quality of a numerical table apart from the prime requirement that the values be free from errors. They include the choice of the quantities, the interval and number of figures for each function; the overall design, layout of the page, the headings, typeface, type-size, spacing and rules. Of course the end result must be a table that is easy to use.
The mathematical tables produced by Comrie showed very clearly the benefits of using white space rather than rules to separate the columns and rows of the numbers. Unfortunately, the demands of economics and the desirability of making all the required data fit into the space available means that this cannot be applied throughout the almanacs produced by HMNAO. A further factor was that the main almanacs were produced jointly with the Nautical Almanac Office of the US Naval Observatory, which in turn had to satisfy the US Navy. Comrie also split the numbers into small groups and changed from level figures to head-and-tail figures in order to reduce the risk of errors in the reading from the tables.
Maker of mathematical tables - the many mathematical tables published by Comrie were acclaimed as the finest tables ever produced, and he was the acknowledged king of the British mathematical table making community in the 1930s. .