Pages: pp. 2-3
When Nevil Maskelyne, the British Astronomer Royal from 1765−1811, wanted to describe a calculation, he would create a grid. Taking a sheet of paper, he would fold it into quarters and divide it into cells with lines drawn with quill-tipped pen. Next to each cell, he would write an arithmetic expression. Take a value from one cell, the expression might suggest, multiply it by a constant and divide the result by the contents of a second cell. The value, thus obtained, should be placed in the adjacent cell. If he needed to explain an idea, he would sketch a graph on the back of the paper.
Maskelyne would work from left to right and from the top of the sheet to the bottom. When he was done, he would have completely described the computation for an ephemeris, a table that might describe the motion of Saturn or Venus or one of the other planets against the background of fixed stars.
Maskelyne rarely did the actual computations for these plans. He generally assigned this work to an assistant, usually called a human computer. These descriptions were called computing plans and were commonly used by the Nautical Almanac and Surveying offices of the 18th and 19th century. Gaspard de Prony used them at the French Bureau de Cadastre in 1790. Charles Henry Davis employed such methods at the office of the American Almanac and Ephemeris in 1848. L.J. Comrie, who liked to write plans on accounting paper, continued the tradition of Maskelyne at the Royal Almanac through the 1920s. Gertrude Blanch, who preferred graph paper with five squares to the inch, created an extensive set of plans for the Mathematical Tables Project between 1938 and 1948.
Even with such a tradition, grid-based computing plans quickly vanished after the introduction of the electronic computation. Among the early pioneers, only Konrad Zuse seems to have considered using such plans to control an electronic device. Today, we generally view his proposed programming language, Plankalkül, as a creative alternative to the sequential languages of the 1950s, a novel idea that went nowhere. Yet, Zuse's language was built upon a computing tradition with at least 150 years of history, a tradition that was still an active craft in the late 1940s and early 1950s.
The old traditions of the computing plans were temporarily displaced by the algebraic languages of the 1950s: Fortran, Cobol, Algol, Lisp, and APL. However, they were revived in 1978 and 1979 by Dan Bricklin, Bob Frankston, and Dan Fylstra with the creation of the modern spreadsheet. The trio released their first product, VisiCalc, in 1979. In less than two years, they had sold 200,000 copies. By 1985, it had become second only to word processing as the dominant application for personal computers.
The connection between VisiCalc and the computing plans of Maskelyne, Blanch, and Zuse is tenuous at best. None of the creators of the spreadsheet have ever pointed to the early human computers as a source of inspiration. Yet, the long history of using grids to describe computation, plus the rise of modern accounting forms in the late 19th century, suggests that grids are an important way of describing calculations and automated grids would, sooner or later, have an important place on electronic computers.
This special issue on the spreadsheet has been organized by Burt Grad and Paul Ceruzzi, who are familiar to regular readers of this journal. Burt and Paul organized a 2006 special issue for Annals on the history of word processing; Burt has also previously organized a 2003 special issue on the start of the software industry (done jointly with Annals editorial board member Luanne Johnson). Burt heads the Software History Institute, which operates in conjunction with the Computer History Museum in Mountain View, California. Burt has been a good friend to the Annals and has produced a wonderful issue. His efforts to recruit contributions from the leaders of the early spreadsheet industry were undaunted and very successful. Paul, of course, brought his keen historian's eye and his sense of context. We are grateful to the efforts of both for their fine work.
In this special theme issue, Martin Campbell-Kelly's "Number Crunching without Programming: The Evolution of Spreadsheet Usability" traces how the modern spreadsheet evolved from a simple calculating aid to an essential business tool. From an "in the trenches" perspective, we have two "recollection" articles by Mitch Kapor and Jonathan Sachs, principal figures in the development of Lotus 1-2-3. These articles, which are based on oral histories conducted by the Software History Center, give a first-person view of the challenges, pressures, and the general milieu of early 1980s software development. And, of course, we have Burt's article on the spreadsheet software that began it all: VisiCalc. This unique narrative is an amalgam: on the basis of several oral histories, it weaves the story of how VisiCalc came about, evolved, and ultimately withered.
Additionally, this issue also features "Computing at the Malta Statistics Office, 1947−1970" by Mario Aloisio. This is an in-depth look at how the Mediterranean island of Malta's central government agency gathered, compiled, and produced vital statistics made possible by the introduction of tabulating equipment in the 1940s.
This issue also marks the return of Anne Fitzpatrick to the editorship of the Anecdotes department. Anne has graciously agreed to complete the work of our recently departed friend and colleague, Laurie Robertson. While we believe that we have been able to recover all the articles that Laurie was editing, we are aware that we may still be missing one or two. If you were preparing an Anecdote for Laurie, I would ask you to contact Anne Fitzpatrick at the department email address, firstname.lastname@example.org.
With this issue we can announce the next editor in chief of the Annals, Jeff Yost. Jeff is the assistant director of the Charles Babbage Institute, University of Minnesota, a long-term member of the Annals editorial board and an accomplished historian. He will begin his term with the first issue of the 30th volume in January 2008. I am proud to have him succeed me as editor in chief and pleased to be able to say that all ideas for articles or special issues should now be directed to him at email@example.com.
Finally, as this issue was going to press, we received word of the passing of one of our long-serving board members, John Todd (1911−2007). John is generally remembered as an applied mathematician but during his career he ran two offices of human computers: the Admiralty Computing Service in London and the Computation Laboratory of the National Bureau of Standards. In these roles, he was the direct descendent of Nevil Maskelyne in that he was responsible for preparing computing plans for human computers, organizing the grids that would guide complex calculations.
Readers may contact David Alan Grier at firstname.lastname@example.org.