Earlier this week Don Knuth turned 70. This posting, inspired by a coordinated series of posts by admirers in mathematics and computer scientists, is my own personal statement of appreciation of his work. Mostly, I’ve been influenced by TeX, which Don Knuth wrote both as a labour of love, and as a means to the end of writing his multi-volume work The Art of Computer Programming (which is also a labour of love).
My first real contact with TeX was in 1987, five years after the first release of what has become the current version of TeX. Now the version number of TeX is, famously, converging to pi, but then it was something like 2.1. Previously I had written my PhD thesis out by hand, and skilled technical typists typed it up page by page, using the admirable IBM golfball typewriters. To obtain a special symbol, the typists had to manually change the golf-ball sized typehead. In 1986/7 I wrote a mathematics paper user the eqn preprocessor and the troff typesetting system that came with Unix. I was able to do this only because I had a lot of help from a technical typist who had several years experience with troff.
So this was my first exposure to TeX. I saw what was clearly the most beautiful output produced using an input syntax that was, compared to the intricacies of troff, simple and clear. What also came through quite quickly was the extraordinary quality of TeX the program. I decided quite soon that I would learn TeX, and that I would in future have the ability to typeset my own mathematical writings. TeX gave me independence. In 1988 I bought a PC and laser printer, mainly so I could run TeX on it. At that time the only TeX distributions for PCs were commercial, and my came on an enormous number of 360Kb 5.25 floppy discs.
A few words about the PC I purchased. It had a 10 MHz 80286 processor, and a 40 Mb hard disk drive. (No, the Mb is not a mistake. It really was 40 Mb, not 40 Gb.) I was advised to buy a powerful machine because TeX was a demanding application, and I could not afford to by the high-end 386 machines that were then coming out. It took about 1 second to process a page.
It is, for me, a testament to the quality of TeX that I am continuing to use the same program today, 20 years later, for typesetting mathematics. At work I am sort of the TeX system manager for the production of mathematics and physics course materials, and we have items that are 15 years old. They still run, with modern distributions of TeX, and still give effectively identical output. I’m trained as a mathematician, and just as 2 + 2 is always equal to 4, so
$x^2$ is always typeset by TeX in a certain way. Documents written in TeX have an enduring quality (as long as they are not stored on 5.25 inch floppy discs).
The arXiv for mathematics and physics preprints is built on TeX and the Web (and many other achievments in computing, such as Unix and the Internet). The majority of its articles are written in TeX. Particularly in mathematics and physics, TeX has made possible electronic communication of complex ideas, and so has rightly become a standard. Today, learning TeX is a technical component of writing a PhD thesis in mathematics. This is very different from my experience over 20 years ago.
We tend, when we look back, to see an historical inevitability to events. In part this is an attempt on our part to understand the past so that we can predict, if not influence, the future. As Niels Bohr remarked “Prediction is very difficult, especially if it’s about the future”. But at the time of great events, when faced with a difficult challenge, we can be subject to doubts and fears.
Today its users take TeX for granted, but neither its existence nor its admirable qualities are solely due to the force of circumstances and history. TeX’s existence and qualities are also due to the diligence, energy, intelligence and wisdom of Don Knuth, and those around him who were also part of the TeX project at Stanford, the American Mathematical Society and elsewhere (including people such as myself who were inspired by TeX in the 1980s).