Skip to content

Dijkstra's Pen: Why a Computer Scientist Wrote by Hand

Abstract

Edsger W. Dijkstra circulated more than 1,300 technical notes, the “EWDs”, over four decades, photocopied and mailed to colleagues who passed them on. Many of the early ones were typed, on his own Hermes typewriters; from about 1979 he wrote almost everything by hand with a fountain pen, in a neat, even script. When he was accused of using his handwriting “as a cheap trick to enhance my visibility”, he answered with a practical reason: mathematical symbols were hard to get on a typewriter’s daisy wheel, but “on my fountain pens I get them for free.”

The EWD Manuscripts

From the early 1960s until 2002, Dijkstra numbered his notes with his initials: EWD1, EWD2, up to EWD1318. They ranged from half-page observations to long technical essays, trip reports and polemics. He sent copies to a list of colleagues, who copied them again; most never appeared in journals.

The notes covered formal verification, programming methodology, concurrency, teaching and the profession. EWD316, “A Short Introduction to the Art of Programming” (1971), was used as teaching material. EWD418 introduced guarded commands and the weakest-precondition semantics he published in 1975. EWD1036, “On the cruelty of really teaching computing science” (1988), argued that computing should be taught as a mathematical discipline.

The collection is digitized and transcribed by the University of Texas at Austin, where Dijkstra taught from 1984 until he retired in 1999; he then returned to the Netherlands and died in Nuenen in 2002.

Typewriter and Pen

In EWD1000 (January 1987) Dijkstra looked back on his writing tools. He had typed his own manuscripts on Hermes machines, which was faster than going through a typist, and had the digits fitted from a different type fount so that he could tell the digit 0 from the capital letter O. The switch to handwriting, he wrote, was practical: special characters, which a mathematician needs constantly, were difficult to obtain for a daisy-wheel typewriter and cost nothing with a pen. He did not adopt a word processor for his own writing.

His handwriting was regular enough that the manuscripts read almost like print, and in the late 1980s Luca Cardelli, then at DEC’s Systems Research Center, turned it into a “Dijkstra” font for the Macintosh. Dijkstra received a letter set in it and took it for handwriting until he heard about the font; he was not amused at first. The habit fitted his general insistence that programs and proofs be thought through before they are written down, the same conviction behind his best-known line: “Program testing can be used to show the presence of bugs, but never to show their absence.”

He received the ACM Turing Award in 1972. See Edsger Dijkstra and Structured Programming.

📚 Sources