Zum Inhalt springen

Dijkstra's Pen: Why the Greatest Computer Scientist Never Used a Word Processor

Zusammenfassung

Edsger W. Dijkstra wrote every paper, memo, and letter of his career by hand with a fountain pen — never using a typewriter or word processor. His 1,300+ EWD (Edsger W. Dijkstra) manuscripts, circulated in handwritten photocopies to colleagues, constitute one of the most influential bodies of work in computer science. His reasoning: a word processor allows revision without cost, which encourages imprecision. The difficulty of handwriting forces clarity of thought before a word is committed to paper.

The EWD Manuscripts

Between the early 1960s and 2002, Dijkstra wrote a series of technical notes that he numbered sequentially with his initials: EWD1, EWD2, … EWD1318 (the last). These ranged from brief half-page observations to 30-page technical essays. He photocopied them and mailed them to a list of colleagues who found them worth reading. They were not published in journals; they circulated through personal networks.

The manuscripts covered formal verification, programming methodology, distributed systems, educational philosophy, and professional ethics. EWD1036, “On the cruelty of really teaching computing science” (1988), argued that computing science should be taught as a rigorous, mathematical discipline. His 1975 paper on guarded commands (EWD418) introduced what became known as predicate-transformer (“weakest precondition”) semantics. EWD316, “A Short Introduction to the Art of Programming” (1971), was used as teaching material at universities for decades.

The entire EWD collection has been digitally preserved at the University of Texas at Austin, where Dijkstra spent the last years of his career.

Why Handwriting

Dijkstra was explicit about his rejection of word processing in several EWD manuscripts. His argument:

  1. Revision without cost produces prose without discipline. A word processor makes it trivially easy to change a sentence; the writer who can revise endlessly never commits to a precise formulation. The pen writer who must cross out and rewrite is incentivized to get the sentence right before writing it.

  2. The appearance of completion is false with digital text. A printed document looks finished even when its ideas are muddled. A handwritten document’s density and corrections are visible; the reader can see where the author struggled.

  3. Mathematics is not served by word processing. Mathematical notation requires care and spatial arrangement that word processors (especially in Dijkstra’s active years) rendered poorly.

His handwriting was extremely neat and consistent — a trained calligraphic hand rather than a hurried scrawl. The EWD manuscripts are legible; the formatting (margins, spacing, inline mathematics) is deliberate. He spent time on the physical document in a way that word processor users typically don’t.

The Broader Philosophy

Dijkstra’s rejection of word processing was of a piece with his broader insistence that programming be treated as a mathematical discipline requiring formal proof rather than empirical testing. Just as he argued that a program tested to work was not proven to work, he argued that a text revised to read well was not proven to think clearly.

His most famous phrase — “Program testing can be used to show the presence of bugs, but never to show their absence” — describes the same epistemological gap between empirical evidence and formal proof that characterized his thinking about writing. You can revise a text until no errors remain visible. You cannot revise it into clarity without having thought clearly in the first place.

He received the ACM Turing Award in 1972. His acceptance speech is available in both handwritten EWD form and printed transcript.

📚 Sources