This is from a talk by Edsger W. Dijkstra, "How Computing Science created a new mathematical style", 4 March 1990:

Almost all formalisms used daily by the classical mathematician are at least ambiguous. But that does not hurt the classical mathematician because he does nothing with a formula without interpreting it and part of his professional competence consists in subconsciously rejecting all unintended interpretations. (If you show him

sin($\alpha+\beta$) versus sin(s+i+n)

it depends on his sense of humour whether he is amused.) The manipulation of uninterpreted formulae requires unambiguous formalisms [...]

Tragically, I am not quite sure what is so funny. Concatenating three English letters together with plus signs to make a function name is amusing I suppose, but I get the feeling there is more going on here, and I am not able to put it into words.

Is anyone able to explain these expressions, and why they could be considered humorous?


The joke is that there is an ambiguity on whether to read it as the polynomial $s\cdot i\cdot n\cdot(s+i+n)$ or as $\sin(s+i+n)$, as opposed to $\sin(\alpha+\beta)$ which is unambiguous because the arguments of the $\sin$ function are from a different alphabet.

  • 6
    $\begingroup$ +1, and not very funny... $\endgroup$
    – Albert
    Oct 11 '19 at 14:11
  • $\begingroup$ This is how I understood it too. And yah, not funny. $\endgroup$ Oct 11 '19 at 14:12
  • 7
    $\begingroup$ @Glougloubarbaki : Well, I guess whether you find it funny "depends on [your] sense of humour". $\endgroup$ Oct 11 '19 at 14:14
  • 2
    $\begingroup$ There is no bass boosted effect, therefore it doesn't qualify as comedy to me. $\endgroup$
    – user239203
    Oct 11 '19 at 14:16
  • 2
    $\begingroup$ I observe that when typeset, $\sin(s+i+b)$ cannot be conflated with $sin(s+i+n)$. But this leads to a recurring error in typesetting: whose calculus book correctly uses upright e (that is, $\mathrm{e}$, not $e$) for Euler's constant and upright pi (that is, $\unicode[Times]{x3C0}$ (rendering depends on details of fonts installed on your computer), not $\pi$) for Archimedes' constant? $\endgroup$ Oct 11 '19 at 14:23

From reading the first answer by Gae. S, it is obvious that some of ambiguity stems from the "invisible multiplication sign". This same author, Edsger W. Dijkstra, has written extensively on the proper choice of notation in other articles, and has criticized the use of this type of implied notation.

There is EWD1059 which I had read previously. Even better is EWD1300 which I had not read yet, but was able to find based on the revelation provided by the earlier answer. This article actually includes another example of this. The author notes that the "invisible multiplication sign" (the dot, when written out) has been hailed as progress through brevity. So, this seems to be another example of him poking fun at the fact that brevity can also yield ambiguity.

(Several weeks after asking the original question, found EWD1115, which is not yet transcribed or searchable on the UT website, but which contains another partial explanation for this on page 3.)

In the context of his other writing, Edsger W. Dijkstra would say that a complicated problem can be made simple by an appropriate choice of notation (and vice versa: poor notation can make a simple problem unwieldy or practically unsolvable). While his examples are seldom laugh-out-loud "funny" and may not quality as comedy, some of the humor is derived from showcasing mini-disasters that could have been easily avoided.

An additional explanation came to me unexpectedly later last evening, when I picked up the book "Surely You're Joking, Mr. Feynman!" for the first time. At the end of the very first chapter he writes:

While I was doing all this trigonometry, I didn't like the symbols for sine, cosine, tangent, and so on. To me, "sin f" looked like s times i times n times f!

Simply stunning.

  • $\begingroup$ Interesting ! A question : how can we access the "top level" of the UT site containing these texts by Dijkstra ? $\endgroup$
    – Jean Marie
    Feb 4 '20 at 12:41
  • $\begingroup$ @JeanMarie Do you mean: cs.utexas.edu/users/EWD $\endgroup$ Feb 4 '20 at 18:41
  • $\begingroup$ Excactly. Thank you very much ! $\endgroup$
    – Jean Marie
    Feb 4 '20 at 18:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.