Donald Knuth is so revered among computer scientists that they won’t cash his checks.

Knuth offers a standard reward of $2.56 (one “hexadecimal dollar”) to the first finder of each error in his published books. Since 1981 he has written more than $20,000 in checks, but most of the recipients have simply framed them as points of pride.

“There’s one man who lives near Frankfurt who would probably have more than $1,000 if he cashed all the checks I’ve sent him,” Knuth said in an October 2001 lecture. “Even if everybody cashed their checks, it would still be more than worth it to me to know that my books are getting better.”

Somebody had told me of a dealer in gin who, having had his attention roused to the enormous waste of liquor caused by the unsteady hands of drunkards, invented a counter which, through a simple set of contrivances, gathered into a common reservoir all the spillings that previously had run to waste. … It struck me, therefore, on reviewing this case, that the more the people drank, the more they would titubate, by which word it was that I expressed the reeling and stumbling of intoxication. … [T]he more they titubated, the more they would spill; and the more they spilt, the more, it is clear, they did not drink. … Yet, again, if they drank nothing worth speaking of, how could they titubate? Clearly they could not; and, not titubating, they could have had no reason for spilling, in which case they must have drunk the whole–that is, they must have drunk to the whole excess imputed, which doing, they were dead drunk, and must have titubated to extremity, which doing, they must have spilt nearly the whole. … ‘And so round again,’ as my lord the bishop pleasantly expresses it, in secula seculorum.

— Thomas de Quincey, Essays on Philosophical Writers, 1856

Beginning poker players are often shown a table like this:

It’s straightforward enough, assigning a hierarchy to the hands based on the likelihood of their appearance. But a strange thing happens when wild cards are introduced. Suppose we add one wild joker:

Now three of a kind is more likely (and thus less valuable) than two pair. Well, can we just reverse their places in the table? No, we can’t, because the wild card permits some players to reinterpret their hands. If you’re holding 6♠ 6♥ 7♣ 10♦ plus the joker, and we change the table, you’ll simply decide you’re holding two pair rather than three of a kind. So will everyone in your position. In fact, if we recalculate the odds with this expectation, we find that two pair has again become the more likely hand (13:1 vs. 34:1).

This can go on all day. Whenever a hand is declared “rare” it becomes popular — and thus not rare. The bottom line is that when wild cards are allowed, it becomes impossible to rank hands based on frequency.

From Julian Havil, Impossible?, 2008.

In 1986 The Mathematical Intelligencer published this story about devising a mnemonic for a famous constant:

For a time I stood pondering on circle sizes. The large computer mainframe quietly processed all of its assembly code. Inside my entire hope lay for figuring out an elusive expansion. Value: pi. Decimals expected soon. I nervously entered a format procedure. The mainframe processed the request. Error. I, again entering it, carefully retyped. This iteration gave zero error printouts in all–success. Intently I waited. Soon, roused by thoughts within me, appeared narrative mnemonics relating digits to verbiage! The idea appeared to exist but only in abbreviated fashion–little phrases typically. Pressing on I then resolved, deciding firmly about a sum of decimals to use–likely around four hundred, presuming the computer code soon halted! Pondering these ideas, words appealed to me. But a problem of zeros did exist. Pondering more, solution subsequently appeared. Zero suggests a punctuation element. Very novel! My thoughts were culminated. No periods, I concluded. All residual marks of punctuation = zeros. First digit expansion answer then came before me. On examining some problems unhappily arose. That imbecilic bug! The printout I possessed showed four nine as foremost decimals. Manifestly troubling. Totally every number looked wrong. Repairing the bug took much effort. A pi mnemonic with letters truly seemed good. Counting of all the letters probably should suffice. Reaching for a record would be helpful. Consequently, I continued, expecting a good final answer from computer. First number slowly displayed on the flat screen–3. Good. Trailing digits apparently were right also. Now my memory scheme must probably be implementable. The technique was chosen, elegant in scheme: by self reference a tale mnemonically helpful was ensured. An able title suddenly existed–“Circle Digits”. Taking pen I began. Words emanated uneasily. I desired more synonyms. Speedily I found my (alongside me) Thesaurus. Rogets is probably an essential in doing this, instantly I decided. I wrote and erased more. The Rogets clearly assisted immensely. My story proceeded (how lovely!) faultlessly. The end, above all, would soon joyfully overtake. So, this memory helper story is incontestably complete. Soon I will locate publisher. There a narrative will I trust immediately appear, producing fame. THE END.

The text explains itself: Count the letters in each word (a punctuation mark other than a period counts as a 0, and a digit stands for itself), and you’ll get the first 402 digits of π.

Astronomers have the light-year, but nuclear physicists need an analogous unit for measuring tiny distances.

Happily, they have one: The Physics Handbook for Science and Engineering defines the “beard-second” as the length the average physicist’s beard grows in one second, or about 5 nanometers.

Google will even make the conversion for you — type 1 inch in beard-seconds into your search box and see what you get.