When Donald Knuth wrote the TeX typesetting system and the books that document it, he attached an unusual guarantee: he would pay readers to find his mistakes. On his Stanford pages he sets the rates plainly. “If you do succeed in finding a previously undiscovered bug in the programs for either TeX or METAFONT, I shall gladly reward you with 0x$80.00 ($327.68) at the Bank of San Serriffe. Corrections to errors in The TeXbook or The METAFONTbook are worth 0x$1.00 ($2.56), as in all my other books.”
The strange-looking amounts are a programmer’s joke that doubles as a unit of account. A book error is worth $2.56, which is exactly one hexadecimal dollar: 256 pennies, written 0x$1.00. The reward for a genuine software bug, 0x$80.00, is that base unit doubled repeatedly until it reaches $327.68. Knuth has long described the program rewards as doubling over time, an open-ended promise that turns finding bugs into a small competitive sport among his readers.
The checks themselves became collector’s items. They are drawn on the Bank of San Serriffe, an invented country named after the typesetting term sans-serif, and Knuth’s page notes that the bank “was founded in October 2008” to keep the accounts. The leaderboard of balances reads like a hall of fame for careful readers, with top finders holding hundreds of hexadecimal dollars. Because a Knuth check is proof that you out-read Donald Knuth, recipients have famously framed the checks rather than cashing them, which is part of why the scheme has cost him far less than the totals suggest.
Knuth is candid that the books were not enough. His page records that the account balances shown “reflect only the rewards for errors reported since the beginning of 2006,” and that “many reward checks were issued during 1968 to 2005, too numerous to mention here.” Decades of corrections have flowed in, each one making the next edition a little more correct.
The bug bounty is more than a quirk. It is a public statement about how Knuth thinks software and scholarship should be made: slowly, openly, and with a standing invitation for anyone to prove him wrong. The same instinct runs through his insistence on literate programming and his decades-long care over The Art of Computer Programming. Paying for bugs simply puts a price, however small and however symbolic, on the truth that correctness is a shared and never-quite-finished project.