Authors should be in a deterministic order irrelevant to contribution status, but alphabetical means that people with last names that start with A always get to be first.
Proposal: order by hash, using the year of publication as a salt.
Mar 7, 2026, 5:04 PM@olynch In the acknowledgments section of my book, some of the lists of contributors are in reverse alphabetical order, to help balance the universe a bit.
Mar 7, 2026, 7:43 PM@olynch I thought this seemed like a good idea, but after thinking on it carefully, I see a serious problem. It's fair, yes, but the division of first-authorships might still be grossly inequitable. Professor Badluck might finish out his distinguished career with no first-authorships, just because of bad luck.
I think maybe a better strategy would be something along the lines of the Fagin-Williams “fair carpool scheduling algorithm”. Each paper is allotted a certain value V, say its length multiplied by the impact factor of its journal. The N-1 other authors each pay V/N "tokens" to the first author A, who has been chosen because A, among all N authors, has the _least_ tokens. Fractional tokens are allowed.
(If there's a tie, choose at random; the loser will have an advantage next time.)
How to fairly allocate V is in interesting question, and if it can't be well-solved it might sink the whole scheme. For carpooling, it's a simple constant.
https://bitsavers.org/pdf/ibm/IBM_Journal_of_Research_and_Development/272/ibmrd2702T.pdf