Zulip Chat Archive

Stream: general

Topic: forgetful inheritance

Rob Lewis (Feb 07 2020 at 13:34):

Just came across this nice paper from the Mathematical Components crowd: https://hal.inria.fr/hal-02463336/file/submitted.pdf

The perfectoid people get called out in the conclusion :)

As explained in Sect. 3.3, type classes also allow for a variant
of forgetful inheritance, although with a less robust proof search. This was also
observed before, e.g., by Buzzard et al. [11, Sect. 3]. Interestingly, it seems that
the authors have not identified that another issue they raise, about completing
groups and rings [11, Sect. 5.3], pertains to the same problem, and can be solved
using forgetful inheritance as well.

Kevin Buzzard (Feb 07 2020 at 14:51):

I am absolutely serious that I would rather this project not be known by "Buzzard et al", especially as one of the "al"s doesn't have a permanent position. We ordered the authors alphabetically, as is standard in mathematics. Was that a mistake?

Jasmin Blanchette (Feb 07 2020 at 15:09):

Kevin, I'm afraid that train has already left. If you want to explicitly promote the junior, you have to move the name first. Computer scientists start using "et al." at n = 3.

Kevin Buzzard (Feb 07 2020 at 15:15):

In maths, 99.9% of the time it's done alphabetically (and also, I guess, most of the time there are n<=2 authors). Would it have been acceptable to re-order our names? In maths this would be considered quite strange.

Greg Langmead (Feb 07 2020 at 15:21):

You could try referring to it yourself as "BCM" and see if it catches on. CS folks often abbreviate papers that way, especially papers with clout. It counteracts some of the "et al" problem.

Jasmin Blanchette (Feb 07 2020 at 15:24):

You can do whatever you like in CS. Junior-first, main-person-who-did-the-work-first, alphabetical, reverse alphabetical, main honcho last. This recent submission of mine combines main-worker-first and reverse alphabetical:

http://matryoshka.gforge.inria.fr/pubs/saturate_report.pdf