I have no opinions about what B_1 should be, but I am concerned about potential confusion: some users will expect one value for B_1, others will expect a different value, and so one group or other will end up being confused when answers don't come out the way they expect. The safest course might be to drop "bernoulli" altogether and have two functions for the two different conventions. Or keep "bernoulli" but have it just print a message saying that there are two conventions, so instead use either "bernoulli_plus" or "bernoulli_minus" (or whatever they might be called). I don't know if the safest course is the right course, though.
(This is how I feel about the "range" of a function in mathematics: some people use this word to mean the image, some people mean it to use the codomain. Since there is serious possibility of confusion, I personally just never use the term at all.) -- John On Tuesday, September 13, 2022 at 4:15:07 AM UTC-7 redde...@gmail.com wrote: > On Tue, 13 Sept 2022, 17:00 David Joyner, <wdjo...@gmail.com> wrote: > >> Let's play nice here, okay? > > > Let me explain what I mean in a nicer way. Not defining B_1 looks good on > the surface given the current discussion, but is really a strictly worse > option than defining B_1 = +½ or -½ because then the n = 1 case has to be > special-cased out of formulae where B_1 is normally used. > > That B_n for odd n at least 3 is zero may provide computational > advantages, but is a non-trivial fact. It is more parsimonious to put the > even, one and greater-odd instances of the Bernoulli numbers into one bag. > > Jeremy Tan / Parcly Taxel > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/6c6859b5-e7b2-4e17-bcf9-30994a7048f4n%40googlegroups.com.