2015-01-05 15:55:33 UTC+1, Peter Bruin wrote:
> According to the definition given in the docstring of chi.bernoulli(), it
>> should be 1/2, not -1/2. This also matches the convention used in
>> Washington's "Cyclotomic Fields", and also what I find in the wikipedia
>> article on generalised bernoulli numbers.
>>
>
> This should indeed be fixed (either the code or the documentation). We
> also have
>
> sage: chi = DirichletGroup(1)[0]
> sage: chi.bernoulli(1, algorithm='definition', cache=False)
> 1/2
>
> With the default algorithm='recurrence' option, the answer -1/2 is caused
> by the fact that Sage uses the bernoulli function, which is often defined
> to take the value -1/2 at 1 ("first Bernoulli numbers"):
>
> sage: bernoulli(1)
> -1/2
>
> I'll open a ticket for this. The question is whether we want
> chi.bernoulli(1) to return 1/2 (matching the definition of generalised
> Bernoulli numbers) or -1/2 (matching the bernoulli function). I would
> personally be in favour of 1/2, but maybe others here have a different
> opinion.
>
Related reading:
http://luschny.de/math/zeta/The-Bernoulli-Manifesto.html
--
You received this message because you are subscribed to the Google Groups
"sage-nt" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send an email to [email protected].
Visit this group at http://groups.google.com/group/sage-nt.
For more options, visit https://groups.google.com/d/optout.
