Jeremy Tan Jie Rui

@mbostock I think this is ready to merge. This will serve as another pull request to include for my CS3282 module.

> Do any browsers support this yet? I’d be more inclined to add it if it had browser support. @mbostock As far as I know, the CSS Color Module Level...

This is the eternal, unquestionable truth: >For any software, defining $B_1=-\frac12$ is a **critical** bug. Therefore, rather than letting the user choose, we should simply fix $B_1=+\frac12$ in the code...

>It's important to be clear here that there is no "correct" or "incorrect" in this situation. There are two possible conventional choices for the value of $B_1$. Totally false. Luschny's...

https://github.com/fredrik-johansson/mpmath/pull/639, once merged, will set $B_1=+\frac12$ in that package, allowing us to not (really) worry about compatibility with mpmath. We might need to ping @fredrik-johansson and tell about this project...

> Also probably we should for evalf use Lushcny's formula to make `bernoulli` a function defined on the complex numbers. Does that mean allowing non-nonnegative-integer input to `bernoulli()` itself as...

>Luschny's definition makes `B` a holomorphic function if I understand correctly. If we're adopting that definition then I think `bernoulli` should allow any complex number as input. There I run...

> Maybe there should be a non-symbolic function in ntheory for calculating the numbers in the bernoulli sequence. Then it could be used by both zeta and also bernoulli. ~~[I...

**Done.** All that remains is to merge my pull request into SymPy 1.11.

@oscarbenjamin #23973