Alexander Voigt

Hi @JP-Ellis and @zhyiyu , I see the same problem: The `photon` and `boson` propagators look pretty jagged. I could imagine one could try to increase the number of samples,...

Hi @zhyiyu , > Can we manually increase sampling points while drawing the diagram? Naively I had guessed that this should be possible. However, I had no luck so far....

Hi @dylanede , that should be straight-forward if one would do what you're suggesting, namely truncating the approximations accordingly. (I guess you're asking for single-precision versions for performance reasons, right?)...

Many thanks in advance for preparing a PR! > I have determined using that script that for the target of 2e-6 relative error, the numbers of coefficients for the numerator...

> Yes, though I did notice that for a given input number of coefficients, the number that comes out is one more than that for both the numerator and denominator,...

I've drafted a single-precision version of the real dilogarithm `li2(float)` in the branch [feature-li2-float](https://github.com/Expander/polylogarithm/tree/feature-li2-float). On my laptop it is faster by a factor ~1.4. I'll see if I can optimize...

Sorry for my slow response. Yes, I can prepare a PR. I guess the CI will check that it works on windows?

My reference value is from Mathematica 13.3.0. I've written the input value `z` as a rational and evaluated `Log[z]` with up to 17 decimal digits: ~~~ z = 499997584357227045897/500000000000000000000 -...

Thank a lot for this excellent comment! You are of course right! :) Just for my understanding: My reference value for `Log[z]` was not correct, because the rational number that...

Implementations in other languages that I found so far include: * Intel's C++ compiler has only `double _Complex cis(double)`. (I guess in C's naming convention one would have `cis` for...