Adam Jermyn

Implemented on the branch `other_pressure_implicit`. I just changed `other_pressure` to support partials. Users can omit them if they wish (to recover the non-implicit version).

We'll try it out for a bit before merging.

I just looked through `code_style.rst` and have two questions: 1. autodiff variables work just fine with `**` and with `pow`. Are we okay allowing `**` with autodiff types, or do...

I think we do want subroutines to be able to read `ierr` (e.g. so they can halt if `ierr /= 0`), so if I've understood correctly we should make `ierr`...

Is there more that needs to happen here, or can we close this issue?

What’s the time-scale from oxygen ignition to collapse versus the thermohaline mixing scale (e.g. r^2 / D_mix)? That would be useful to know for thinking about this... -Adam On May...

If I'm reading the einsum source correctly the iterator limitation should only affect individual contractions. The example I provided, for instance, can be done with just numpy dot or operations...

Oh I see: that's what it does when optimize=False. When einsum is called with optimize='greedy' or optimize='optimal' it instead computes a series of intermediate objects that to avoid doing just...

Ah in that case there's an inconsistency in the documentation. The [docs](https://docs.scipy.org/doc/numpy/reference/generated/numpy.einsum.html) for einsum say that optimize defaults to False in the function signature but that it defaults to True...

Okay perfect. Then it's just the issue with characters in c_einsum.