Bob Carpenter

To me, `_for_each` sounds like it's saying the behavior is a loop. While that's technically true and therefore the name makes sense, I find it too long and not quite...

I've even tested that Schur decomposition works as part of the original complex PR. Here's a version that worked using the two functions approach: https://github.com/stan-dev/math/commit/6490ee33e650d1aefe4e44126a39cb1e2975e1e7 But then for some reason...

I know it works because I used the Schur decomposition to test complex numbers. Here's an implementation that used to work before the math library started drifting faster than I...

From [Wikipedia on Schur decomposition](https://en.wikipedia.org/wiki/Schur_decomposition): > Although every square matrix has a Schur decomposition, in general this decomposition is not unique. Luckily, we don't need to test prim against existing...

Exactly!

Ideally, there would also be a `_lcdf` and `_lccdf` function to complement the `_lpdf` and `_rng`. We have to decide whether to restrict input to (-pi, pi) or to treat...

Nice. See inline. > On Oct 2, 2016, at 6:46 AM, benlambertStan [email protected] wrote: > > Hi Bob, > > Ok, I have now cloned the Stan/math repository and have...

Thanks, @andrjohns. Is there code that produces compiler errors now? If so, it'd help to say what they are. That may be the `std::max` call, but I couldn't tell. The...

Thanks, @spinkney. Right now, we're just autodiffing through the Eigen implementations of complex functions, but this is what we'll need to add analytic derivatives.

I think there's a general problem with forward-mode autodiff of complex numbers interacting with Eigen. They're failing in my FFT PR as well when all of the reverse mode things...