Seth Axen

`gradient(ba, f, args...)` just looks cleaner than `gradient(f, ba, args...)`, since it puts the functions with its args. Also, in general these functions should dispatch on the backend, not `f`,...

@phipsgabler what do you see as the pros and cons of the curried form vs the one in https://github.com/JuliaDiff/AbstractDifferentiation.jl/issues/33#issuecomment-1018923309?

Not certain if this package wants to support complex AD (with Jacobians as the primitive, complex support is a little awkward), but to me it seems like, if the AD...

That's right, and even with both complex inputs and outputs, it can be supported by interleaving the real and imaginary parts into one real vector, but having the eltype be...

Several of these are added because the version of the docstring in Base julia explicitly mentions function behavior for complex numbers, which does not describe behavior for quaternions. So not...

On Julia v1.9 this could be done by setting up a weakdep of Symbolics on Quaternions or the other way around. I'm not certain which makes more sense.

Since https://github.com/JuliaDiff/ChainRules.jl/pull/540 has stalled, we should probably just defensively use `ChainRulesCore.@opt_out` for functions that we implement and that we know depend on the quaternion product so that ChainRules-compatible ADs don't...

I think this is one of the next things we should do. It involves adding ChainRulesCore as a dependency, but it's already an indirect dependency of much of the ecosystem,...

Here's a minimal example. We'd need to think more carefully about indexing for a real implementation, which might mean we restrict inputs to `StridedMatrix` types and triangular matrices: ```julia julia>...

Thanks for the issue! I don't think the solution is to restrict the more accurate branch to `AbstractFloat`, because then we would hit the less accurate branch whenever using wrappers...