Eric Agol

https://github.com/ericagol/NbodyGradient.jl/blob/685014cfa11f9876c498d1e6e17b1be050ba4932/src/integrator/ahl21/ahl21.jl#L1238-L1240 Can write `-\beta(G_2\eta+G_1r_0)^2` to simplify three terms.

https://github.com/ericagol/NbodyGradient.jl/blob/685014cfa11f9876c498d1e6e17b1be050ba4932/src/integrator/ahl21/ahl21.jl#L1258 We can factor out `k` and include `c_{15}/k` in the brackets.

https://github.com/ericagol/NbodyGradient.jl/blob/685014cfa11f9876c498d1e6e17b1be050ba4932/src/integrator/ahl21/ahl21.jl#L1309 This line can be simplified to `-k/r^3*(r_0G_1+\eta G_2)`

Efficient use of arrays

7

Could it help to use StaticArrays? Might it help to use a more specific type than Real? For the Jacobian propagation, should we use views? https://stackoverflow.com/questions/38777469/function-for-reshape-view/38778099

Need to investigate what circumstances are leading to this.

Need to figure out how to handle derivatives in e=0 case.

TTV computation comparisons

13

- [x] 1). Compare with TTVFast. - [x] 2). Compare speed of integration with C code. - [x] 3). Compare precision with Rebound (if possible). - [ ] 4). Check...

Implement photodynamics

1

- [ ] Allow integrator to compute positions and velocities (and derivatives) at pre-specified times. - [x] Code up a differentiable light curve solver (such as Pàl's). - [ ]...

Refactor code to allow Autodiff

1

- [x] Redefine Float64's as Reals, and copy types for type stability. - [ ] Implement Julia Autodiff and compare with analytic differentiation.

- [ ] In the conclusions where we discuss missing physics, we should cite https://arxiv.org/abs/2103.12767 and POSIDINUS codes.