Rachel Kurchin

`add_atom`/`remove_atom` functions, example with mutability?

2

Nothing should stop someone from having a subtype of `AbstractSystem` that's mutable, and it could be useful for setting up geometries for simulations, especially if they're rather large and the...

Library of examples

1

Now that AtomsBase is starting to get wider usage, I think a library of examples in the docs would be a good thing to have. Starting this issue to brainstorm...

What type parameters should `AbstractSystem` have?

2

Currently we have one `D` for number of dimensions and one `S` for the species type. There could be more (e.g. adding `Unitful` types to specify units for length, velocity,...

should there be a `distance` method?

2

I'm imagining a primary syntax something like `distance(sys::AbstractSystem, atom_ind_1, atom_ind_2)` that would return the distance between two particles in the structure, and maybe also single-index and no-index options that would...

make sure AtomsBase plays nice with Crystalline.jl/Bravais.jl

4

https://github.com/thchr/Crystalline.jl because this would just be really cool for geometry building etc. ...depending on how in-depth this goes, it could easily end up as a separate package...this issue is mainly...

Should we add `mass` and `momentum` functions?

1

@cortner suggested this...TBD if it belongs at the top level of the interface...

Improve convenience constructors for included implementations

1

They're a bit too rigid right now and should ideally be flexible to anything (different types of vectors, tuples, etc.) that might be thrown at them to specify systems.

best names for `FastSystem` and `FlexibleSystem`?

1

They were originally `SoASystem` and `AoSSystem` and were changed to the present names. Thoughts?

Support Cartesian indexing

1

In lattice-based systems such as those @Gregstrq works on, it will eventually be desirable to support Cartesian and not just linear indexing into `AbstractSystem` objects.

some kind of `wrap` function for PBC's?

12

Would it make sense to have this be part of the interface? If so, we would need to be able to dispatch on boundary condition types, which might require rejiggering...