Leandro Martínez

Yeah, I could have simplfied that. The idea there is that each particle has a radius, and the interaction potential is zero if the distance between the particles is greater...

> The downside here is that deleting and adding agents during a simulation becomes much more complicated. Yes, certaintly there is some gained flexibility with the current `Agents` approach, although...

Chimming in: `StaticArrays` also support `FieldVector`, which is a very practical way to represent positions. As far as I can see, mostly of what is changed here are signatures, so...

I don't have very strong feelings about that. FieldVectors are convenient because you access each position by name (i.e. `pos.x`, `pos.y`). Otherwise is to allow more ecosystem integration. For example,...

Without complicating too much, the only thing that I would consider more in depth is to support `FieldVector`. Then those suggestions are useful, and the signature could be an alias...

Yes, in this example I can do that (in general I'm not sure, there are some customized problems that may require the current interface). I will do that here and...

> Why are `forces` in the `Properties` struct, and positions in the `CellListMapData` struct? What I tried is to separate what would be necessary for the use of CellListMap from...

> Wouldn't it make sense to wrap all these stuff into a struct and provide functions that operate on this struct, and take in positions and forces vectors and update...

@Datseris: I have significantly changed the example, after creating a new interface for the use of CellListMap which will simplify its, particularly in these examples. The branch containing the new...

> @lmiq please ping me when you think I should give another review and merge Sure, I'll tell you when I release the corresponding version of CellListMap.