Guillaume Fraux

Yeah, HDF5 could be nice too here! There is a standard format for using HDF5 storage for MD data called [h5md](http://nongnu.org/h5md/), we could implement that too. > The latest version...

> What do you think is the best approach to do that? For NetCDF, chemfiles already supports the trajectory convention. Adding support for the Restart convention should not be too...

> Where/how to keep track of degrees of freedom? > Where would we implement information about constraints? Could they be part of Interactions? We could store both of them in...

This could be a solution. Or we could store them aside with the Bond, or in a RigidBond struct: ```rust struct Molecule { bonds: Vec, angles: Vec, rigid_bonds: Vec, rigid_angles:...

Yep, that is true! We will need support for rigid molecules in CBMC too then. Do you know if this is possible?

Creating bonds will be hard: how could we know which atoms are in the same molecules if we don't have the bonds between them? Checking bonds might be hard too:...

> Ok. I thought we'd supply angles and dihedrals here as well. We don't need to, as we can determine them from the list of bonds only. And yes, there...

Hey there ! I'd really like to do something like this in my code. What is the advancement of this point ? There was a PR (#12) some time ago,...

Well, it would be a good thing to have it in this package too : you can fall back to finite differences if there is no way to differentiate analytically...

> One thing that would be nice at some point is to try to unify interfaces between the packages so that it's easier to transition from Calculus to ReverseDiffSparse. Sure...