Guillaume Fraux

> This is a reminder to add support for storing these isotopes in these formats. We could also add a warning so that we don't silently drop non-integer masses. At...

Yes, units conversion will be an issue for this =/

> Another note: Avogadro uses Eigen for storing Vectors. Because `sizeof(Vector3D) == sizeof(Vector3)` and `std::is_pod && std::is_pod`, it should be enough to cast a `Vector3D*` to a `Vector3*` for the...

Yeah, cells lists (that the blog post mention rather than neighbours lists) might be a good idea here. Thanks for the head up!

We don't support POSCAR yet, but it would make for a great addition! Nobody is working on it at the moment AFAIK. It seems to be a free-form line oriented...

That's something we still have to document. The files live in `https://github.com/chemfiles/tests-data`, so you can make a PR there. Then, they are downloaded in https://github.com/chemfiles/chemfiles/blob/76e77998739420fef2a2080212d148da9ac8d28b/tests/CMakeLists.txt#L1-L11 So you can update the...

Since we are currently using the VM molfile reader for DCD files (until #370 is implemented), we should update the documentation! This part of the doc is generated directly from...

> I feel like this would introduce rather unpredictable behavior. After all, this is a completely valid orthorhombic cell. Software like Lammps can have all sorts of different units which...

Let's keep this open to track similar changes for PDB & friends

I think that topology reading is not registered with MDA, since the chemfiles adapter is implemented as a coordinate reader/writer. Although the conversion from chemfiles to MDA topologies is already...