Feedback on the manual

[antoinewdg]

As discussed in #111 I'll list here some points I had (have) difficulty with when reading the manual, from the point of view of someone who is not an expert in molecular simulation. I'll update it along the way.

• Input files > System : the .pdb files are described as being useful to specify the bonds, while they can also be used (and are in the integration tests) to specify the molecule positions and the cell. I'm still not clear what happens if we pass both a .xyz and a .pdb and in here

• Input files > Interactions > Restrictions why is this whole part not in Input files > Pair interactions ?

• Input files > Pair interactions references to literature about the different types of cutoff would a plus

  • Tail correction: "if the pair distribution function $g(r)$ is 1 after the cutoff" what is g ? what is the rho that appears later ?
  • Potentials computation: mentioning the cutoff in this part creates confusion. Since the table computation is the only alternative computation for now, I would just remove this section, Table interpolation is self-sufficient
  • Table interpolation: can this be used on Coulomb potential ? This should be precised. Literature would also be nice here.

This is of course entirely subjective so feel free to disagree :).

[g-bauer]

Thanks a lot for the feedback. Some points you mention are due to "adding things on the fly" which leads to repetitions (e.g. the restrictions parts appear in two subsections).

"if the pair distribution function $g(r)$ is 1 after the cutoff" what is g ? what is the rho that appears later ?

These are points we should clarify, I agree. But I think instead of adding these informations directly, we should keep the input section rather light and focus on explaining the input quantities the user has to make. Additional information about the theory should be added in the "conceptual guides" section. For the user, I think we should make sure that the input part is polished.

[Luthaf]

"if the pair distribution function $g(r)$ is 1 after the cutoff" what is g ? what is the rho that appears later ?

FYI, g(r) is the radial distribution function and rho the density.

---

I think we should just define the quantities we are using here, and redirect people to other part of the manual/other tutorials.

Concerning the restriction, they are here because we can use restrictions with both pair interactions and coulombic interactions. We could move them in the pair interaction section and then link from the coulomb part to there.

Table interpolation: can this be used on Coulomb potential ? This should be precised.

Theoretically yes, at least for the real part of Ewald and for the Wolf method. In practise, it is not implemented in the code. I've never seen it done though. If one want a tabulated method to use with electrostatic, we already have (S)PME and PPPM.