Non-dimensionalization of parameters in code not documented

[brunobeltran]

Luis's discovery that DT is ignored in #15 is part of a much larger problem of undocumented scaling of the parameters used in the simulation.

We all love non-dimensionlization but we all agree that it should be documented, and not just in some paper that you have to look up and carefully parse.

I've started working on the Gaussian chain part, which will be the easiest.

Closing #15 as this supercedes it.