Parameter D in Eigenbasis Calculation from Pivoted Cholesky Output

[thogerig]

Fixes #164

The mass matrix parameter D is in our situation a scalar, which is set to 1.0. Since there is no occasion to be found where we would have to change this, I would also propose to remove the exposure of this parameter.

In a future case, in which the current calculation needs to be adapted with a different D, I would propose to define it as a separate function.

[marcelluethi]

@thogerig Thanks a lot for addressing this issue.

I agree that this parameter should not be an argument that the user can chose.
Yet, as it seems to be important in the calculation and seems to have a meaning, I would prefer not to drop it altogether.
How about introducing it as a value val D = 1 inside the function? We would then be reminded whenever we look at the method, that there is such a parameter D which can play a role. Ideally, we would also add a comment explaining its effect and why we keep it fixed.

[thogerig]

@marcelluethi Thanks for your input.

To me this seems like a good idea. I have updated the pull request with a reintroduction of D and a description with a pointer to further literature.

[marcelluethi]

I find this comment somewhat confusing. Why is the matrix a scalar in this case? Where in the paper can we find the explanation? At least a first quick look at the paper does not reveal a discussion regarding the parameter D.