Update citation policy for theory used

[scarlehoff]

The theories should include a bibtex file with the relevant citations so that if someone uses a theory you can check which software should you cite.

I'm adding this issue here because the citation policy is in the docs of nnpdf but perhaps we could have a pineko sanitize theory or something which will create the .tgz file and looking at the pineappl grids inside prepare the citations?

Or we just copy the entire citation list that we will have for 4.1 into the docs. I just realized that things like EKO should also be referred there.
We don't want to get sued!

[felixhekhorn]

first the obvious statements:

• all FK tables (beyond 4.0) use EKO and should cite EKO according to the citation policy given at the repo
• all FK tables (beyond 4.0) use PineAPPL and should cite PineAPPL according to the citation policy given at the repo

then only grids remain, for which I can see an easy option: the interfaces in pinefarm should provide a citation string, which pinefarm then adds to the grids (together with version numbers etc.) - so for me this should be a pinefarm issue.

(There is also a too complicated point of view: we could also let pinecards add dedicated citations, e.g. to specific codes in mg5 or similar - but this is too much I would say ... for DIS we are not citing all the literature that went into DIS; those references are just inside yadism)

[scarlehoff]

For reference, I meant this citation policy here: https://docs.nnpdf.science/get-started/cite.html which is still that of 4.0. People looking at that won't look into pinefarm, pineko, etc so I think at least the general policy should be as easily visible as possible.

[felixhekhorn]

but the list there is quasi the point I was worried about. If people ask us explicitly to be listed there, we can do that - but that we do manually case by case. I would not do this for everything, because it would be very hard to draw a line on what to cite and what not (beyond the program itself). Just to name a completely random example: should we always cite my thesis just because it contains some genuine new coefficient functions, which we use in all DIS calculation? or should we cite the paper on which my thesis is 100% (more or less) built? should we cite Numpy and friends (without we were completely lost)?

just to stress: I'm happy to cite whoever, but I wonder if we need an automatic tool for that

[scarlehoff]

What I'm suggesting is that we should change that list to nnpdf / eko / pineappl and then perhaps a script that can automatically read the citations from the theory fktables.

The last sentence about 4.1 was meant to be snarky since I remembered what happened back in the day with eko :P. Indeed, if people start asking to be put there, we should put the whole 4.1 list of citations, numpy included.