Condidering commits with more than one author (co-authors)

[lidix91]

When a commit have co-authors, LD only consider one of them, the commits should be shared among the authors, at least the authors that are part of the team.

Possible cases:

• more than one author, all of them part of the team --> shared authorship
• more than one author, not all of them part of the team. Example: three co-authors, only two part of the team
  • option A: shared authorship among authors part of the team (no anonimous commits): 50-50%
  • option B: shared authorship among al the authors (this option create anonious commits): 33-33-33%

LD-Connect needs to identify the number of authors, LD-Eval needs to consider the number of authors in the formula

Note: Consider when commits has "Claude" (or any AI development tool) as co-author. Should be consider IA tools as co-authors? How do we treat a commit user+Claude, the user has de 50% of authorship or 100% because Claude is a tool used by the user. Is it possible to distinguish when a co-author is a AI tool?

[XavierFranch]

This is definitively a must-have feature. I was not aware that you can have co-authors in GitHub commits. With this feature, we can get rid of the recommendation of switching among accounts when they work in pairs, which is something I hate to ask for (artificial).

About external committers, I think that if all are humans, this is something will not happen, right? All co-authors should be from the team. If any committer-id is not from the team, it should be treated as an error, IMO.

Special case of GenAI. I would not consider them, at least in a first version. I would instruct the students not to declare an AI tool as co-author. But it is not a strong opinion; I could be convinced otherwise.