Hi Nicolas, If I understand what existed and what is proposed, then I vote for the category Groupoids() and no arguments.
A standard definition of a groupoid in category theory is a category in which every morphism is an isomorphism. Thus it is possible that this was intended as a constructor for any such category (whose underlying objects are in some given set), but a multigraph would probably be a more precise input (and efficient representation). But there are many categories which might turn out to be groupoids, and it is unlikely that one constructor would suffice for their study. One still needs functors between categories to represent their morphisms (in the category Groupoids() of all groupoids). Those groupoids which do have a concrete representation (e.g. as a multigraph) and morphisms between them provide sufficient motivation to create the category Groupoids. The fact that each of its objects can be viewed as a category is an approach that can be explored later (if someone wants to build an effective framework for functors, natural transformations, etc.). The original code was set up to give a framework for category theory to organize mathematical constructs. The new category framework should be modular and flexible enough to delete unnecessary categories, or add a new one and experiment with its morphisms of functors from certain core categories. At present a constructor of a specific category, which happens to be a groupoid, is more of a motivating example for what one should be able to do rather than a core part of the categories framework. Cheers, David --~--~---------~--~----~------------~-------~--~----~ To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
