On Tue, Sep 27, 2022 at 9:32 PM Asaf Bartov <abartov@wikimedia.org> wrote:

I agree this makes sense.

Thus, if X really hates Y's implementation of a factorial function, X can write their own (or fork!), and both can live side by side, and re-users will get to choose which implementation they prefer. That alone can significantly pre-empt a lot of potential conflict.

Good point, this could be a low-stakes shift. Anyone who wants a new version of a function should have a persistent instance of that fork to point to, with a name + version history that distinguish it from others in the same family tree.