> What are your thoughts?

Sigh. I don't object to extending to `k < 0` and `k > n`, but once we've made that extension it's impossible to undo if we decide we'd rather have had the error checking. I'd really like to see some convincing use-cases. Quotes from Concrete Mathematics (fine book though it is) don't amount to use-cases.

I'd say leave it as-is for 3.8, see what the reaction is, and maybe relax constraints in 3.9 if that seems appropriate. But if a majority of others really want to make the change now, that's okay with me.