the difference between foo _ and foo(_) (or foo(_, _) and so on) has always been a puzzler factory.

if the choice is between foo _ and foo(_), then losing the first option isn't bad, but if the choice is between foo _ and foo(_, _, _, _, _, _) then it starts feeling different.

but in the wild, perhaps most uses of foo _ would also compile simply as foo because a function type is already expected?

personally, I think I tend to write foo _ even when just foo would compile, just to explicitly signal to the reader that foo is definitely not being called.