Josh Berkus <josh@agliodbs.com> writes:
> Tom, just to be perfectly clear about why I see Procedures as a way of
> resolving parameter ambiguity, my idea is that:
> FUNCTIONS will support overloading but will not support named parameter
> calling;
> PROCEDURES will support named parameter calling but not support overloading.
Understood, but this seems like a bad design to me, because it's
non-orthogonal.
> Particularly, I'm concerned about adding any more code to the
> evaluation of a function call, out of fear that it will have a
> significant performance impact due to increased time to evaluate
> built-in functions.
I think that named params would have no significant extra cost *when not
used*, so I'm not sure the above concern is a big deal. (I do worry
about the cost implications of defaultable parameters, however, as that
seems likely to expand the search space for a matching function quite a
bit.)
regards, tom lane