Well, I think we established that "===" wouldn't work.

Making multi-dispatch work with class hierarchies could turn out to be very
expensive if done naively.  Runtime would potentially have to traverse the
inheritance list of each argument specified, until a matching method was
found supporting arguments of the specified classes.
Maybe a cache could be used that also hashes on the supplied argument
classes list.  So generally, the hit is only taken first time round.

--
Justin Johnson

"William Djaja Tjokroaminata" <billtj / z.glue.umd.edu> wrote in message
news:amvagt$b69$1 / grapevine.wam.umd.edu...
> Well, you know, when the passed objects do not match the expected classes
> according to the Object#instance_of? criterion in any of the
> functions.  When you start to use Object#is_kind_of?, then you need to
> decide which function is the "closest".
>
> Bulat suggested using "===" so any expression will do.  But then do we
> just go from the first declared function till the end?  Then function
> declaration place has a new meaning, which is a new semantic for Ruby.
>
> Regards,
>
> Bill
>
============================================================================
> Justin Johnson <justinj / mobiusent.com> wrote:
>
> > What sort of ambiguities?