At 08:07 AM 8/26/2001 +1000, Damian Conway wrote:
>General purpose delegation has to be so general-purpose that it's hard
>to see that *any* (virtual) hardware support is feasible.
True, true. My OO experience is reasonably limited (in my day, we only had
integers. And they only had 8 bits. And we *liked* it that way... :) so I
wasn't sure if there was something both subtle and glaringly obvious that I
was missing.
Also, I'm OK with a target less than 100%. Heck, if we could say "Dispatch
Method A is used by 60% of classes that do delegation" I'd be thrilled to
weld support for it into the interpreter. I'm never going to get it all,
but if I can get a reasonably large chunk with some ease, I'm happy.
>As you pointed out, the scary delegation capabilities of Class::Delegation
>would seem to defy low-level support, because they require too many
>high-level decisions to be made before a method can be delegated.
Yup. But not everyone goes to quite that extreme, and for all I know there
could be One True Delegation Way that some OO language uses. (Or, more
likely, there are three or four languages with One True Way, all different
and incompatible)
Sounds like punting is the best thing at the moment. That's cool, I like to
punt.
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
