On Saturday 05 April 2008 17:10:57 Larry Wall wrote: > On Fri, Apr 04, 2008 at 09:41:26PM -0500, John M. Dlugosz wrote:
> > I suppose any object would do, it doesn't have to be "but undefined", or > > created using that Class{hash} syntax? > > Possibly. Haven't really thought through the ramifications, and it > feels a bit like the problem of confusing clone with new. It would > be pretty easy for idioms to arise that create a lot of useless > temporary objects. You can only sweep so much under the carpet of > "the optimizer could fix it"... > > In any case, the programmer really needs to keep straight when > an object is being used for its value vs when it's being used > for its type. Maybe that's not an issue here. Parrot-wise, right now that's somewhat more expensive than it should be. The carpet's not as large as I wish it were. (However it's possible not to require instantiating a new object or class for type checks, let's do that!) > We have another possible level of granularity there too, insofar as we > could trust a single multi or a proto that represents all multis in > its scope. Maybe a proto exported from within a class automatically > conveys trust to all normal multis (of the same name) declared the > import scope. Something about that I like. > > Also, since classes can be re-opened, anyone can grab trust from any > > class anyway, just by declaring a new method inside that class. > Well, sure. But whenever the programmer does that, the Perl 6 compiler > will automatically send email to the programmer's supervisor. :) If you mean "The World's Most Maintainable Programming Language", that was an April Fool's joke. -- c