Mitchell N Charity <[EMAIL PROTECTED]> wrote: > ... Our focus will > shift to making parrot actually work. To giving it > working exceptions, io, embeddablility, and so on.
> Given Leo's numbers, it seemed a possibility worth suggesting. > I can certainly imagine other policies. Like focusing on performance > and using it to drive interest and obtain additional people to get > things working. But we should be clear on what we are doing. Well, making it work and making it work correctly is of course one of the major goals, or the primary goal. But sometimes I just don't know yet, how it should really work, like the whole experimental code from events over exceptions to threads. Or the current discussion WRT Continuation. No one did answer til now, how they should really work. OTOH: We have CPS subroutines since almost a year. Performance sucked until today (s. fib.* benchmarks, which are plain function calls). If Parrot's function (or method) calls are 2-3 times slower then any other interpreter, no one will have much interest to use Parrot. Not even me. Some optimizations might be premature but I think that some are just necessary to create interest that makes the project keep running. And one final note: It's of course a lot of fun to speed things up considerably. > Mitchell leo
