John Meacham wrote (snipped): > I should say that efficiency is the only thing I have been concerned > about in this conversation.
Well, I think my proposal can be made efficient if you try, using hash tables.
There is a reason why I think top-level <- is inefficient, because it imposes a single scope for variables on the program; variables really are global. Imagine a very highly parallel program running on lots of processors, then it will be important to make sure you do not have all the processors having to constantly refer to the same global variable. Really what you need is finer control over what copies of a variable each processor uses, which is one reason why I think the extra indirection introduced by my proposal is a good idea.
Consider for example Data.Unique. If you are generating lots of unique identifiers in a very highly parallel system it might well be much better for each processor to generate its own and then add a processor identifier than to have to have a central dispatcher which has to be consulted each time. _______________________________________________ Haskell mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell
