tag:blogger.com,1999:blog-10039526.post111171774660752752..comments2007-04-16T14:27:43.424+10:00Comments on David Kemp's Blog: Futures and Eventual Values Part 1David Kemphttp://www.blogger.com/profile/12250289605528269244noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-10039526.post-1114372639300664062005-04-25T05:57:00.000+10:002005-04-25T05:57:00.000+10:00AFAIK, the distinction between "eventual values" a...AFAIK, the distinction between "eventual values" and futures is typing. With MultiLisp's futures, you can't distinguish between an int or a "future int". Everything works the same. "Eventual value", however, are a distinct type that you have to extract the result from using a getter. Therefore, Java 1.5's implementation of Future is really an eventual value. Implementing real MultiLisp-style futures would require that the runtime magically hide futures from the programmer:<BR/><BR/>int a = future (expression) ;<BR/>x = a // runtime blocks until a is computed<BR/><BR/>MultiLisp gets credit for both ideas, but really they solved the harder problem.Anonymoushttp://rover.cs.northwestern.edu/~surana/blognoreply@blogger.com