tag:blogger.com,1999:blog-21500237.post116181872767090719..comments2023-04-29T04:08:32.247-04:00Comments on Praise, Curse, and Recurse: Exploring HaskellPaul R. Pottshttp://www.blogger.com/profile/04401509483200614806noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-21500237.post-33392833933377117152006-12-11T12:27:00.000-05:002006-12-11T12:27:00.000-05:00Hi Miles,
I was just intending to point out a cas...Hi Miles,<br /><br />I was just intending to point out a case where a dynamic language performed a lot better when types are provided so that generic dispatch does not have to occur and the compiler can generate more straightforward code. The code in question was a toy implementation of the old standby problem, the Knight's Tour. I could send it to you -- I was building it with d2c. There were some specific issues with d2c where it would not generate efficient code unless you introduced typed intermediate variables. They may have been fixed by now, or may perform better in OpenDylan (using the old Harlequin code base).Paul R. Pottshttps://www.blogger.com/profile/04401509483200614806noreply@blogger.comtag:blogger.com,1999:blog-21500237.post-21234506184564817922006-12-11T07:51:00.000-05:002006-12-11T07:51:00.000-05:00[Er, sorry, clicked "Publish" too soon on the last...[Er, sorry, clicked "Publish" too soon on the last attempt]<br /><br /><i>[M]y experience with Dylan leads me to believe that typing is actually valuable; in Dylan, for example, it can make a difference between code that uses full runtime dispatch and takes twenty minutes to run, or a version that takes a few seconds.</i><br /><br />Wow. I'm coming to Haskell from Perl, and I'm finding the type system a rich source of pain and frustration, so this really caught my eye. Could you post the relevant code?Unknownhttps://www.blogger.com/profile/07136909835648629963noreply@blogger.com