Conversation

Notices

1. Embed this notice
   alcinnz (alcinnz@floss.social)'s status on Wednesday, 29-Jan-2025 04:54:11 JST alcinnz

   Computer science is based on foundational assumptions about which operations are fast vs which are slow. So when I design hypothetical hardware, that *can* change these underlying assumptions. But we needn't go that far!

   Merely promising that you want modify any data changes performance considerations! Which is the route Haskell takes.

   I quite like it, though I would advise picking compute-intensive tasks for Haskell! That's what its suited to!

   • Embed this notice
     alcinnz (alcinnz@floss.social)'s status on Wednesday, 29-Jan-2025 11:37:46 JST alcinnz

     in reply to

     An example: Wouldn't all the short-lived copies you're making to avoid modifying data be inefficient? Not with GHC's copying/arena-GC...

     Fundamentally in functional programming we prefer trees over arrays, since that minimizes how much memory we're duplicating.

     2/1!