News for hackers. Stuff that matters.
Re:Timeline for Logic, λ-Calculus, and Programming Language ... by pwm on Tuesday September 03, 2019 @08:58AM (#20866019)
(On my phone walking so very quick and dirty post) purely functional DSs in general will cost you an extra O(log n) cause of their underlying tree representation for immutability (immutability achieved by copying paths in the tree to retain old data). Also some DSs are very difficult to express in a purely functional way.
Re:Timeline for Logic, λ-Calculus, and Programming Language ... by adjkant on Tuesday September 03, 2019 @12:31PM (#20868257)
While true on the first part, these days many cases don't have insane primary efficiency needs that would be affected by that log(n). As always though, you choose the tools based on the job at hand, not the tools. In my experience, functional philosophy often is about the gained mental and consistency advantages achieved by the features offered.
Re:Timeline for Logic, λ-Calculus, and Programming Language ... by pwm on Tuesday September 03, 2019 @12:50PM (#20868475)
Don’t get me wrong, you are preaching to the choir (I write Haskell at work), just thought it’s worth mentioning this to the OP. In practice sophisticated compilers like GHC produce insanely fast binaries and is getting better and better.
Re:Timeline for Logic, λ-Calculus, and Programming Language ... by continuational on Tuesday September 03, 2019 @05:13PM (#20871237)
Array indexing is not O(1) on any modern machine due to the cache hierachy.