News for hackers. Stuff that matters.
Re:Recursive fibonacci benchmark using top languages on GitHub by gnulinux on Friday September 28, 2018 @08:35AM (#18092801)
Why do we need language level memorization again? In python memoizing a function is just one line `@lru_cache()` I can't see why this needs an improvement by adding yet another thing to the core language.
Re:Recursive fibonacci benchmark using top languages on GitHub by thaunatos on Friday September 28, 2018 @08:55AM (#18092907)
Just adding a @lru_cache() to a python isn't guaranteed to be correct; for example, if you're reading from an API the response could change between calls.

Skip tracks side effects, and will either (a) memoize automatically a pure function or (b) recognize that a function is impure and avoid the memoization.

Re:Recursive fibonacci benchmark using top languages on GitHub by ulucs on Friday September 28, 2018 @09:05AM (#18092961)
Some time ago, facebook introduced a javascript compiler which pre-compiled intermediate values. I think the point was to extend the spirit to the backend apps. If memoization is first class, implementing pre-compilation should be a blast