On Friday, May 15, 2015 at 4:39:33 PM UTC, Kuba Roth wrote: > > Awesome answer!
You answer is awesome too. > I too agree bringing up these numbers have little sense, at least form the > perspective of solving "real world" problems. Julia has much more to offer > then raw speed. > So if i were worried about the performance I probably would focus on > advocating better support for shared memory concurency since in this area > Julia can do better. Fortran being 3.75 times faster than Julia in a restricted case (doubly recursive - fib) seems worrying. Neither language are mostly for recursion(?). You can always rewrite recursion as a loop (and vise versa - but may not want to..). Recursion is elegant for many things. I'm not sure I would want to rewrite double recursion even if I could (can you always?). I agree it has more than raw speed, clean small source code and development times in relation to that (smaller). That is a huge win. Mostly all of my posts look at how Julia could be even better. I would hate for there being a good complementary language (I want Julia to excel for everything - I'm also greedy - just more than the original developers/goals), but especially for speed reasons only. -- Palli.