When it comes to fine distinctions, the game is a sad joke. It's an arms race toward writing nearly machine-language code as best your language supports it. T

When it comes to fine distinctions, the game is a sad joke. It's an arms race toward writing nearly machine-language code as best your language supports it. There are Haskell programs that are doing a significant amount of work as essentially "peek" and "poke" inside a memory buffer. There are Java programs that use variables q,r,s,t,u,v,w,x,y,z rather than two arrays of size 5. I guess if you have a legitimate need for getting the most performance on some fixed task (and abstraction, readability, and extensibility can all be sacrificed), without using a call to C, then both are fair.