4 ms·

In fact you might also want the un-optimized code around and run the optimized version "against" the un-optimized version (in tests) to see if they actually do

In fact you might also want the un-optimized code around and run the optimized version "against" the un-optimized version (in tests) to see if they actually do the same thing.

joshka 11y ago
Extract an interface (IFoo). Implement Foo : IFoo and FasterFoo : IFoo. Write tests to confirm all implementations of IFoo act equivalently. Write performance characteristic tests around FasterFoo.
- lomnakkus 11y ago
  Exactly.