I can imagine CPU and memory usage are still side effects. These side effects seem often overlooked. perhaps saying “in functional programming, side effects are minimised” is closer to the truth
-
This Tweet is unavailable.
-
This Tweet is unavailable.
-
That is an interesting view. How does that interpretation fit in with referential transparency? Surely the number of times you call a function affects the CPU usage, but which is not encoded in the return/effect
It means that you can replace expressions with a value, without affecting program outcome. Like in this IO program: https://github.com/data61/fp-course/blob/master/README.markdown#demonstrate-io-maintains-referential-transparency …
