Functional Language Summary
I’ve been hearing a lot about functional programming lately (and the circle of life continues); I found Functional Programming For The Rest of Us to be a nice summary. Here’s what I got from it:
Atoms of FP:
- All “variables” are immutable, often called “symbols”
- Program state is kept in functions (specifically, arguments passed on the stack to functions), not variables
- Functions are first class citizens, i.e. functions can be passed as arguments
- “currying” is a convenience syntax for adapting a function to an alternate function signature
- “closures” are functions that are allowed mutable state and access to state outside their lexical scope to bridge functional and non-functional languages
Implications:
- Functions cannot cause side effects (“variables” are immutable)
- FP is great for unit testing (only have to test outputs against inputs — don’t have to test side effects)
- FP is great for debugging (no need to worry about external state affecting function results — results are only based on the input)
- No need for multi-threaded locks, as state is immutable
- This makes functional programs automatically parallelizeable
- Can hot swap new function definitions w/o effecting existing instances
- Don’t need to evaluate a function ’til the results are needed
FP sounds great! Why do we mess around w/ anything else?!?