Although it’s really a fraction of the complexity of e.g. elliptic curves, most of the implementations I’ve read look decidedly like magic, mysteriously multiplying values by enchanted constants, and shuffling bits like The Sorcerer’s Apprentice in Fantasia. Even the paper does not explain why and how its design decisions lead to faster code!

Still, after reverse-engineering what the implementations were doing, I grew convinced that cryptography code could be perfectly understandable if only we commented it.

source: L