Has anyone written some sort of comprehensive analysis of how protocols were designed to be used, how people ended up using them instead, and what lessons can be drawn from it?

Across multiple different case studies, ideally, not just one protocol.