This seems like a perfect scenario to try out Bret Victor's "explorable explanations"; i.e. not just documenting assumptions & code inline with your reasoning,

This seems like a perfect scenario to try out Bret Victor's "explorable explanations"; i.e. not just documenting assumptions & code inline with your reasoning, but making the document itself explorable and reactive: http://worrydream.com/#!/ExplorableExplanations http://worrydream.com/#!/ExplorableExplanations