I liked a lot of what (friend of ODF) Stuart Eccles had to say here about the process of prototyping at MadeByMany. He talks about how prototyping is an on-going process that might involve dozens of prototypes, each with a speciifc learning goal (typically the most important thing that needs to be identified at that stage of the project) that could relate to anything from user experience, idea evaluation, technical feasibility, business model development or cost forecasting. The common error is to try and cram too many learning goals into a single prototype, which dilutes the learning and also the ability to select the optimum form of prototype to achieve that learning at the least cost.
There was also a useful delineation in this HBR piece between rapid prototyping (quantity over quality, dozens of sketches, wireframes, scenarios, or models, quickly created to get a feel for something), pilots and technical prototypes (which might be more of a fully formed experience, or closer to the final version), and 'live prototyping' which involves releasing still-rough concepts into as natural a context as possible and designed to replace techniques like focus groups (which I dislike anyway), surveys and so on.
One other (not insignificant) point about prototyping – in my experience even the most basic working prototype is a far more compelling way to secure senior buy-in for a project than a dry powerpoint presentation.
P.S. MadeByMany are hosting Lean Day London – sadly I can't attend but it looks really good. And my good friends at AdaptiveLab run two day Market Sprints which is a concentrated burst of prototype/business case building which I'd also recommend