[mindcrime]

Certainly, never start with the solution and work back to find the problem.

I'm going to play contrarian here and say I slightly disagree. OK, I mean, yes, as a general rule you don't want to start with a solution and work backwards. That I agree with. I just don't consider it an absolute. And I think that as an exercise done occasionally it can be worthwhile to look at some "solution" (that is, some bit of potentially usable technology) and work backwards and ask "What could I do with this?" I find that that can lead to some interesting ideas in its own right.

The danger, of course, is the possibility of getting overly enamored with the technology for its own sake, and wind up building something nobody needs or wants. But things like Steve Blank's "Customer Development" process can be used to help avoid situations like that, or at least before you invest very much in such an un-needed "solution".