I'm with you on the domain topic; reading the extreme example I'd add repeatability of the performance (on a different problem of the same domain) as an equally important factor to distinguish the developer from the inventor.