> I'll see your Joel Spolsky and raise you Possibly not necessary. The underlying theme with Joel's point, if you look past the hyperbole, is that you don't w

> I'll see your Joel Spolsky and raise you Possibly not necessary. The underlying theme with Joel's point, if you look past the hyperbole, is that you don't want to take successful software and start over, re-implementing everything from the ground up to do the same thing more efficiently or with better architecture or aesthetics. "Plan to throw one away" on the other hand, is about going through the building process to better recognize the problems you actually need to solve rather than the ones you thought you needed to solve. Assuming you rewrote your engine because it wasn't solving the problems you really needed it to solve, the conflict between the two maxims isn't significant.