To solve a problem, try to build as general problems -- or parameters -- into the cognitive model as possible. This gives it more flexibility. Many people think that this lowers the flexibility. Wrong.
Every problem defines some axis of acceptable solutions. In other wrods every problem is a parameter in a problem space. With one axis in the space, the range of parameter settings, where each denotes a logically possible solution, is linearly representable. Note that the set of logically posible solutions is a uperset of empirically possible solution. To arrive at the later a second parameter must be added into the space, namely that denotable by the problem: let the solution be possible to carry out in the world. [Personally I leave this one out until last. Personal choice. makes me feel less like an engineer when I think because I get to wander all the fairy tale landscapes at no extra cost before selecting a real solution!]
[ ]