I think that BrowserUK's point is that a problem can be related to a known hard problem, but be very doable. Or there is a variation on it which is what you really need that is doable. The CS major looks at the problem and knows the theory about why it is hard. But if you just try it, often you'll get an answer.
in reply to Re^2: Empirically solving complex problems
in thread Empirically solving complex problems
Sometimes you won't. Often you'll get a fast answer to the wrong problem. But success comes often enough that it is often worth not giving up too easily.