in reply to Empirically solving complex problems
Does anybody else think like this?
And I have the same feelings about a lot of other stuff, including most of CS.
A problem described can be related to Knapsack or The Travalling Salesman Problem--correctly described as NP-hard or NP-complete--but it can still be solved in real time using Perl when the CS majors give up. I've done this many times.
Graph theory is, IMO, an abstraction too far. Graph terminology--nodes, edges, traversal etc.--in many cases, simply serves to obscure the details of the problems they describe, and in doing so, maks things which are straight forward to understand and implement, hard or impossible to do, because you can't see the trees cos the forest is in the way:)