
So I spent most of Thanksgiving weekend relaxing. And by that I mean "studying algorithms" because I really am a geek. And I feel much more relaxed now.
Anyway, I studied max flow and bipartite matching. I realized that bipartite matching is the first algorithm complicated enough that I should try some "implementation practice"  by which I mean simply repeatedly typing in the algorithm from scratch to see how long it takes to get it working. After the first one (which was a bit of time), the implementation times I had for FordFulkerson were (in a row): 19 minutes, 11 minutes, 7 minutes. I tried also with a standard BFS and got 11 minutes, 5 minutes, 3 minutes.
How curious...
I'm not sure how useful this exercise is, but it appears to have some value. I haven't done part two of the experiment, which is to repeat the experiment in a couple days and see if I have times similar to or better than the first.
Anyway, I was struck by the shape of the curve. Thoughts? 