After yawning during the first 10 installments of the series, i’ve just watched a really beautiful talk (with accompanying slides and literate code) by guest star Graham Hutton in the C9 Lectures webcast [0]. Graham devotes it to show how to solve the countdown puzzle for numbers (find ways of combining n integers to obtain a target one, using basic arithmetic operations) in Haskell. The lecture is a very nice show off of list comprehensions and how they allow the elegant expression of solutions to certain kind of problems (basically, tree searches in this case), without precluding incremental optimizations (pruning of the search space in the problem at hand). In addition, you’ll see a very simple, but enlightening, example of program fusion as a technique to boost performance. The code is beautiful and straightforward and, most notably, retains those qualities even after optimization. Hutton is a good speaker and teacher, and the lecture is accessible to anyone with a bare-bones knowledge of Haskell. And the fun doesn’t end with the video: there’s also an accompanying paper, most aptly included in the functional pearls series, providing all the details, code and proofs. A discussion is also present in Graham’s book, which, to judge from the quality of the paper, must be an excellent way to introduce yourself to Haskell and functional programming.
—-
[0] I keep reading praise for these lectures, and everybody seems to adore E. Meijer. Yet i must confess that i find these lectures incredibly bad: to my taste, the lecturer is dull, often misses the point, keeps drawing dubious parallelisms and, every now and then, just makes plain (and, i’m tempted to say, glaring) conceptual errors. At the risk of losing all my credibility, if any, i wholeheartedly recommend you not to spend any time in the first ten lectures of this series. I’d take any talk by SPJ any day: that is passion and insight. Or by Hutton, for that matter.
