On 18 Jul 2007, at 8:52 pm, Bjorn Bringert wrote:
Well, the original poster wanted advice on how to improve his Haskell style, not algorithmic complexity. I think that the appropriate response to that is to show different ways to write the same program in idiomatic Haskell.
(a) I gave some of that; I wrote my solution before seeing anyone else's. (b) I find it hard to imagine a state of mind in which algorithmic complexity is seen as irrelevant to style. I am reminded of the bad old days when Quintus had customers who were infuriated because writing an exponential-time algorithm in a few lines of Prolog didn't mean it ran fast on large examples. Their code was short, so it HAD to be good code, which meant the slowness had to be our fault. Not so! (c) The key point in my posting was the reference to Gries' paper, in which he derives an imperative program in Dijkstra's notation USING A CALCULATIONAL STYLE, very like the bananas-lenses-and- barbed wire stuff popular in some parts of the functional community.
/Björn