Tuesday, June 17, 2008

Answers to a prayer and some comments

Having been online for a few days in a row seems to good to be true; but then my father always told me, "be careful what you pray for." I have this odd feeling of being naked and exposed when I'm using a browser. Being inside a simple http session, without any hint of encryption disturbs me. I'm afraid I 've lost my innocence as a web traveler. I have to keep a streaming view of traffic close at hand to venture to far out into the frontier. I had a comment come in sometime ago that I only now was able to publish. The reader makes the observation that the flatten algorithms would not work with dotted pairs. This is true, and further they will not eliminate embedded nils, that structure will remain. I haven't looked at my original post yet, however, I read through my comments at the top of the source code, and stipulate that the list need to be a proper list without embedded nils. This was the class of functions that I was studying, by design. Another comment presented an algorithm that the author believes will run in O(n) time, which uses the `list?' predicate to check each `car' encountered in the input list to be flattened. Unfortunately, my system configuration is so dramatically different than it was when I ran the original benchmarks, that I'll have to reconstruct the results to make any side by side comparisons. However, `list?' is a predicate I avoid if at all possible, since it requires walking the entire candidate list to validate that is not cyclic. So, without doing the trials, my guess is that it's behavior would be highly dependent on the structure of the input list. It would be quite nice if I stayed online long enough to gather the Schemes of the day, and actually post some data. Take care, and as always, be safe. --kyle