clore@cis.ohio-state.edu (michael a clore) (05/08/91)
The following message from Bill Poucher was distributed to ACMSTCHP. Subject: Time:5:52 PM OFFICE MEMO Programming Contest Apology Date:5/6/91 Every so often I see a comment about the Programming Contest not being a realistic measure of team software development performance. I agree. It is not a realistic measure of team software development performance. In fact, should someone successfully defend a measure of team software development performance, they could probably solve the Arab-Israeli problem overnight. The Programming Contest is a programming contest, pure and simple. It tests who can develop the most solutions to abstractly-posed problems in the least amount of time - sort of. It certainly develops an appreciation for team dynamics and encourages synthesizing a survey of algorithms typical of those seen in a Computer Science program. But the contest is meant to be fun (first), challenging (second) and meaningful only in the fact that it is fun and challenging. In the early years of the contest, there were special prizes for the best structured code. Funny thing, the winning team almost always wrote the best structured code, so the award was dropped as being redundant. So, writing robust, structured code seems to pay off in the contest. So does cooperation and practice analysing problem statements. Developing a reasonable test plan helps too, because most teams that "nearly" solve a problem miss it because they improperly handle boundary conditions. The contest problem set has challenged computer science curricula by introducing coding problems and graphics problems before they were broadly covered by the curricula. It is rare when recursion is not a needed tool. Furthermore, team dynamics must be integrated into any winning team management strategy - a topic overlooked by many Computer Science curricula. There have been maintenance problems posed in the past. (Fix the following program to do what you think that it was designed to do.) Funny thing. Just like in real life, nobody wanted to do those problems. So we rarely offer them anymore. The contest is a game requiring enough skill so that if you win, you can claim intellectual superiority, and if you lose, you know that it was just dumb luck. While it certainly is not perfect, I hope that it never becomes so. But, I would trust the Contest Finals results more than the results of the GRE Advanced Computer Science exam as an indicator of initiative, breadth of exposure to algorithms, and excitement in pursuing computing. Bill Poucher, ACM Scholastic Programming Contest Director -- Michael Allen Clore "It is unworthy of excellent (persons) to lose OSU-ACM Chair hours like slaves in the labour of calculation" Baron Gottfried Wilhelm von Leibniz (1646-1716) <CLORE@OHSTPHRM.PHARMACY.OHIO-STATE.EDU> or <CLORE@OHSTPHRM.BITNET>