Re:Crobots / corewars (Score 1)

I have vast experience in both Core War and in teaching introductory computing courses. I often see Core War recommended as a possible curriculum activity for high school students. I cringe. I would NOT introduce Core War to my students.

Motivation is easy, because the concept is cool: killer program written by evil genius lays barrage of numeric bombs. Also, tournaments can be a blast... if they are administered exactly right. After that, there are big problems.

Typically, the innovations in Core War come from an elite group of players who "get it" and have a "clue". These are the heros who invent all of the revolutionary new strategies. Everyone else can only copy to compete. Most redcode programs are highly derived from past accomplishments.

Furthermore, Core War is played differently now than in 1984 when it was introduced by A.K. Dewdney's articles in Scientific American (yes, I had his permission to scan them). Originally, programmers wanted to try intelligent strategies in complex programs (e.g. self-correcting code). However, the tournament winners today are highly-optimized compact codes, usually running in a tight loop. There may be some interesting mathematics in the optimization process; genetic algorithms are sometimes used. However, the assembly code itself is not very algorithmic.

The space of competitive warriors has been thoroughly explored. The best players may still find some new tricks. However, everyone must work harder and harder to discover and implement interesting new ideas. Again, most of the tournament submissions are highly redundant -- they are the few successful strategies that everyone knows about.

Finally, the current resources haven't evolved beyond the original Scientific American articles. My experiences with confused newbies suggest that the current online tutorials are inadequate.

Overall, I doubt that Core War will benefit students much in the long-term.

-- David M. Moore