Biological Neg-Entropy
Some of you might have heard that Jonathan Schaeffer and his team at the U. of Alberta recently solved the game of checkers. It made big news in the computer science world.
I first met Jon at the First Computer Olympiad in London (organized by the famous David Levy of chess and computer-chess fame) at which Jon’s program won the gold medal and mine won the silver.
Jon and his team eventually computed the eight-piece endgame database for checkers, and later my colleague Ed Trice and I computed it as well. Jon and I compared results, and it turned out that his database had errors that had evaded his error-detection scheme. This scheme produced internally consistent results, despite the errors. Later, Jon detected errors in my database, which were traced back to a scratch on a CD that evaded my error-detection scheme.
All the errors were eventually traced to data transfer anomalies and not the generative computational algorithms, so CRC (cyclic redundancy check) methods were used to solve the problem.
Read More ›