What makes a chess program original? Revisiting the Rybka case

Abstract In this article I will consider the controversial Rybka case, which created a considerable stir in the computer chess community in 2011. Rybka was, from its initial public release in 2005 until it was surpassed in 2010, by far the strongest chess engine ever seen. The question of whether Rybka was substantially or slightly copied from other sources, or a completely original work, is still passionately being debated on forums and blogs. Since the actual source code was not available the determination of whether or not the program was original is complex and involves technical topics such as reverse engineering, program analysis, the abstraction–filtration–comparison test and black box testing such as ponder-hit analysis. The question relates to practices for software development, legal issues like copyright law and touches on innovation and market leadership. Drawing on this array of topics we review the question if and to what extent Rybka was a clone/derivative. The main conclusion is that the view computer experts take on this question is colored by two different views of software development, one that fits programming in an academic environment, one that fits programming in an industry environment. This article is party based on my four part ChessBase article, but includes some new analysis.