A VIEW FROM PROGRAMMING
David N. Perkins Steven Schwartz Rebecca Simmons
Like cognition of all sorts, effective problem solving poses profound and provocative puzzles about the relationship between the general and the particular. How, for instance, does one's general knowledge of a domain (for example, knowledge of Newton's laws in physics or FOR-NEXT loops in programming) inform the solving of a specific problem in that domain? To what extent does knowledge that cuts across domains, for example, general problem-solving heuristics and self-management techniques, inform the resolution of a specific problem?
Although these are important questions to ask, some might aver that contemporary cognitive science has gone a long way toward answering them already. Findings from the extensive research on expertise argue that good problem solving within a domain depends mostly on a large repertoire of highly "compiled" domainspecific schemata, what is sometimes called "local knowledge." General crosscutting methods -- "weak methods," as they have been named -- do not count for much. The question about the relation between general and local knowledge dissolves into the resolution that "local knowledge wins," accounting for nearly all