Guidelines for reporting results of computational experiments. Report of the ad hoc committee

This report on guidelines for reporting results of computational experiments is written by subcommittee of the Committee on Algorithms (COAL). The authors have two goals: (1)to present a concise set of principles to guide authors, editors, and referees in assessing computational tests, and (2)to clarify existing guidelines in certain areas. They do not provide a step-by-step procedure, but rather three principles that form a basis for performing and judging computational experiments. These principles are: (1)the results presented must be sufficient to justify the claims made; (2)there must be sufficient detail to allow reproducibility of results; and (3)it is not the referee’s duty to reproduce the results.