On Heuristic Models, Assumptions, and Parameters

Study of the interaction between computation and society often focuses on how researchers model social and physical systems in order to specify problems and propose solutions. However, the social effects of computing can depend just as much on obscure and opaque technical caveats, choices, and qualifiers. These artifacts are products of the particular algorithmic techniques and theory applied to solve a problem once it has been modeled, and their nature can imperil thorough sociotechnical scrutiny of the often discretionary decisions made to manage them. We describe three classes of objects used to encode these choices and qualifiers: heuristic models, assumptions, and parameters, and discuss selection of the last for differential privacy as an illustrative example. We raise six reasons these objects may be hazardous to comprehensive analysis of computing and argue they deserve deliberate consideration as researchers explain scientific work.