SoD-HCER: Comprehensibility as a Design Criterion

University Of Wyoming, Laramie WY

Investigators

Abstract

Directorate for Computer and Information Science and Engineering (CISE) Division Computer and Network Systems (CNS) Science of Design (SoD) Program Proposal Number: 0613919 P/I: Ruben Gamboa Institution: University of Wyoming Title: "SoD-HCER: Comprehensibility as a Design Criterion" Award: $ 157,428 Duration: 24 months The focus of this project is on software design comprehensibility where an objective measure of comprehensibility is proposed that can be determined by the use of automated techniques for detecting program constraints. Using comprehensibility as a criterion for evaluating designs poses a major intellectual challenge because comprehensibility is a subjective notion that complicates the task of detecting it automatically. The proposers' approach toward detecting comprehensibility is to modify two program analysis tools (Daikon and AbsInt) so that those tools can be used to comprehend "good" designs. The project has two main research goals: a) Determine if comprehensibility captures the informal notion of "good" design. This goal represents the empirical validation (or refutation) of the main thesis, that comprehensibility is the key quality shared by good designs; and, b) Identify design features that make designs more comprehensible. The intent here is to analyze different programs that are freely accessible on the Internet to identify designs that are easily comprehended by the tools, and then to see what design features they have in common. This is a highly speculative goal, but it is the main scientific question raised by the research. Moreover, it is one that can have a broad impact on the practice of programming. The thesis of this project, that comprehensibility can be used as a single criterion for evaluating design quality and that this criterion can be measured objectively by using automated tools that infer meaning from programs, provides a single and testable criterion to measure quality. This research brings the art of software design one step closer to a scientific footing. An early benefit of this project will be the identification of design features that lead to greater comprehensibility, hence to better design quality. Moreover, the integration of automated tools to measure design quality objectively is an aspect of this proposal that is unique and innovative. If successful, the project may contribute to the body of knowledge of the science of design by presenting "good" design precepts for practitioners to follow. In addition, the proposed research has the potential to make a significant impact both on computer science education and on the practice of programming in general. If this project is successful in establishing that comprehensibility is an adequate measure of design quality, students and professional designers will have a new and objective way to judge different designs. Program Manager: Anita J. La Salle Date: June 28, 2006