Steps Towards Efficiently Implementing Program Mutation Systems. The High Level Design of a Distributed Mutation System

Program mutation is a promising method for testing the functional correctness of programs. Its major criticism has been a lack of an efficient implementation method. This report presents four methods which have the potential to dramatically improve the time needed to perform program mutation: distri...

Full description

Saved in:
Bibliographic Details
Main Authors: Sayward,Frederick G, Lipton,Richard J
Format: Report
Language:English
Subjects:
COMMUNICATIONS NETWORKS
COMPILERS
COMPUTER ARCHITECTURE
COMPUTER COMMUNICATIONS
Computer Hardware
COMPUTER LOGIC
COMPUTER PROGRAM VERIFICATION
COMPUTER PROGRAMMING
Computer Programming and Software
Computer Systems
Distributed data processing
DYNAMIC PROGRAMMING
FLOW CHARTING
FORTRAN
HIGH LEVEL LANGUAGES
MUTATIONS
PARALLEL PROCESSING
Program mutation
SEMANTICS
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Program mutation is a promising method for testing the functional correctness of programs. Its major criticism has been a lack of an efficient implementation method. This report presents four methods which have the potential to dramatically improve the time needed to perform program mutation: distributed computation, automatically detecting equivalent mutants, partial mutant execution, and a 'weak' version of program mutation. The presentation of the four potential speed-up methods is given in terms of a high-level design of a prototype distributed program mutation system. One goal of this system will be to analyze how the parallelism inherent in program mutation can be exploited on emerging distributed computer systems consisting of many processors. Another goal is to perform experiments to measure the gain in efficiency realizable by a distributed program mutation system and to compare the relative strength of weak program mutation against program mutation. In the report we also indicate areas which should receive special attention in the detailed design of a prototype distributed mutation system if these goals are to be met. Prepared in cooperation with Yale Univ., New Haven, CT. Dept. of Computer Science.