PROCESS AND MODEL FOR DETERMINING THE REUSABILITY OF AN EXISTING SOFTWARE ASSET

DOD WILL DEVELOP AND RE-ENGINEER BILLIONS OF LINES-OF-CODE OVER THE NEXT FEW YEARS TO MEET OPERATIONAL REQUIREMENTS. SOFTWARE REUSE CAN BE A MAJOR FACTOR IN REDUCING THE BOST OF DEVELOPMENT, RE-ENGINEERING, AND QUALITY IMPROVEMENT. KEY TO SOFTWARE REUSE IS THE ABILITY TO COST EFFECTIVELY IDENTIFY AND RE-ENGINEER EXISTING SOFTWARE ASSETS FOR REUSE AS WELL AS TO COST EFFECTIVELY DEVELOP NEW ASSETS FOR FUTURE REUSE. THIS STUDY ADDRESSES THE PROBLEM OF CHARACTERIZING SOFTWARE ATTRIBUTES THAT DETERMINE THE REUSABILITY OF SOFTWARE ASSETS AND FOR- MULATING NUMERICAL FACTORS THAT ARE PREDICTIVE INDICATORS OF REUSE. THE STUDY INCLUDES FORMULATION OF AN EXPERIMENTAL INEAR REGRESSION MODEL THAT PROVIDES ESTIMATORS BASED ON DEFINED METRICS FOR A SET OF REUSE INDICATORS AS WELL AS AN OVERALL MEASURE OF REUSABILITY. THE REUSE INDICATORS MEASURE THE SUITABILITY OF THE ASSETS FOR REUSE AND POINT TO THE STRATEGY FOR RE-ENGINEERING THE ASSET FOR REUSE. THE MODEL ALSO PROVIDES A SINGLE REUSE FACTOR SUMMARIZING THE INDICATORS THAT PROVIDE A PREDICTION OF REUSE SUCCESS. THE MODEL CAN ALSO BE USED TO ESTIMATE THE REUSE INDICATORS AND OVER- ALL REUSE FACTOR AT ANY POINT DURING THE DEVELOPMENT PROCESS. THE DEVELOPMENT PROCESS CAN THEN BE MODIFIED TO IMPROVE THE FUTURE RE- USABILITY OF THE ASSET.