SOFTWARE QUALITY AND PRODUCTIVITY IS EXTREMELY DEPENDENT ON MEASUREMENT APPLIED THROUGH THE USE OF QUANTIFIED METRICS. A QUANTITATIVE APPROACH OF ASSESSING SOFTWARE QUALITY AND PRODUCTIVITY IS INTRODUCED. THE LINEAR MODEL THAT ACCOMMODATES BOTH YIELD (E.G. LOC, FUNCTION POINTS, # BUGS) AND TRAIT (USABILITY, REUSABILITY, ETC.) CAN BETTER REFLECT THE TOTALITY OF SOFTWARE, YET KEEPING ENOUGH FLEXIBILITY TO EMPHASIZE SELECTED ATTRIBUTES VIA COEFFICIENTS. THE LINEAR MODEL COMPUTES OUTCOME, AND SOFTWARE PRODUCTIVITY IS DEFINED AS THE CHANGE IN OUTCOME AS THE RESULT OF EVENT (E.G. TOOL INTRODUCTION, PROCEDURE CHANGE, ETC.). SOFTWARE PRODUCTIVITY CALCULATED AS SUCH APPROXIMATES T-DISTRIBUTION. THE APPROACH THUS HAS WIDE APPLICATION.
| Number of Pages : | 6 |
| Published : | 06/01/1990 |
