A step toward quality control in computer programming: Understanding the psychology of the management of computer programmers

Abstract

AT THE 1972 ANNUAL MEETING OF THE AMERICAN ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE, SPEAKING BEFORE THE ACM-SPONSORED SESSION ON BETTER COMPUTING FOR RESEARCHERS, DR. RUTH DAVIS, DIRECTOR OF THE CENTER FOR COMPUTER SCIENCE AND TECHNOLOGY OF THE NATIONAL BUREAU OF STANDARDS, STATED THAT THERE WERE SOME 2000 TYPES' OF COMPUTER APPLICATIONS, SUCH AS PAYROLL, WRITING MUSIC, SOLVING DIFFERENTIAL EQUATIONS, ETC. OF THESE, FORTY PERCENT WERE SCIENTIFIC AND ENGINEERING APPLICATIONS (AND THIS KIND OF APPLICATION IS DECREASING RELATIVE TO THE TOTAL), BUSINESS AND MANUFACTURING ACCOUNTED FOR THIRTY PERCENT, AND HUMANITIES APPLICATIONS ALSO TOTALED THIRTY PERCENT. THE ONE COMMON THREAD IN EACH APPLICATION IS THAT MANAGERS CANNOT CONTROL THE QUALITY OF THE PROGRAMS, AND THEY CANNOT CONTROL THE KINDS OF TESTING EMPLOYED, THE PROGRAMS THAT ARE PRODUCED USUALLY HAVE NOT BEEN TESTED THROUGHOUT THE FULL RANGE OF POSSIBLE INPUTS. THE TIME HAS COME, DUE TO THE OFTEN TREMENDOUS COSTS OF SYSTEM FAILURES, TO EMPLOY QUALITY CONTROL PROCEDURES DURING PRODUCTION OF PROGRAMS, SINCE QUALITY CONTROL PROCEDURES ARE AN INTEGRAL PART OF MOST BUSINESS AND PRODUCTION ORGANIZATIONS, ONE MIGHT WONDER WHY SIMILAR PROCEDURES HAVE NOT BEEN USED IN DATA PROCESSING SO FAR.