AN EXPLORATORY INVESTIGATION OF PROGRAMMER PERFORMANCE UNDER ON-LINE AND OFF-LINE CONDITIONS.

Abstract

An experiment conducted at System Development Corporation compared the program debugging performance of programmers working under conditions of on-line and off-line access to a computer. This paper describes that experiment; it is the first known study measuring the performance of programmers under controlled conditions for a standard task. Two groups of six programmers each, comprising a total sample of 12 subjects, debugged two types of programs under on-line and off-line conditions in accordance with a Latin-Square experimental design. The on-line condition was the normal mode of operation for the SDC Time-Sharing System; the off-line condition was a simulated closed-shop with a two-hour turn around time. Statistically significant results indicated faster debugging under on-line conditions. Perhaps the most important practical finding of this study, overshadowing on-line/off-line differences, concerned the large and striking individual differences in programmer performance. Attempts are made to relate observed individual differences to objective measures of programmer experience and proficiency through factorial techniques. In line with the exploratory objectives of this study, methodological problems encountered in designing an conducting this type of experiment are described, limitations of the findings are pointed out, hypotheses are presented to account for the results and suggestions are made for further research. (Author)

Document Details

Document Type
Technical Report
Publication Date
Sep 02, 1966
Accession Number
AD0640654

Entities

People

  • E. E. Grant
  • H. Sackman

Organizations

  • System Development Corporation

Tags

DTIC Thesaurus Topics

  • Combinatorial Analysis
  • Computers
  • Computing-Related Activities
  • Corporations
  • Data Science
  • Debugging
  • Experimental Design
  • Hypotheses
  • Information Science
  • Mathematical Analysis
  • Mathematics
  • Standards

Fields of Study

  • Engineering

Readers

  • Computer Science.
  • Instructional Design and Training Evaluation.
  • Systems Analysis and Design