Problems with software complexity measurement

Authors:
Robert L. Sedlmeyer;Joseph K. Kearney;William B. Thompson;Michael A. Adler;Michael A. Gray
Affiliations:
Department of Computer Technology, Indiana University-Purdue University, Fort Wayne, Fort Wayne, Indiana and Computer Science Department, University of Minnesota, Minneapolis, MN;Department of Computer Science, University of Iowa, Iowa City, Iowa and Computer Science Department, University of Minnesota, Minneapolis, MN;Computer Science Department, University of Minnesota, Minneapolis, MN;Control Data Corporation;Control Data Corporation
Venue:
CSC '85 Proceedings of the 1985 ACM thirteenth annual conference on Computer Science
Year:
1985

Citing 10
Cited 0

A Review and Evaluation of Software Science

ACM Computing Surveys (CSUR)
A genealogy of control structures

Communications of the ACM
Elements of Software Science (Operating and programming systems series)

Elements of Software Science (Operating and programming systems series)
Methods for improving controlled experimentation in software engineering

ICSE '81 Proceedings of the 5th international conference on Software engineering
Third time charm: Stronger prediction of programmer performance by software complexity metrics

ICSE '79 Proceedings of the 4th international conference on Software engineering
A model for program complexity analysis

ICSE '78 Proceedings of the 3rd international conference on Software engineering
Using a behavioral theory of program comprehension in software engineering

ICSE '78 Proceedings of the 3rd international conference on Software engineering
An experimental investigation of the effect of program structure on program understanding

Proceedings of an ACM conference on Language design for reliable software
Measurement of program complexity by the pair: (Cyclomatic Number, Operator Count)

ACM SIGPLAN Notices
Reliable software through composite design

Reliable software through composite design

Quantified Score

Hi-index	0.00

Visualization

Abstract

Theoretical and methodological problems in the development of program-based measures of software complexity are enumerated. Three issues are analyzed: the potential confounding influence of non-program factors; the desirable features of complexity measures, and the difficulties in validating complexity measures. The analysis suggests that researchers must find ways to control the effects of non-program factors, perhaps by restricting their focus to a well-defined programming environment. Further, both the designers and users of complexity measures must be aware of the inherent limitations of such tools.