guest editor's introduction

1
J. SYSTEMS SOFTWARE 209 1994; 24:209 Guest Editor’s Introduction Warren Harrison Department of Computer Science, Portland State Uniuersity, Portland, Oregon It has been my pleasure over the past 5 years to be involved in the organization and coordination of the Annual Oregon Workshop on Software Metrics (AOWSM). It has been gratifying to watch the Workshop grow from a mainly regional affair held on the Portland State University campus to a truly international event held annually at the Silver Falls Conference Center in the Oregon forest. The four articles in this special issue of the Jour- nal of Systems and Software were selected from the dozen that were presented at the AOWSM in March 1993. Each article has been expanded in response to feedback from participants of AOWSM ‘93 and postworkshop reviews. Two of the articles (“Alternative Approaches for the Use of Metrics to Order Programs by Complex- ity” by Khoshgoftaar et al. and “Real-Time Software Metrics” by Cook and Roesch) deal extensively with the idea of complexity domains. Briefly, the idea is that virtually all software complexity metrics can be put into one of a small number of categories such as size domain, control flow domain, and data domain. The implication is that when considering the com- plexity of a program, use of multiple metrics in the same complexity domain may be redundant, but, more importantly, assessment by a set of metrics that does not include all the complexity domains is incomplete. Khoshgoftaar et al. propose a relative complexity metric derived through the use of principal compo- nents analysis. Their article establishes the viability of the new metric by examining the relationship between the metric and faults in a military software system. Cook and Roesch attempt to characterize the evolution of real-time software systems. Al- though real-time systems appear to evolve in a man- ner similar to other software, they seem to demon- strate additional complexity domains that are not as evident in non-real-time software. In particular, their paper suggests not only that the information flow domain is important for real-time systems, but that the complexity domain is comprised of two subdomains: an information inflow domain and an information outflow domain. The other two articles that appear in this special issue (“A Software Metric Set for Program Mainte- nance Management” by Stark et al. and “Construc- tion and Testing of Polynomials Predicting Software Maintainability” by Oman and Hagemeister) focus on the role of measurement in the software mainte- nance activity. It is interesting to note that these two articles also make use of the concept of complexity domains in the process of arriving at minimal sets of maintenance metrics. Stark et al. describe a set of metrics that are used in software developed for the NASA space shuttle program. Their article describes the 13-metric set and addresses their implementation. Oman and Hagemeister describe the development of a series of polynomial models that have been shown to be highly correlated with assessments of program main- tainability made by an extensive questionnaire-style instrument used by the U.S. Air Force. The authors of these four articles, as well as the authors of all the other papers presented at AOWSM ‘93, should be appreciated not only for the advances they are making in the development and use of software measurement, but also for their willingness to share the fruits of their labor with others, both in workshops like AOWSM ‘93 and archival journals such as the Journal of Systems and Software. Their contributions to the field cannot be overestimated. 0 Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010 0164.1212/94/$7.00

Upload: warren-harrison

Post on 21-Jun-2016

216 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: Guest editor's introduction

J. SYSTEMS SOFTWARE 209 1994; 24:209

Guest Editor’s Introduction

Warren Harrison Department of Computer Science, Portland State Uniuersity, Portland, Oregon

It has been my pleasure over the past 5 years to be involved in the organization and coordination of the Annual Oregon Workshop on Software Metrics (AOWSM). It has been gratifying to watch the Workshop grow from a mainly regional affair held on the Portland State University campus to a truly international event held annually at the Silver Falls Conference Center in the Oregon forest.

The four articles in this special issue of the Jour- nal of Systems and Software were selected from the dozen that were presented at the AOWSM in March 1993. Each article has been expanded in response to feedback from participants of AOWSM ‘93 and postworkshop reviews.

Two of the articles (“Alternative Approaches for the Use of Metrics to Order Programs by Complex- ity” by Khoshgoftaar et al. and “Real-Time Software Metrics” by Cook and Roesch) deal extensively with the idea of complexity domains. Briefly, the idea is that virtually all software complexity metrics can be put into one of a small number of categories such as size domain, control flow domain, and data domain. The implication is that when considering the com- plexity of a program, use of multiple metrics in the same complexity domain may be redundant, but, more importantly, assessment by a set of metrics that does not include all the complexity domains is incomplete.

Khoshgoftaar et al. propose a relative complexity metric derived through the use of principal compo- nents analysis. Their article establishes the viability of the new metric by examining the relationship between the metric and faults in a military software system. Cook and Roesch attempt to characterize the evolution of real-time software systems. Al- though real-time systems appear to evolve in a man-

ner similar to other software, they seem to demon- strate additional complexity domains that are not as evident in non-real-time software. In particular, their paper suggests not only that the information flow domain is important for real-time systems, but that the complexity domain is comprised of two subdomains: an information inflow domain and an information outflow domain.

The other two articles that appear in this special issue (“A Software Metric Set for Program Mainte- nance Management” by Stark et al. and “Construc- tion and Testing of Polynomials Predicting Software Maintainability” by Oman and Hagemeister) focus on the role of measurement in the software mainte- nance activity. It is interesting to note that these two articles also make use of the concept of complexity domains in the process of arriving at minimal sets of maintenance metrics.

Stark et al. describe a set of metrics that are used in software developed for the NASA space shuttle program. Their article describes the 13-metric set and addresses their implementation. Oman and Hagemeister describe the development of a series of polynomial models that have been shown to be highly correlated with assessments of program main- tainability made by an extensive questionnaire-style instrument used by the U.S. Air Force.

The authors of these four articles, as well as the authors of all the other papers presented at AOWSM ‘93, should be appreciated not only for the advances they are making in the development and use of software measurement, but also for their willingness to share the fruits of their labor with others, both in workshops like AOWSM ‘93 and archival journals such as the Journal of Systems and Software. Their contributions to the field cannot be overestimated.

0 Elsevier Science Inc.

655 Avenue of the Americas, New York, NY 10010 0164.1212/94/$7.00