SOFTWARE QUALITY METRICS

page 1

SOFTWARE QUALITY METRICS

for

Object Oriented

System Environments

SATC-TR-95-1001

JUNE 1995

National Aeronautics and Space Administration

Goddard Space Flight Center, Greenbelt Maryland 20771

Software

Assurance

Technology

Center

page 2

TABLE OF CONTENTS

I. Introduction

II. Metrics Overview

III. Object-Oriented Overview

IV. Metrics for Object-Oriented Systems

A. Metric Evaluation Criteria

B. Traditional Metrics for Object-Oriented Systems

1. Methods

Metric 1: Cyclomatic Complexity

Metric 2: Lines of Code

C. Object-Oriented Specific Metrics

1. Classes

a. Methods

Metric 3: Weighted Methods per Class (WMC)

b. Messages

Metric 4: Response for a Class (RFC)

c. Cohesion

Metric 5: Lack of Cohesion of Methods (LCOM)

d. Coupling

Metric 6: Coupling Between Object Classes (CBO)

2. Inheritance

Metric 7: Depth of Inheritance Tree (DIT)

Metric 8: Number of Children (NOC)

D. Example

E. Summary

1. Metric Summary

2. COTS for Recommended Object-Oriented Metrics

V. Conclusions

Appendix A: Comprehensive Listing of Object-Oriented Metrics

Appendix B: Object-Oriented Detailed Discussion

1. Object

a. States/Attributes

b. Operations

c. Object Example

2. Class

3. Inheritance

4. Messages

5. Cohesion

6. Polymorphism

7. Object-Oriented Languages

8. Terminology

Appendix C: Cots Packages

Appendix D: Object-Oriented References

page 3

TABLES

Table 1: SATC Metrics for Object-oriented Systems

Table 2: Key Object-Oriented Definitions

Table 3: Summary of Recommended Object-Oriented Metrics

Table 4: Object-Oriented Metrics Supported by COTS

FIGURES

Figure 1: Pictorial Description of Key Object-Oriented Terms

Figure 2: Geometric Classes with Attribute, Operations and Methods

Figure 3: Pseudocode for Perimeter of an Equilateral Triangle

Figure 4: Notation of an Object

Figure 5: Sample Object

Figure 6: Conceptual View

Figure 7: Class with Objects

Figure 8: Inheritance Network

Figure 9: Message Passing

page 4

Software Quality Metrics for Object-Oriented System Environments

I. INTRODUCTION

Object-oriented analysis and design are popular concepts in today’s software development environment. They

are often heralded as the silver bullet for solving software problems,; while in reality there is no silver bullet,

object-oriented has proved its value for systems that must be maintained and modified. Object-oriented

software development requires a different approach from more traditional functional decomposition and data

flow development methods. While the functional and data analysis approaches commence by considering the

systems behavior and/or data separately; object-oriented analysis approaches the problem by looking or system

entities that combine them. Object-oriented analysis and design focuses on objects as the primary agents

involved in a computation; each class of data and related operations are collected into a single system entity.

The concepts of software metrics are well established, and many metrics relating to product quality have been

developed and used. The SATC applies a model for evaluating software quality that has four goals:

(1) Stability of Requirements and Design, (2) Product Quality, (3) Testing Effectively, and

(4) Implementation Effectively. With object-oriented analysis and design methodologies gaining popularity, it

is time to start investigating object-oriented metrics with respect to these goals. We are interested in the

answer to the following questions:

· What concepts and structures in object-oriented affect the quality of the software?

· Can traditional metrics measure the critical object-oriented structures?

· If so, are the threshold values for the metrics the same for object-oriented designs as for

functional/data designs?

· Which of the many new metrics found in literature are useful to measure the critical concepts in

object-oriented?

This report summarizes results of the SATC’s research on metrics for object-oriented systems. We start with a

brief discussion of the metrics recommended by the SATC for object-oriented systems. These metrics include

modifications of “traditional” metrics as well as “new” metrics for specific object-oriented structures. Since the

object-oriented metrics require a cursory understanding of the object-oriented concepts, Section III presents a

pictorial representation of the basic object-oriented structures and defines the key terms. A more extensive

explanation of the object-oriented structures is in Appendix B and is referenced by Section III. In Section IV,

we discuss the proposed object-oriented metrics with respect to the SATC Software Quality Model,

specifically, their relationship to the attributes of quality (Goal 2: Product Quality -Structure/Architecture,

Reuse, Maintainability). In the summary, we will address the availability of COTS packages to facilitate the

collection of these metrics. Details on the COTS packages are given in Appendix C.

II. OVERVIEW - OBJECT-ORIENTED METRICS

In this report, the SATC documents its research into the current status of object-oriented metrics. The

research was done by surveying the literature on object-oriented metrics then applying the SATC experience in

traditional software metrics to select the object-oriented metrics that support the goal of measuring design and

code quality. In addition, we required that a metric be feasible and have a clear relationship to the objectoriented

structures being measured. At this time, many proposed object-oriented metrics lack a theoretical

basis and have not yet been validated. Some of these metrics are too labor intensive to collect, or are too

dependent on the implementation technology. The object-oriented metrics proposed by the SATC can be

related to desirable software qualities.

The SATC’s approach to identifying a set of object-oriented metrics was to focus on the primary, critical

constructs of object-oriented design and to select metrics that apply to those areas. The suggested metrics are

page 5

...