Candidate metrics for object-oriented software within a taxonomy framework

Fernando Brito E Abreu,Rogério Carapuça

doi:10.1016/0164-1212(94)90099-x

Abstract

This article offers an overview of the state of the art in object-oriented (OO) metrics as well as some new contributions. The usefulness of metrics is reviewed. The inappropriateness of “traditional” metrics to encompass development under the OO paradigm is discussed. A framework for classifying metrics is suggested. Metrics are classified along two vectors: category and granularity, the usefulness and rationale behind each category are presented. Candidate metrics are suggested within the proposed framework. Finally, some research directions that require further effort are identified.

Highlights

Object-Oriented (OO) technology has been pointed out as being one of the keys for solving the software crisis problem
A considerable number of metrics for object-oriented software development has been suggested in this paper within a proposed classification framework (TAPROOT)
This was defined across two independent vectors: category and granularity

Summary

- Introduction

Object-Oriented (OO) technology has been pointed out as being one of the keys for solving the software crisis problem. Examples are the Goal Question Metric (GQM) (Rombach et al, 1987) and other frameworks proposed by some ECC ESPRIT projects (PYRAMID, 1991) (AMI, 1991) They suggest a step-by-step approach that starts with the identification of business objectives (goals), proceeds with the definition of intuitive and well-understood attributes of the software products and process that may represent those goals and derives the mapping from those attributes to a numbering system. Examples quantification of attributes of the length in words of the user manual, lines of software development deliverables source code, number of relations in a database quantification of attributes of the design duration, coding effort, maintenance software development process cost, average effort for the application of 1 test mixture of product and process cost per function point, time to deliver n LOC, metrics average monthly failure rate per I/O interface. A sample of 128 references was overviewed and the following categories arose: Metric's Category No of refs

Design

Method Class System

Findings

- Conclusions and Future Trends