Elsevier

Knowledge-Based Systems

Volume 11, Issue 2, 12 October 1998, Pages 125-143
Knowledge-Based Systems

Requirements for the knowledge-based support of software engineering measurement plans

https://doi.org/10.1016/S0950-7051(98)00043-4Get rights and content

Abstract

In order to improve the quality of software systems, measurement programs have been implemented in many companies to support process improvement activities. The planning and implementation of a successful measurement program requires, in practice, a significant amount of effect. Cost may be reduced and quality of measurement may be improved by providing knowledge-based support and reusing experiences gathered on past measurement programs. In this article, we state the requirements for the knowledge-based support of planning measurement programs based on the Goal/Question/Metric paradigm. Reuse opportunities are precisely identified, the knowledge to be captured for effective reuse is identified and structured, and reuse scenarios are provided.

Introduction

For the continuous improvement of software quality and productivity within organizations, capturing and reusing explicit software development know-how is essential. What represents relevant software know-how differs among software organizations, depending on their specific development environment and objectives. Therefore, this software know-how has to be derived by organization-specific measurement programs, taking into account the environment characteristics and company-specific improvement goals, e.g. reduce development effort by 10% during the next year. For continuous improvement of software processes and products, the establishment of goal-orientated measurement programs is essential for deriving relevant quantitative and qualitative data on software processes and products. In this context, the Goal/Question/Metric Paradigm (GQM) 1, 5is of particular interest, since it helps in defining and implementing operational and measurable software improvement goals. However, measurement programs are known to be still difficult to plan and implement, especially when one lacks the required experience. Similarly to software development know-how, in order to institutionalize systematic and continuous learning in the organization, measurement know-how has also to be explicitly captured, modeled and reused to support the planning of future measurement programs. Therefore, experiences regarding the use of GQM, the know-how related to its products and processes, needs to be captured and made reusable organization-wide.

In this paper, we focus on requirements and principles for the knowledge-based support of planning GQM-based measurement programs. A motivation for the reuse of experiences and know-how from past measurement programs is given in Section 2. The potential benefits of experience-based support are discussed in Section 3. The knowledge which should be captured in a reusable form for future measurement programs is identified in Section 4. Section 5addresses the modeling and representation of this knowledge in an experience base. Scenarios for the reuse of the knowledge are provided in Section 6. Conclusions and future research directions are discussed in Section 7.

Section snippets

Motivation

Building up organization-specific software development know-how, the company has to learn continuously from its software projects by developing tailored, context-specific quality and resource models based on quantitative and qualitative data derived through measurement programs. To be effective and efficient, measurement programs must also be tailored to the characteristics of the specific organization, the software processes and products and company-specific goals. Therefore, one of the major

What activities need knowledge-based support?

In this section we describe the GQM planning process for which we claim knowledge-based support ought to be provided. The GQM approach is a specific technology for goal-orientated measurement in software projects. In order to be tailored to the needs, GQM-based measurement programs have to be specified precisely and explicitly by a detailed measurement goal. The measures are derived in a top-down fashion, based on goals via a set of questions and quality/resource models. This refinement is

What kind of knowledge is reusable?

In the previous section, we provide an overview on the GQM planning process and the activities that need knowledge-based support. In this section, we examine the opportunities in terms of knowledge reuse for the GQM planning process. A taxonomy of reusable components in the planning phase is defined in detail in the next subsections.

Representation and packaging of knowledge

In the last section, we identify the knowledge which can be reused and support the planning of measurement programs. In this section we examine how the knowledge gained on individual projects has to be packaged and represented.

In order to allow intelligent and efficient reuse, additional knowledge has to be captured. This additional knowledge will be used to identify adequate experiences with respect to the characteristics of a new project. It can be classified into the following categories:

Reuse scenarios

In Section 4, we described which kind of knowledge can be reused in order to support the GQM planning process. In order to illustrate the potential reuse of this knowledge, we provide some scenarios focusing on particular steps of the planning of a measurement program.

Conclusions

For continuous learning in a software organization, the software development know-how available from past projects has to be packaged and reused in new projects. In this paper, we focus on gained software measurement know-how. We show that there is a great potential for reducing the effort related to the measurement planning phase and for improving the adequacy, consistency and effectiveness of a measurement plan through reuse. Within the context of the GQM paradigm for software measurement, we

References (13)

  • L.C. Briand, C.M. Differding, H.D. Rombach, Practical Guidelines for Measurement-Based Process Improvement. Software...
  • L.C. Briand, J.W. Daly, J. Wüst, A unified framework for coupling measurement in object-orientated systems. Technical...
  • L.C. Briand, S. Morasca, V.R. Basili, Defining and validating high-level design metrics. Technical Report CS-TR-3301,...
  • L. Briand, W. Melo, C. Seaman, V. Basli, Characterising and assessing a large-scale software maintenance organisation....
  • V.R. Basili et al.

    The TAME project: towards improvement-orientated software environments

    IEEE Transactions on Software Engineering SE-

    (1988)
  • The CEMP Consortium, Customised establishment of measurement programs. Final Report, ESSI Project No. 10358, Germany,...
There are more references available in the full text version of this article.

Cited by (0)

View full text