Lehner/Dumke/ Abran (Eds.) Software Metries GABLER EDITION WISSENSCHAFT Information Engineering und IV- Controll in g Herausgegeben von Professor Dr. Franz Lehner Die Schriften reihe präsentiert aktuelle Forschungsergebnisse der Wirtschaftsinformatik sowie interdisziplinäre Ansätze aus Informatik und Betriebswirtschaftslehre. Ein zentrales Anliegen ist dabei die Pflege der Verbindung zwischen Theorie und Praxis durch eine an wendungsorientierte Darstellung sowie durch die Aktualität der Bei träge. Mit der inhaltlichen Orientierung an Fragen des Information Engineerings und des IV-Controllings soll insbesondere ein Beitrag zur theoretischen Fundierung und Weiterentwicklung eines wichtigen Teilbereichs der Wirtschaftsinformatik geleistet werden. Franz Lehner/Reiner Dumke/ Alain Abran (Eds.) Software Metries Research and Practice in Software Measurement DeutscherUniversitäts Verlag Die Deutsche Bibliothek -CIP-Einheitsaufnahme Software Metrics : research and practice in software Measurement / Hrsg.: Franz lehner .... -Wiesbaden: Dt. Univ.-Verl.; Wiesbaden: Gabler, 1997 (Gabler Edition Wissenschaft : Information Engineering und IV-Controlling) ISBN 978-3-8244-6518-7 ISBN 978-3-322-99929-0 (eBook) DOI 10.1007/978-3-322-99929-0 Der Deutsche Universităts-Verlag und der Gabler Verlag sind Unternehmen der Bertelsmann Fachinformation. Gabler Verlag, Deutscher Universităts-Verlag, Wiesbaden © Betriebswirtschaftlicher Verlag Dr. Th. Gabler GmbH, Wiesbaden 1997 lektorat: Claudia Splittgerber Dos Werk einschlieBlich aller seiner Teile ist urheberrechtlich geschUtzt. Jede Verwertung auBerhalb der engen Grenzen des Urheberrechts gesetzes ist ohne Zustimmung des Verlages unzulăss!9 und strafbar. Das gilt insbesondere fur Vervielfăltigungen, Ubersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elek tronischen Systemen. http://www.gabler-online.de H6chste inhaltliche und technische Qualităt unserer Produkte ist unser ZieI. Bei der Produktion und Auslieferung unserer Bucher wollen wir die Umwelt schonen: Dieses Buch ist auf săurefrei­ em und chlorfrei gebleichtem Papier gedruckl. Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in diesem Werk berechtigt auch ohne besondere Kennzeichnung nicht zu der Annahme, dal3 solche Namen im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei zu betrachten wăren und daher von jedermann benutzt werden durften. ISBN 978-3-8244-6518-7 Foreword There ubiquity of software in the development of complex systems and high technology consumer products is steadily increasing in a competitive environment that is now global in scope. Thus software developers are faced with the challenge of making software systems and products of ever greater quality and safety, while at the same time being faced with the growing pressure of costs reduction in order to gain and maintain competitive advantages. As in any scientific and engineering discipline, reliable measurement is essential for talking on such a challenge. "Software measurement is an excellent abstraction mechanism for learning what works and what doesn't" (Victor Basili). Measurement of both software processes and products provides a large amount of basic information for the evaluation of the software development processes or the software products themselves. Examples of recent successes in software measurement span multiple areas, such as evaluation of new development methods and paradigms, quality and management improvement programs, tool-supporting initiatives and company-wide measurement programs. The German Computer Science Interest (GI) Group of Software Metrics, the (GI) Interest Group of Quality Improvement of Object-Oriented Systems and the Canadian Interest Group in Software Metrics (CIM) have attended to these concems in the recent years. Research initiatives were directed initially to the definition of software metrics and then to validation of the software metrics themselves. This was followed by more and more investigation into practical applications of software metrics and by critical analysis of the benefits and weaknesses of software measurement programs. Key findings in this area of software engineering have been published in some important books, such as Zuse's Software Complexity, AIlerburg and Abran's Metries in Software Evolution, Durnke and Zuse's Theory and Practice of Software Measurement, Durnke et al.'s Software Quality with Metric Tools, Lehner's Software Maintenance and Ebert and Durnke's Software Metries in Practice. VI Foreword This new book inciudes key papers presented at the 6th International Workshop on Software Metrics in Regensburg (Germany), September 1996. It is a coIlection of theoretical studies in the field of software measlirement as weIl as experience reports on the application of software metrics in Canadian, Austrian, Belgian and Gerrnan companies and universities. Some of these papers and reports describe new software measurement applications and paradigms for knowledge-based techniques, maintenance service evaluation, factor analysis discussions and neural-fuzzy applications. Others address the object-oriented paradigm and discuss the application ofthe Function Point approach to an object-oriented design method, the evaluation of the Java development environment, the analysis of quality and productivity improvements of object-oriented systems, as weIl as the definition of the metrics of ciass libraries. Other papers offer a different perspective, presenting a software measurement education system designed to help improve the lack of training in this field, for example, or they inciude experience reports about the implemention of measurement programs in industrial environments. This book will be of interest to software engineering researchers, as weIl as to practitioners in the areas of project management and quality improvement prograrns, for both software maintenance and software development in general. Alain Abran Reiner Durnke Franz Lehner Table of Contents Foreword ................................................................................................................................... v Table of Contents .................................................................................................................... vn I. Quality and Measurement of Object-Oriented Software ........................... 1 R. Hubig, I. Morschel Quality and Produetivity Improvement in Objeet-Oriented Software Development ................. 3 M. Hitz, S. Stiller Automatie Extraetion of Objeet -Oriented Software Metries ................................................... 15 R.Dumke Really Objeet-Oriented Software Metrics ................................................................................ 27 T. Fetcke, A. Abran, T.-H. Nguyen Mapping the OO-Jaeobson Approach to Funetion Point Analysis .......................................... 59 11. Internet and World Wide Web ................................................................... 75 A. Winkler, R. Dumke, R. Koeppe, G. Kompf Efficieney and Maintainability of JAVA Applications ............................................................ 77 B. Duhamel, G. St-Amant, A. Abran Design of a Measure to Assess Complianee of Internet Web Sites with Privacy Laws ........... 95 111. Software Metrics and Measurement ...................................................... 109 M. Maya, A. Abran, P. Bourque Measuring the Size of Small Funetional Enhaneements to Software .................................... 111 D. Schmelz, M. Schmelz The Use ofFaetor Analysis in the Area of Software Metries ................................................ 123 E. Baisch, C. Ebert On a Neural-Fuzzy Teehnique with GA-Optimization for Software Quality Models ........... 131 H. Zuse The Software Measure Information System: ZD-MlS ........................................................... 139 VIll Table of Contents IV. Quality Improvement and Validation Process ...................................... 165 F. Houdek Software Quality Improvement by Using an Experience Factory .......................................... 167 J.-M. Deshamais, P. Mords Validation Process in Software Engineering: An Example with Function Point.. ................. 183 C. Ebert Applying Knowledge-Based Techniques to Software Quality Management. ........................ 193 A. Mittelmann Implementation of a Measurement Plan in an Industrial Environment.. ................................ 211 List of Contributors ................................................................................................................ 229 I. Quality and Measurement of Object Oriented Software Quality and Productivity Improvement in Object-Oriented Software Development Roland Hubig, Ivan Morschel Abstract The quality movement in the last years has a drarnatic impact on all industrial sectors, reaching recently the systems and software industry. Reuse of products, processes and experiences origi nating from the system life-cycle is seen today as a feasible solution to the problem of developing higher quality systems at a lower cost. Managers and developers have to decide among different approaches to improve productivity and software quality. This paper relates the experiences in introducing., establishing and validating a goal-oriented approach to measurement and control of the quality and productivity in an industrial environment. Firstly, we will describe the basic quality methods applied like QIP and software metrics, then the main characteristics of the envi ronment at the Daimler-Benz Aerospace AG (DASA) in Ulm. Finally, we present the experi ences gained with the improvement of the industrial object- oriented systems development. 1. Introduction Software systems are present in almost every activity or artifact in our society, being tardy per ceptible when software problems arise. However, software is often not completely understood or developed in a controlled way. In particular, there is a certain difficulty in dealing with software quality or productivity. What means in essence quality software? How can we produce quality software? How to sys tematically introduce new technologies in running projects ? How to effectivelly select and use appropriate metrics to measure quality of software during development ? In the industry there are different approaches like CMM, Bootstrap, AMI, etc. supporting assessment of software. Of par ticular interest is the Quality Improvement Paradigm which involves the use of a goal-oriented approach to empirically, i.e. based on the characteristics of the target environment, improve software quality. It gives support for analyzing and controlling quality of product and processes during their development. Specially by reducing the arnount of rework and reusing software arti facts across segments of many different projects and by planning, sustaining and continually im-