McGill University

Institute Name: McGill University, School of Computer Science

Description: The School of Computer Science at McGill University is widely recognised as a leading institute in research and advanced training in computer science and software engineering. It has 17 professors and 45 doctoral and 81 masters students. It teaches a number of courses at the Bachelor's an Master's level. The Department has extensive links with industry.

The School is involved in research in a number of areas: Computational Geometry, VLSI, Computer Architecture, Parallelism, Program semantics, Robotics, Networks, Artificial Intelligence, Databases, and Software Engineering. The area which will contribute to this project is software engineering, directed by Madhavji.

The Software Engineering Laboratory focuses on techniques, methods, models and tools for the development and improvement of software engineering processes, and on controlled experiments and case studies to investigate phenomena in software processes. Example research projects include: causal models, process assessment criteria, congruence evaluation, process generalisation, organisation readiness, process elicitation, tool insertion, personal software process, process feedback, and process measurement instruments. The laboratory comprises 1 permanent academic staff, 1 visiting academic staff, 3 research associates, 5 exchange students, 9 research students.

A substantial part of the research in the laboratory has been dedicated to studies in requirements engineering. In the Macroscope project (funded by a consortium of organisations in Quebec), for example, research has resulted in causal models of requirements engineering processes, criteria for assessing the success of a requirements engineering process, congruence evaluation of requirements engineering processes, and generalisation of requirements engineering processes. In the PRS project (funded by IBM Canada Ltd.), research has resulted in formal models of the division-wide requirements engineering processes, and impact analysis of tool insertion in the requirements engineering process. In the Namima project (funded by the Networks of Centres of Excellence managing body, PRECARN), we are investigating requirements for evolving processes. A recent project, funded by Transport Canada, is aimed at investigating the advantages and disadvantages of formally defined models of the requirements engineering and change processes.

Below are descriptions of four requirements-engineering-centred activities during 1996: Barriers to the implementation of a requirements change process; Process Modelling; Requirements Changes; and User-centred Design.

• Barriers to the implementation of a requirements change process: A case study was carried out by El Emam and Madhavji on investigating the barriers to the implementation of a requirements change process in a systems engineering organization using a a large real-time system. Our study also identified suggested solutions for overcoming these barriers. The results indicate that the barriers to the implementation of the change process, in descending order of priority, include: weak senior management commitment, resistance to change by the users (procurers, project managers, and end users), weaknesses in the requirements documentation, and weaknesses in the change process itself. The suggestions for improvement, in descending order of priority, are: senior management action, education, process changes, and customization of the requirements documents. In addition, the approach used in this study may be of utility to other researchers investigating problems with the implementation of requirements processes for large scale systems.
• Process Modelling: An exercise in modelling an industrial-scale process for changing the Technical Baseline for an Air Traffic Control System was carried out by Hoeltje and Madhavji. Source documents were in a natural language. Final models were in represented in the X-Elicit tool (built at McGill) and Statemate (from iLogix). The preliminary results of this study indicate that process modelling has helped improve numerous inconsistencies found in the source documents. It was also found that many process attributes were missing in the source desciptions although these existed implicitly in the organisational culture.
• Requirements Changes: A case study of requirements changes due to an evolving environment was carried out by Nanda and Madhavji. The system studied was a "process congruence evaluation system" (CES). CES was a proof of concept system, developed over two years with three person-years of effort. It was part of an environment which was constantly evolving, and this change in the environment dictated the fitness of the CES system in the environment. We assessed the state of the system requirements as well as that of the environment at different times during system development and re-development. From this experience, the key contribution of this work is a detailed insight into how environmental evolution affects system survivability in terms of requirements changes.
• User-centred Design: In an attempt to map out a "process architecture" of an organisation, El Fakih and Madhavji are investigating how software development processes interact with "contextual" processes such as the "common checkpoint process", "investment process" and the "user-centred design (UCD) process". A component of this study focuses on how the UCD and the contextual processes feed information into the requirements engineering processes within the software development projects, and vice versa. The goal here includes studying communication patterns, focusing on both product and process data.

Researchers associated with the project:

Prof. Nazim H. Madhavji, McGill University

Dr. Khaled El Emam, CRIM

Dirk Hoeltje, McGill University

Vivek Nanda, McGill University

Khaled El Fakih, McGill University

Brief CVs of Key Researchers:

Name: Prof. Nazim H. Madhavji

Nationality: U.K. & Canada

Academic Qualifications & Experience: Nazim Madhavji is a professor at the School of Computer Science, McGill University. In 1993, he was appointed as Research Director of the Software Process Programme at Centre de Recherche Informatique de Montreal (CRIM). His research interests are in software engineering, software processes, requirements engineering, project management, software environments and programming languages. He was head of the software process research project in the Macroscope initiative involving many organisations in Quebec. Also, he was leading ProM Canada, the Canadian component of a Canada-Germany joint research project in software processes involving McGill/CRIM, GMD (German National Research Centre in Computer Science) and FernUniversitaet Hagen. He is Technical Co-ordinator and a principal investigator in the Process Reuse Project involving IBM Canada Ltd., Carnegie Mellon University, McGill University, Columbia University, and University of Maryland. He is a principal investigator in a software evolution project, called Namima, involving McGill University, University of Toronto, University of Alberta, University of Victoria and IBM Canada. He is Chair of the IEEE TCSE committee on software process. He was Program Co-chair of CASE'92, and Co-chair of International Workshop on Software Process Evolution, held in Quebec in January, 1993. He is co-chair of the international workshop on Software Evolution Processes and Measurements, held in Victoria in September 1994. He is and has been a member of Program Committees of many international software engineering and software process conferences. He has chaired sessions at international conferences on the subject of software processes and software environments. He was a Guest editor of the September 1991 IEE/BCS Software Engineering Journal special issue on Software Process and Its Support. He was a Guest editor of the December 1993 IEEE Transactions on Software Engineering special issue on the Evolution of Software Processes. He is on the Advisory Editorial Board of the Journal of Software Maintenance. He has led Quebec and Canadian missions to foreign countries on the subject of software engineering, software processes and CASE technologies. He is a consultant to several organisations in the field of software engineering and process technology.

Selected Publications:

Madhavji, N.H.(1992): Environment Evolution: The Prism Model of Changes, IEEE Transactions on Software Engineering, vol. 18, no. 5, pp. 380-392

Madhavji, N.H., Hoeltje, D., Hong, W. and Bruckhaus, T.(1994): Elicit: A Method for Eliciting Process Models, Proc. 3rd Int. Conf. on Software Process, Reston, Virginia, pp. 111-122

El Emam, K. and Madhavji, N.H.(1995): Requirements Engineering Practices in Information Systems Development: A Multiple Case Study, To appear, Proc. International Symposium on Requirements Engineering, York, England

El Emam, K. and Madhavji, N.H.(1995): Measuring the Success of Requirements Engineering Processes, To appear, Proc. International Symposium on Requirements Engineering, York, England

Perez, G., El Emam, K., and Madhavji, N.H.(1995?): A System for Evaluating Congruence of Process Models, submitted to IEEE Transactions on Software Engineering.

El Emam, K. and Madhavji, N.H.(1995?): Criteria for Assessing the Success of a Requirements Engineering Process, submitted to ACM Transactions on Software Engineering and Methodology.

Contact Details:

Principal Investigator: Dr. Nazim Madhavji,

Address: School of Computer Science, McGil University, 3480 University Street, Montreal, H3A 2A7, Quebec, CANADA

Telephone number: + 1 514 398 3740

Telefax number: + 1 514 398 3883

EMail: madhavji@opus.cs.mcgill.ca

List of Members