ICSP5: Computer Supported Organizational Work

Abstract: Method engineering is concerned with techniques for modeling systems during the analysis and design phases of software development. Method engineering includes the creation of notations that represent the concepts to be modeled, definition of semantics for those notations, and development of processes for building and validating the models. Effective use of methods is aided by automated tools that support the methods. A method engineering toolset should include a metamethod mechanism that supports the creation of a tool to support any modeling approach that is defined.

This talk gives an overview of methods engineering and customization of methods, relates these aspects to the overall software development process, and describes metamodeling tools that can be used to support a defined method.

Biography: Anthony I. (Tony) Wasserman is Principal of Software Methods and Tools, which provides consulting on software development strategies. He was the founder of Interactive Development Environments (IDE), which developed the innovative Software through Pictures modeling environment. Prior to starting IDE, Tony was a University of California professor, where he developed the User Software Engineering methodology and supporting tools. He is a Fellow of oth ACM and IEEE.

Carlo Ghezzi, Politecnico di Milano
Software processes: a retrospective and a path to the future

Abstract: Software engineering focuses on producing quality products through quality processes. The attention to processes dates back to the early 70's, when software engineers realized that the desired qualities (such as reliability, efficiency, evolvability, ease of use, etc.) could only be injected into the products by following a disciplined process that would also make the production process predictable and economical. Most of the software process work, however, remained in an informal stage until the late 80's. From then on, a new era started, in which the software process was recognized by researchers as a subject that needed to be scrutinized on its own in a scientific manner, to understand its foundations, develop models, identify methods, provide tool support, and help menage its progress.

The presentation will try to characterize the main approaches to software processes that were followed historically by software engineering, to identify the strenghts and weaknesses, the motivations and the misconceptions that lead to the continuous evolution of the field. This will lead us to an understanding of where we are now and will be the basis for a discussion of a research agenda for the future. Based on the author's personal experience and bias, the cornerstones on which this agenda can be defined, are:

1. The software process is a cooperative design activity. This involves supporting human interactions and negotiations, recording design decisions, managing deviations from the defined processes, acting in the presence of inconsistencies, etc.. Activity automation and process mechanization are important, but they should not be over-emphasized.

2. The infrastructure providing process support must enable distribution and mobility of users.

3. Process support must include support to specific software development methods. Such methods will be incresingly based on semantically precise notations.

Biography: Carlo Ghezzi is a Professor of Software Engineering in the Department of Electronics and Information of Politecnico di Milano, where he served as Department Chair. He received his Dr.Eng. degree in Electrical Engineering from Politecnico di Milano, where he spent most of his professional life. He taught at the Universities of Padova (Italy) and North Carolina at Chapel Hill (USA). He spent sabbatical periods in the USA at UCLA and UCSB. He was a Guest Professor at The Escuela Superior Latino-Americana de Informatica (ESLAI), Argentina, in 1990, at the University of Klagenfurt and at the Tecnical University of Vienna, Austria, in 1996. Presently, he is a Guest Professor at the Universita` della Svizzera Italiana, Lugano, Switzerland, 1997-1998.

Ghezzi's research interests are in software engineering and programming languages. He is currently particularly interested in application the theoretical, methodological, and technological issues involved in developing network applications. He is a co-author of over 120 scientific papers and 8 books. Among these, Programming Language Concepts (co-author M. Jazayeri), J. Wiley and Sons, New York, NY, 3rd Edition, 1997; Fundamentals of Software Engineering (co-authors M. Jazayeri and D. Mandrioli), Prentice Hall, Englewood Cliffs, NJ, 1991; and Theoretical Foundations of Computer Science (co-author D. Mandrioli), J. Wiley and Sons, New York, NY, 1987.

Ghezzi was the Program Chair of the 2nd European Software Engineering Conference (ESEC-2), Program Co-Chair of the 6th IEEE Workshop on Software Specification and Design, Program Co-chair of the IEEE 14th Int.l Conference on Software Engineering (ICSE-14), and Program Chair of the 9th Int.l Software Process Workshop (ISPW-9). He will be General Chair of ICSE in the year 2000. He is a member of the board of directors of ISPA. He is a member of the editorial board of IEEE Transactions on Software Engineering, Trends in Software, Software Process Improvement and Practice, Theory and Practice of Object Systems.

Abstract: Workflow systems are used today to run mission critical applications in many enterprises. Consequently, the corresponding systems must be robust, scaleable and highly available, and an appropriate application development environment is needed. We will sketch the exploitation of database and transaction technologies to achieve the required system properties, and outline the blueprint of a development environment for workflow applications. Vice versa, we will show that workflow technology provides the proper base for transaction handling in multi-database environments, and that it can be used for enforcing dynamic integrity rules. Furthermore, workflow technology is the base for a new software structure paradigm which might become as influential as the client/ server paradigm.

Biography: Dr. Frank Leymann is an IBM Senior Technical Staff Member. Since joining IBM in 1984 he contributed to a variety of product development projects from areas like repositories, database systems, tools for application development and database tuning, and workflow systems. He is the lead architect of FlowMark, IBM's production workflow system. Frank published papers in various journals and conference proceedings, filed a multitude of patents, and works as a technology consultant for large customers from all over the world.

Abstract: Collaboration takes place whenever humans and/or computer applications work together to accomplish common/compatible goals. Collaboration Management Infrastructure (CMI) is required to promote, support, and manage collaboration. In this talk we propose novel CMI technology based on a comprehensive process- and service-oriented solutions. We discuss specific CMI applications that are currently being developed by the industry and government, and illustrate that such applications present key collaboration process and corresponding infrastructure requirements that cannot be fully supported by existing workflow, groupware, and business process (re)engineering technology.

To address the collaboration process management requirements of these driving applications, CMI must develop new process models and technology for collaboration coordination and awareness. Coordination involves controlling the order of specific activities performed by humans and applications that participate in a collaboration, and the use and movement of resources in a collaboration. The CMI coordination model provides for both synchronous and asynchronous activities, allows partial process specification, permits dynamic process change and extension, and supports process templates and context resources. Awareness involves communication of collaboration-related information and events among the collaboration participants, and composition of such events to provide higher-level of abstractions as needed to achieve specific collaboration goals. The CMI awareness model supports process monitoring, allows participants to subscribe to specific process events and automatically construct a participant-specific view/abstraction of the process.

Services are reusable components that advertise and implement service agreements. In the CMI service model, services are reusable process model objects, such as process resources, activities, and (sub)processes that generate awareness information and events. Therefore, CMI services can be coordinated and monitored as normal process resources or activities. However, service lifecycle management and advertisement are independent of any specific process. Process instances that require a particular service must discover it and request it. CMI services define service agreements that include descriptions of service levels and corresponding quality. In addition, they provide operations for requesting and monitoring specific level(s) of service quality. The CMI service model can support composition of virtual services in virtual enterprise environments.

CMI combines the coordination, awareness, and service models in a comprehensive core plus extensions model. In particular, the coordination, awareness, and service models are provided as extensions of a common core that ensures sufficient interoperability between them. This allows selective use of any combination of these models (as needed by each driving application), interoperability with models used in commercial workflow systems and groupware tools, and further extensibility to support new requirements, such as mobility and migration. A CMI proof of concept prototype currently being developed at MCC has a loosely-coupled, component-based architecture, and uses commercial workflow and groupware components to provide existing capabilities required by the CMI model.

Biography: Dimitrios Georgakopoulos is currently a Senior Member of Technical Staff and the technical leader of the Collaboration Management Infrastructure Project (CMI) at MCC. Before coming to MCC, Dimitrios was a Principal Member of Technical Staff at GTE Laboratories and the initiator of the Workflow Management Infrastructure project. He led the development of the Word Wide Workflow system for enterprise-wide and multi-organizational business process management. Before that, he also worked on distributed object computing and advanced transaction processing. At At GTE, Dimitrios provided consulting support to various GTE business units. He played a principal role in introducing workflow and distributed object technologies to GTE business units, and received an Excellence award for his work. Prior to GTE, Dimitrios worked at Bellcore in the area of multidatabase systems.

D. Georgakopoulos received the B.S. degree in mathematics from Aristotle University, Thessaloniki, Greece, in 1982, and the M.S. and Ph.D. degrees from the University of Houston, Houston, Texas, in 1986 and 1990, respectively. Dimitrios is the Program Co-Chair of the International Joint Conference on Work Activities Coordination (WACC'99), and the Program Chair of the Ninth International Workshop on Research Issues in Data Engineering: Information Technology for Virtual Enterprises (RIDE'99). Dr. Georgakopoulos has received two IEEE Computer Society Outstanding Paper Awards in 1991 and 1994, respectively. He and his Transaction Specification and Management Environment project at GTE had been nominated for the 1994 Computerworld Smithsonian Award in Science.