135

135_Poels_Geert.htm

Moving from OOAD towards COAD

Think Tank 75 minutes

Do we need Component-Oriented Analysis & Design methods ?

Geert Poels

Guido Dedene

Recent years saw the emergence of a new paradigm in software development, i.e., the component-oriented & component-based paradigm. The move towards components has largely been industry-driven. At the beginning of the new millenium technology has progressed so far that there are no technical hindrances to a widespread use of componentware worth mentioning. The world of components seems to be firmly grounded on technologies and standards (e.g., CORBA, DCOM and JavaBeans), on component frameworks (e.g., Boca (OMG) and the SAP R/3 Business Framework), on COTS, etc. However, methodology hasn't caughted up on the new trend. Although some work is observed in the field of component-based software development/engineering (CBSD/E), there isn't much methodological support for component-oriented software development/engineering (COSD/E), apart from the work on component-oriented programming (COP). Has anyone heard from component-oriented analysis & design (COAD) yet? We haven't.

The goal of this think-tank session is to discuss the need for a rigorous approach towards the development of software components and to identify requirements for a COAD method. Our starting point are the current OOAD methods, as we see Object-Orientation as a sine qua non for Component-Orientation. However, is Object-Orientation alone enough? What else is needed to transform the current OOAD methods into COAD methods? What specific methodological support is currently missing in our approaches towards Component-Orientation?

The think-tank session tries to answer questions of this kind and gives participants a forum to express their needs and requirements for component-oriented methodology. It is time that researchers take the new paradigm seriously!

If you would like to come to this session, you may find it useful to read this paper first: Do we need Component-Oriented Analysis & Design methods?

Geert Poels (geert.poels@econ.kuleuven.ac.be)

Catholic University of Leuven, Dept. Applied Economic

Dr. Poels is a Postdoctoral Fellow of the Fund for Scientific Research - Flanders, working with the information systems group of the Dept. of Applied Economics at the Catholic University of Leuven. His research interests include object-oriented software measurement and quantitative approaches in component-oriented software engineering. Dr. Poels is Program Chair of the 2nd European Software Measurement Conference (Amsterdam, 6-8 October 1999).

Guido Dedene (guido.dedene@econ.kuleuven.ac.be)

Catholic University of Leuven, Dept. Applied Economic

Dr. Dedene is a Professor of management information systems at the Dept. of Applied Economics, Catholic University of Leuven. His lecturing and research activities concentrate on formal methods for systems development and quantitative systems management techniques. He is co-author of the book "Object-Oriented Enterprise Modelling with MERODE", published in the summer of 1999.

Topics

Benefits

Participants will learn about the differences between Object-Orientation and Component-Orientation (and the differences between component-based development and component-oriented development). They will also learn about the sorry state of the contemporary methodological support for COSD/E. But more important, they will have the opportunity to express their specific needs and requirements regarding such methodological support. The session is not a tutorial but an interactive event in which we, as researchers, might also learn something from the participants.

Session: Think Tank 75 minutes Level: intermediate
Audience: The session is intended for all practitioners planning/making/completing the move from Object-Orientation towards Component-Orientation, or all practitioners otherwise interested in the component paradigm. We equally welcome all researchers (academia and others) struggling with the same questions as we have. We expect practitioners to have a basic knowledge of objects and components, just enough to appreciate the differences between both worlds. We do not require participants to be experts in (specific) OOAD methods. Max 15

Material

The material presented during the introduction will be made available to the participants. This will include examples in which the requirements of the OO and CO worlds seem to be conflicting. We also prepare a list of questions to gauge the reactions and sentiments of the participants. This list should be made available on the web page of the session.

Delivery

We plan to produce a poster that can act as a sort of position paper for the think-tank participants. We hope this poster will express what practitioners expect from a COAD method or the COSD/E methodology in general. At the very least the poster should answer the question raised by the session leaders: do we need component-oriented analysis & design methods?

We further plan to circulate the position paper among researchers (academic and others) working on the component-oriented paradigm, and to submit it to workshops dealing with COSD/E.

Format

The first part necessarily includes a bit of lecturing to clarify the issues raised. The second part can be much more informal, although we will initially structure the discussion using a list of questions prepared in advance.

135

135_Poels_Geert.htm