Specification of behavioral semantics in object-oriented information modeling

Abstract

Foreword This workshop report is organized in five sections: 1) purpose of workshop, 2) logistics, 3) technical presentations, 4) recommendations, and 5) list of participants. To understand an enterprise, make its components reusable and semantically interoperable, precise specifications of behavioral semantics are essential. Further, as objects do not exist in isolation, modeling an enterprise involves modeling the collective behavior of objects that make up the enterprise. The purpose of the workshop was to explore behavioral modeling concepts, especially modeling the collective behavior of objects (using declarative constructs), and behavioral abstraction and refinement approaches for aggregation and decomposition. Some topics of particular interest included formal specification of behavioral semantics and (attempts at) the standardization of information modeling concepts and the specifications of reusable components. The call for participation welcomed contributions by both researchers and practitioners, as we hoped to achieve a cross-fertilization of formal and heuristic/informal specification approaches, and an outline of open practical and theoretical questions that would advance the state of the art and the practice. We received a total of thirty submissions varying in length from 2 pages to full length papers. Those submissions were collected-and some manually edited and typeset!-in a 169 page proceedings edited by Haim Kilov and Bill Harvey, produced by Bill Harvey's home institution, proceedings were sent to the participants two weeks prior to the workshop. Fifteen submissions were selected for presentation during the workshop, under two major tracks: 1) Issues, and 2) Solutions. Under solutions, three loosely recurring themes were identified: 1) modeling of collective behavior, 2) the use of formal methods, and 3) types/roles and operations/events. After some last minute changes, we (28 people at the beginning, a bit less 7 hours later) sat through 12 presentations averaging 10 min each, punctuated with two coffee breaks (did anyone find out where coffee was served, if any?), a late-lunch break, and two " thematic " open discussions. There then followed a no holds barred, two-hour cross-fire, by the end of which we all learned how to pronounce each other's names correctly. This workshop builds on a workshop titled " Object-Oriented Reasoning in Information Modeling, " organized by Haim Kilov and Bill Harvey in OOPSLA'92 [Kilov & Harvey, 19921. Some of the participants had already participated in the '92 workshop. All of the participants this year expressed the desire to " do this again. " 3. Technical Discussions This section is divided into three …