7. Concluding Remarks

14 However, the concept of (central) state is not always adequate, for instance if the system is truly distributed, i.e., without some central coordination. Here we come to the limits of our object model: as it is, it does not capture truly distributed cases. It is good practice to identify large portions of the system where a central state makes sense, for instance the sites of the distributed system, so here we can use our approach. Giving a logics and semantics for an entire truly distributed system, however, would require to substitute our process model by another one, involving true concurrency and distributed states. Petri nets may be a good idea. We are conndent that it is possible to substitute other process models into our approach. The conceptual and theoretical work presented in this paper is part of a coordinated eeort which also comprises practical work, developing object speciication languages OBLOG SSE87, CSS89] and TROLL JSHS91]. Work on prototype implementations is in progress. The E.S.D.I. company in Lisbon is developing OBLOG into a commercial product. Languages and tools for an object system design and speciication approach should be based on a systematic methodology CY91, Gr91, Lo93]. Work on methodology was not in the focus of our activities so far, but we expect that it will become more important in the future. Another promising line of research is to incorporate \knowledge" into our approach. Our situation concept (sets of facts) is open for generalization to other kinds of formulas , e.g., deduction rules. Also, default reasoning is becoming very interesting Br92]. We expect that object{oriented and knowledge{based approaches can be integrated along our lines. Acknowledgements Thanks to all IS-CORE colleagues who contributed to the development of ideas presented here. In particular, Cristina Sernadas, Gunter Saake and Ralf Jungclaus participated in discussing the basic ideas of objects and object descriptions. Felix Costa's contributions to object semantics as well as Jose Fiadeiro's and Tom Maibaum's contributions to object logic are gratefully acknowledged. We have taken some inspiration from Gunter Saake's Habilitationsschrift Sa92] and Ralf Jungclaus' Doktorarbeit Ju93]. 13 6 Modularization It is commonplace that modularization is of paramount importance to software construction and reconstruction. The object concept itself is a sort of modularization principle, but a rather in-the-small one. For eeective software reuse, we need an in-the-large concept which makes it possible to put modules into a library, nd the ones we need …