DEVELOPMENT OF SEMANTICALLY INTEROPERABLE INFORMATION SYSTEMS

Leonid Kalinichenko, Russian Academy of Sciences

Abstract:

The gap between existing Object Analysis and Design (OAD) methods applying mostly top-down technique and the demand of the middleware interoperation architectures and methodologies for the development based on a composition of interoperating components remains to be large.

A number of various computational, data and knowledge models based on an object paradigm is continuously increasing. These models are used for development of software and data services, information systems and their subsystems that technically can easily become components of the middleware. Such heterogeneity and lack of well-defined semantics of the respected models creates a big obstacle for their interoperability.

But probably the largest obstacle for the interoperability of components consists in the application semantics of components technically interrelated through the middleware. Reconciliation of their application concept base that is an obvious prerequisite for their interoperation constitutes a problem.

The approaches to fill in the gaps mentioned are briefly overviewed. We strictly distinguish between application semantics and object model semantics.

Using the "canonical" model we believe that terms widely used in various object models need to be carefully defined to give them precise semantics. To give the canonical model exact meaning, we construct a mapping of this object model into the B AMN notation providing precise meaning for the language. Thus, we get the semi -formal object model and its formal counterpart that we can use together as a common paradigm for:

- uniform representation of various object models;

- uniform specification of pre-existing components;

- different models used on the phases of requirement planning, analysis and design of the information systems.

The concept of refinement of the specifications relying on the pre-existing components becomes inherent for the model.

An approach of mapping of existing object models into the canonical one is discussed.

Application semantics of components we consider separately in frame of the ontological approach. Semi-formal and formal (model- based) specifications for concepts are provided. We base the ontological model we introduce on the canonical object model.

Component-based information design issues in the semantically interoperable environment are briefly discussed.

Type specifications and their reducts are chosen as the basic units of specification manipulation. The 'algebra' of type specifications is introduced for that. Reducts of the component type specifications are used as minimal fragments potentially reusable for the respected reducts of the analysis model types. Process of design of the semantically interoperable information systems is discussed.


Short resume of Leonid A. Kalinichenko

Leonid A. Kalinichenko received Ph.D. degree from the Institute of Cybernetics, Kiev, Ukraine, in 1968 and the degree of Doctor of Sciences from the Moscow State University in 1985.

He was a staff scientist at the Institute of Cybernetics of the Ukrainian Academy of Sciences (Kiev) until 1969. Since 1969 to 1985 he chaired the department of information systems at the Institute of Control Computers in Moscow. Since 1985 he is a chair of the department for the systems for knowledge integration at the Institute for Problems of Informatics of the Russian Academy of Sciences in Moscow. Since 1977 he teaches at the Moscow University (Computer science department) including seminars on data models, multidatabase management, heterogeneous database integration, object technologies. Got degree of professor in 1990.

His current research interests include: interoperable heterogeneous information resource management, semantic interoperability, deductive object-oriented databases, object models, object analysis and design methods. Currently he leads the SYNTHESIS research project covering various aspects of semantically interoperable information systems specification and development.

He is author of the book "Methods and tools for heterogeneous database integration" (Moscow, 1983), and co-author of the four books : "SLANG - programming system for discrete event system simulation" (Kiev, 1969), "Computers with advanced interpretation systems" (Kiev, 1970), "Computer Networks (Moscow, 1977), "Database and Knowledge base Machines" (Moscow, 1990).

L.A. Kalinichenko is a member of the ACM and of the IEEE CS, the Chairman of the Moscow ACM Sigmod Chapter, a Member of the Editorial Board of an International Journal "Distributed and Parallel Databases", Kluwer Academic Publishers.