Object-oriented Data Modelling

A declarative model of OOP which supports basic elements of OOP programming and allows program synthesis from specifications and program transformation is developed. Optimisation methods applicable within the deductive OOP model are developed on the basis of the definition of partial deduction (PD) for structural synthesis of programs. The declarative OOP model is used for modeling dynamic aspects of OO systems (databases, IS) as well as for expressing general knowledge about application domain [1].

A methodology for specification of semantics of evolution in object-oriented databases (OODB) is proposed. It is based on Horn logic as meta-language for specification of semantics of both schema and object evolution in OODB. Partial Deduction is used as technique for specialisation of a set of general constraints on database and its schema and to derive conditions that must be satisfied to guarantee the validity of managing evolution in OODB. Implementation principles of the methodology are based on the object-oriented language NUT. Different types of transformations are defined that allow derivation of predicates from the descriptions of classes and objects used in the NUT system [2, 3].

The main contribution of this research is the application of Galois approach to dynamic classification of objects in OODB. First, we defined an object model as the Galois lattice of binary relationship between objects and classes they belong to. Second, we have shown that an object classifier can be built on the basis of construction of such a model. We argued that such an object classifier can be easily used for building tools for visualisation and querying of meta-information concerning context of object base [4].

In this research we consider automatic discovery of OO schema and schema conflicts. We propose an approach which is intended to detect conflicts between schema and underlying object base. Our approach is distinguishable by first discovering schema from given set of objects and then to comparing it to the existing schema for discovery of schema conflicts. The approach is based on the comparison of the two inheritance lattices: one constructed on the basis of class library and another discovered from object base using one of the general conceptual classification methods.

The particular goal of the research on object-oriented persistence is to construct a conceptual framework for persistence of objects and design persistence manager for the NUT system. Up to now, the NUT system is lacking support for persistent objects. We take an approach to extend this programming system with persistent store of objects. It is nontrivial task to construct a consistent model of persistence for NUT given the wide variety of functions for working with objects in Object Memory, but not having generally solved problem of Object Identity.

An object-oriented interface to the Informix-OnLine database system is developed. The guiding philosophy underlying the interface is to support code reuse for persistent variables. Unlike the usual object-oriented interfaces, the data model mappings for relational -to-object and object-to-relational have been separated. This resulted to very flexible interface.

Matskin M, and Haav H.-M., A Deductive Object-Oriented Approach to Information System Modeling, In: J. Eder, L.A. Kalinichenko (Eds) Advances in Databases and Information Systems, Workshops In Computing Series, Springer, 1996, pp 459-479
Haav H-M, Specifying Semantics of Evolution in Object-Oriented Databases Using Partial Deduction, In: U. W. Lipeck, B. Thalheim (eds.), Modelling Database Dynamics. Selected Papers from the Fourth International Workshop on Foundations of Models and Languages for Data and Objects, Volkse, Germany, Springer-Verlag, 1993 pp 48-63
Haav H-M, A Meta-language for Specification of Evolving Class and Object Lattices in Object-oriented Databases, In. J. Eder and L.A. Kalinitchenko (Eds) East/West Database Workshop, Proc. of the Second International East-West Database Workshop, Klagenfurt,1994 , Workshops In Computing Series, Springer-Verlag 1995, pp 147-159
Haav H-M, An Object Classifier Based on Galois Approach, In: H. Kangassalo, J. F. Nilsson, H. Jaakkola, S. Ohsuga, Information Modelling and Knowledge Bases VIII, IOS Press, 1997, pp 309-321