Pavlo Zhezhnych Personal Site and Blog

The thesis is dedicated to problems of temporal parameter modeling in information systems based on relational databases. Scientists investigate a temporal database area during more than 20 years and proposed more than 40 different relational and object-oriented temporal data models. These models were not realized in popular commercial DBMSs due to their complexity. The most research directions of the thesis are concentrated on qualitative temporal relational data model building and its implementation with relational data model.
At the beginning problems of time-depended information storing in databases were considered. We emphasize three requirements that should be followed during data domain modeling: temporal completeness, temporal density and temporal isomorphism. Time-depended information was classified so whole data domain is divided on object. Data domain facts are bounded with objects and events referring to these objects. Each object has characteristics, which can be changed with events. An object information state is used to calculate values of these characteristics. An information state has to satisfy following requirements: historical consecution and historical density.
In the second chapter we propose temporal relational data model based on a notion of temporal relation. The temporal relation was defined as Cartesian product of attributes’ and time parameter’s domains. Beside the abstract temporal relations informational relations are considered. To know object characteristics’ values at a time moment a relation state has to be calculated. To choose a set of events referring to the object and happened at the time moment a relation slice has to be calculated. Analyzing informational relations we classified temporal functional dependencies: state dependencies, slice dependencies and dependencies from time parameter. Following these dependencies’ definitions new formulations of first, second, third and Boyce-Codd normal forms were proposed. Also object-temporal and synchronized normal forms were introduced. Using state and slice operations the conventional relational algebra was extended to temporal data model requirements. We considered straight, static and temporal extensions of relational operators. At the end of second chapter the temporal relational data model was generalized up to any number of temporal parameters. For generalized relations two new operations were introduced: temporal state and temporal slice. Functional dependencies based on whole set of temporal parameters and its subset are considered. So we propose temporal normal forms using whole set of temporal parameters and its subset as well.
In the third chapter an object-temporal methodic of information system design was considered. The methodic is based on a concept of transformed temporal relation. A transformed relation is a convenient relation but with temporal interpretation. A transformed relation has state and slice, temporal state and temporal slice. A transformed relation state can be calculated using several rules. For this reason we introduce class of state calculation aggregation rules and implemented it with relational algebra operators. Also operators of temporal relational algebra were implemented. The second part of the chapter was dedicated to information system architecture design. We distinguish three levels of the architecture: a conceptual level (an abstract model level), a logical level (a user level) and a physical level (a data store level). For temporal databases the conceptual model represents a general schema of data domain objects and relationships between them. The logical level is an information schema of database. The physical level is a database stored on data medium.
In the fourth chapter we described an information system “Electronic museum” designed according to object-temporal methodic. At a conceptual level the information system description includes functional structure (represented with Data Flow Diagrams), classes hierarchy and conceptual data schemas (represented with Entity-Relationship Diagrams). At a logical level database schema was given. It describes logical structure of tables and views have to be stored in the database. The system was implemented with DBMS Microsoft Access 2000 using SQL (Structured Query Language) and VBA (Visual Basic for Applications). It works at Lviv Historical Museum and proves the scientific and practical results of the thesis to be true.