Description

A wide range of database applications manage time-varying, or temporal, data. These include financial applications such as portfolio management, accounting, and banking; record-keeping applications, including personnel, medical-record, and inventory; and travel applications such as airline, train, and hotel reservations and schedule management.

The proposed research activities are based on two quite different observations. First, in order to exploit effectively a temporal database management system, its data model must have an associated design methodology. Second, independently of whether a temporal DBMS is available, providing built-in temporal support in the data model used for conceptual database design yields simpler conceptual schemas that better capture the modeled reality. We feel that there is a lack of a fully satisfactory, user-friendly design methodology for the design of databases with temporal data.

It is generally accepted that database design is to proceed in two steps. In the first, the reality to be modeled is described using a conceptual data model. A conceptual data model has a rich variety of modeling constructs and allows the designer to capture the important aspects of the modeled reality in a conceptual schema that is easy to understand. In the second step, the conceptual schema is mapped to a database schema of the data model of the available database system. With this division, the first part of database design is independent of the underlying database system–it is possible to map a conceptual schema to a wide variety of implementation data models. In practice, some variation of the ER model is typically used as the conceptual model, and the implementation data model is the relation model (e.g., some vendor-dependent variation of the SQL-89 or SQL-92 standards).

Preliminary studies have indicated that if time is made an fundamental part of the enhanced ER model, this leads to a simplification of the designed (temporal) ER schemas. We expect these simplified temporal ER schemas to be much easier to grasp and then also easier to maintain than traditional ER schemas.

We consequently propose to develop a temporally enhanced ER model, to be TIMEER, that, in building on and addressing identified shortcomings of the existing proposals, provides better means for designing databases containing time-varying data.

Transformations of TIMEER schemas to logical schemas on various particular implementation platforms, perhaps including SQL-92, TSQL2, and ATSQL2, will be developed.

The resulting design methodology supports the convenient capture of the temporal aspects of data. It is maximally independent of the available implementation platform for the database. This provides for platform independence and reuse of designs. And the methodology aims to fully exploit the functionality available on the given implementation platform.

Further readings:

	H. Gregersen and C. S. Jensen, Temporal Entity-Relationship Models–a Survey, [.ps.gz]. A revised version is accepted for publication in IEEE Transactions on Knowledge and Data Engineering.
	H. Gregersen, C. S. Jensen and Leo Mark, Evaluating Temporally Extended ER Models Published in the proceedings from the The Second CAiSE*97/IFIP8.1 International Workshop on Evaluation of Modeling Methods in Systems Analysis and Design. Barcelona, Spain, June 1997.
	H. Gregersen and C. S. Jensen, Conceptual Modeling of Time-Varying Information [.ps.gz].
	H. Gregersen, L. Mark, and C. S. Jensen, From Temporal ER models To Relations. Submitted for publication.