Abstract
A ‘cubic’ view of database with representation by time-stamping of tuples has been largely favoured so far in the research on modelling of time in database systems. In this paper, we propose a ‘state’ oriented view of historical databases. The salient features of our proposal are its simplicity, closeness to the consistency with classical relational model, and efficient implementability. We propose an algebra for historical which contains classical as well as some new operators. The operators are simple to comprehend, unlike in other research proposals. We are also able to formulate a completeness criteria for the proposed model. Finally, we extend the popular SQL query language for use with historical databases. Again, the extensions are consistent with the simple basis of standard SQL. They are minimal in number, yet very powerful and expressive. We illustrate algebra operators and extended SQL with many examples.