Form of a Set of Time Points. This Set Can Used as a Domain 4 Comparison with Other Ongoing Research

for time point variables, such as the variable t above. This approach of using time points is more declarative than ours because it hides the implementation detail of using a set of pairs of time intervals and values to implement the partial function from time points to values. It is not clear, however , whether this higher level of abstraction leaves more opportunities for optimization or whether it is more user-friendly. We believe that their approach is not well-suited to version control, since version control inherently requires to associate one time interval with a version. Another temporal OODB model is proposed by Cheng and Gadia [3]. Their temporal OODB language, called TempSQL, is based on temporal expressions, which map boolean predicates to temporal domains. For example, the temporal expression e.salary>15K]] returns the time period (lifespan) in which the salary of the employee e was greater than 15K. Such temporal expressions have also been used in the temporal query language GORDAS proposed by Elmasri and Wuu for the extended Entity-Relationship model [6]. 5 Conclusion We have presented an extension to the ODMG-93 model that incorporates a time dimension to data. In contrast to other models, time dimension in our model is an orthogonal property of data types. We have also presented an extension to OQL to accommodate time information. These language extensions are minimal, uniform, and easy to understand. As a future work, we are planning to extend our model with parameterized types for representing transaction time and bi-temporal queries, and for representing time-series data and calendars. In addition, we are planning to design an optimization framework to evaluate the TOQL language constructs efficiently using currently proposed temporal evaluation techniques. We are also planning to work on the specification of the temporal integrity constraints and their use on semantic query optimization. parameterized type Temporal h T i f Structure f Time Interval valid time, Integer version, T value g Temporal value; attribute Listh Temporal value i history; attribute Integer size; attribute Temporal value latest version; Temporal value version projection (in Short version) raises (no such version); Temporal value time projection (in Time time) raises (not deened at that time); Temporalh T i interval projection (in Time Interval i); new version (in T value, in Time event); close version (in Time event); g; The TOQL syntactic constructs are translated into the following OQL expressions: time]) struct(start: time; end: time) time 1 ; …