Abstract Programmers are increasingly interested in developing applications that can be used internationally. Part of the internationalization effort is the ability to engineer applications to use dates and times that conform to local calendars yet can inter‐operate with dates and times in other calendars, for instance between the Gregorian and Islamic calendars. $\tau$ ZAMAN is a system that provides a natural language‐ and calendar‐independent framework for integrating multiple calendars. $\tau$ ZAMAN performs ‘runtime‐binding’ of calendars and language support. A running $\tau$ ZAMAN system dynamically loads calendars and language support tables from XML‐formatted files. Loading a calendar integrates it with other, already loaded calendars, enabling users of $\tau$ ZAMAN to add, compare, and convert times between multiple calendars. $\tau$ ZAMAN also provides a flexible, calendar‐independent framework for parsing temporal literals. Literals can be input and output in XML or plain text, using user‐defined formats, and in different languages and character sets. Finally, $\tau$ ZAMAN is a client/server system, enabling shared access to calendar servers spread throughout the Web. This paper describes the architecture of $\tau$ ZAMAN and experimentally quantifies the cost of using a calendar server to translate and manipulate dates. Copyright © 2006 John Wiley & Sons, Ltd.

Aggregate computation, such as selecting the minimum attribute value of a relation, is expensive, especially in a temporal database. We describe the basic techniques behind computing aggregates in conventional databases and show that these techniques are not efficient when applied to temporal databases. We examine the problem of computing constant intervals (intervals of time for which the aggregate value is constant) used for temporal grouping. We introduce two new algorithms for computing temporal aggregates: the aggregation tree and the k-ordered aggregation tree. An empirical comparison demonstrates that the choice of algorithm depends in part on the amount of memory available, the number of tuples in the underlying relation, and the degree to which the tuples are ordered. This study shows that the simplest strategy is to first sort the underlying relation, then apply the k-ordered aggregation tree algorithm with k=1.>