------------------------------------------------------------------------------ ****************************************************************************** ------------------------------------------------------------------------------ B O S T O N U N I V E R S I T Y Computer Science Department C O L L O Q U I U M Wednesday, March 19, 1997 3:00 pm (Coffee served at 2:30 pm, Room MCS 137) Room MCS 135 ------------------------------------------------------------------------------ ****************************************************************************** ------------------------------------------------------------------------------ On the Efficient Maintenance of Temporal Integrity in Knowledge Bases Dimitris Plexousakis Kansas State University http://www.cis.ksu.edu/~dimitris/home.html The maintenance of semantic integrity has been recognized as a cornerstone issue for the development of databases and knowledge bases alike. Despite the extensive research conducted during the last two decades, semantic integrity maintenance has yet to become a practical technology. Furthermore, the need for modeling evolving domains has given rise to challenging research issues relating to the incorporation of time in knowledge bases. In this talk we address the problem of maintaining the integrity of temporal deductive knowledge bases. We argue that existing approaches in either temporal or deductive databases do not address the problem in a satisfactory manner nor do they deal with all the issues involved in a unified framework. We propose an assertion language that permits us to express different types of temporal assertions that are not expressible in other formalisms. We define the notion of temporal constraint satisfaction in a bitemporal context. We follow two orthogonal approaches to integrity maintenance, both using a combination of compile-time and run-time optimization steps. The former, termed temporal integrity monitoring, operates in two phases: compilation and evaluation. The result of the compilation phase is the derivation of simplified forms of constraints that suffice to be evaluated at run-time. The evaluation phase performs additional simplifications with respect to the transactions taking place at run-time by combining the previously generated simplified forms. The latter of the proposals, termed transaction modification, delegates the task of integrity maintenance to determinate transactions specified in terms of precondition/postcondition pairs. It suggests additions to the transaction's postcondition whose effect is to maintain the constraints. The proposed method establishes the theoretical foundations for the development of a tool that assists the database designer by providing valuable feedback concerning the safety of transactions. Both proposals demonstrate that considerable savings can be incurred by performing compile-time simplifications. We argue that this is a viable approach towards making constraint maintenance a practical technology in time-critical applications. Host: Prof. Azer Bestavros ------------------------------------------------------------------------------ For colloquium info, including directions, see http://www.cs.bu.edu/colloquium For more information contact Prof. David Yates ------------------------------------------------------------------------------