Document Type


Publication Date



Abstract Large-scale information systems typically contain autonomous information sources (ISs) that dy- namically modify their content interfaces as well as their query services regardless of the data ware- houses (views) that are built on top of them Current view technology fails to provide adaptation techniques for such changes giving support to only static views in the sense that views become unde- fined when ISs undergo capability changes We propose to address this new view evolution problem - which we call view synchronization - by allowing view definitions to be dynamically evolved when they become undefined The foundations of our approach to view synchronization include the Evolvable- SQL view definition language (E-SQL) the model for information source description (MISD) and the concept of legal view rewritings In this paper we now introduce the concept of the strongest synch-equivalent view definition that explicitly defines the evolution semantics associated with an E-SQL view definition Plus we propose a strategy and proofs of correctness for transforming any user-specified E-SQL view definition into the strongest E-SQL query We also present the Complex View Synchronization (CVS) algorithm that fully exploits the constraints defined in MISD by al- lowing relation substitution to be done by a sequence of joins among candidate relations Examples illustrating this multi-step approach are given throughout the paper