The problem of providing database systems customised to the specific needs of an organisation or application is addressed. The authors demonstrate that such customisations are expensive in existing database systems owing to the crosscutting nature of customisable features, and they propose a customisation approach based on `aspect-oriented programming' techniques which allow separation of crosscutting features using special constructs known as `aspects'. Changes to the features encapsulated by aspects are localised, making cost-effective customisation possible (at compile-time and run-time) at both the DBMS level and the database level.
This paper reports pioneering work on the application of aspect-orientation to database systems. The work reported represents a key stage in the evolution of Lancaster's leadership in AOSD. It emerged from the Rashid and Sawyer's earlier work on O-O database schema evolution and fed (by drawing on Sawyer's work on RE) into Rashid's work on early aspects. To date the paper has been cited 21 times. RAE_import_type : Journal article RAE_uoa_type : Computer Science and Informatics