Uma abordagem baseada em contexto para reescrita de consultas a bancos de dados relacionais

Abstract

With the development of Internet technologies and their large scale usage, database queries have evolved from situations where users had a known profile and submitted their queries on desktop computers through dedicated applications to situations with decentralized and massive information distribution. However, there are a variety of factors which may influence the appropriateness and relevance of the answers provided to users' queries, such as: personal preferences, location (e.g. home, work, hotel, plane), climate (e.g. rainy, sunny), and device used. These factors are not usually taken into consideration when submitting queries to databases. This set of relevant conditions and influences that make a situation unique and understandable is called context. Contextual elements are defined as any data, information or knowledge that allows us to characterize an entity in a domain. The set of instantiated contextual elements needed to support a task at a given time denotes the context. Although database management systems (DBMS) are a mature, widely known technology and are considered to be the most efficient systems to manage and respond to structured data queries, they currently lack the resources needed to meet the demand for information that is adapted to the current context. This work proposes an approach named Texere, that endows conventional relational database queries with context sensitivity. The solution is based on the analysis of contextual elements associated with a query (arising from the application, device, user and physical and computational environments) and the use of directives for query rewriting able to alter queries devoid of context awareness to others whose answers are more appropriate to the context in which they were made and thus are more relevant to the user. For the evaluation of this work, experiments were performed with a prototype. The results produced by the users’ evaluation points to the feasibility of the approach and gains in adaptation to the context and relevance of the answers produced by the rewritten queries.