Implementing a Query Language for Context-Dependent Semistructured Data

In today’s global environment, the structure and presentation of information may depend on the underlying context of the user. To address this issue, in previous work we have proposed multidimensional semistructured data (MSSD), where an information entity can have alternative variants, or facets, each holding under some world, and MOEM, a data model suitable for representing MSSD. In this paper we briefly present MQL, a query language for MSSD that supports context-driven queries, and we attempt to motivate the direct use of context in data models and query languages by comparing MOEM and MQL with equivalent, context-unaware forms of representing and querying information. Specifically, we implemented an evaluation process for MQL during which MQL queries are translated to equivalent Lorel queries, and MOEM databases are transformed to corresponding OEM databases. The comparison between the two query languages and data models demonstrates the benefits of treating context as first-class citizen. We illustrate this query translation process using a cross-world MQL query, which has no direct counterpart in contextunaware query languages and data models.