Expression language for search computing queries

Search Computing queries make use of possibly complex expressions to represent join conditions, selection predicates and ranking criteria over the data that is returned by the Web services. Apart from basic constructs such as conjunctions, disjunctions, relational, arithmetic and set operators, etc., expressions need to support the full expressivity of the SeCo data model and to provide more complex operators to compose service results, such as similarity and proximity operators and, more generally, fuzzy operators that returns a matching score instead of a simple truth value.
The goal of the proposed thesis is to develop an expressive and flexible expression language for Search Computing applications, by:

• identifying the join, selection and ranking requirements of Search Computing applications;
• devising an expression language satisfying these requirements and formally defining its semantics;
• implementing a validating parser and an efficient evaluator for the expression language;
• providing support for user-defined functions and predicates to be used to extend the language.

Previous knowledge of formal grammars (BNF) and artificial language parsing techniques is strongly recommended.