VOQL

Abstract

Expressing complex query expressions in a concise and intuitive way has been a challenge in the design of visual object-oriented database query languages. However, partly due to lack of well-defined semantics and syntax of the languages and partly due to improper representation of their visual constructs, most previous visual query languages fail to meet the challenge. Specifically, they fail to handle the semantic complexity and ambiguities resulting from multi-valued path expressions, which are important to properly model the features of object-oriented query languages- especially with more than one nesting level. In this dissertation, we propose a visual query language for object-oriented databases called Visual Object-Oriented Query Language VOQL. VOQL has such good features as excellent expressive power for object-oriented queries, well-defined semantics, and concise and intuitive syntax, which few previous visual object-oriented query languages have. By employing the visual notation capable of representing set inclusion and functional relationships, the semantics of multi-valued path expressions can be properly specified in VOQL queries. In particular, we introduce the visual query construct called the VOQL path expression constructed from the query primitives such as blobs, nested blobs, binding edges, and flattening edges to enhance expressive power. Blobs and nested blobs are used to denote sets of objects that path expressions represent. Binding edges and flattening edges are designed to visually simulate the notion of variable binding and dot functions in path expressions. Based on the query constructs of VOQL, we provide the syntactic rules to define the syntactic components such as VOQL terms, VOQL formulas, and VOQL query expressions in a way as in tuple relational calculus. We illustrate the semantics of such components by providing syntax-directed translation to corresponding counterparts in Object-Oriented Predicate Calculus (OOPC). In combin...