An Abstract Data Model for the Tabulator Data Browser

One of the main features of the RDF data model is its ability for modeling highly interconnected data in an extensible way. In this sense, given the graph-like nature of RDF, browsing RDF data arises as a natural requirement. At the moment, there are many applications for browsing RDF, for example Palm-DAML, RDF Author, IsaViz, CS-Aktive, mSpace, BrownSauce, DAM viewer. These applications, commonly named Semantic Web Data Browsers, use different visualization models (e.g. circle-and-arrow diagrams), each one having its advantages and disadvantages. The Tabulator [2] is a semantic data browser which provides the user with different views (e.g. outlines and tables) for visualizing, querying and browsing of RDF data. The central component of the Tabulator is the outliner view, a tree-like environment that provides a powerful interface for presenting RDF data as a nested structure of resources. Moreover, the user is able to express queries through the definition of graphical patterns. The objective of this paper is to show that underlying the outliner view, there is a powerful data model which needs to be studied formally to take advantage of its properties. In this sense, we present and analyse the outliner view. We define a general framework based on the notion of nested graphs, a concept that was studied in the area of databases by the hypernode model [9; 7; 12]. An hypernode extends the plain structure of a graph to a nested structure, allowing a simple and flexible representation of nested complex objects. We study nested graphs in the context of complex objects in databases, in particular the notion of equivalence. In this direction, the information capacity of a nested graph is defined in terms of its plain representation called a graphset. We treat data relativism [5] when analyzing the issue of transforming nested graphs into graphsets and viceversa. Relevant properties for visualization of nested graphs are defined. Finally, we present an algebraic query language oriented to capture the basic queries supported by the outliner-view. This language consists in two basic operators for selection and projection over a set of nested graphs. Both operators are based on matching of nested graph patterns and are closed under sets of nested graphs. The contribution of this paper is to define formally an abstract data model which allows us to study properties and characteristics of the Tabulator. We show that nested graphs, graphsets and RDF graphs are three equivalent abstract representations for RDF data. Using this, we propose that the nested graph model can be considered an abstract representation for the outliner view of the tabulator RDF browser.