Putting Integrated Information in Context: Superimposing Conceptual Models with SPARCE

Abstract

A person working with diverse information sources—with possibly different formats and information models—may recognize and wish to express conceptual structures that are not explicitly present in those sources. Rather than replicate the portions of interest and recast them into a single, combined data source, we leave base information where it is and superimpose a conceptual model that is appropriate to the task at hand. This superimposed model can be distinct from the model(s) employed by the sources in the base layer. An application that superimposes a new conceptual model over diverse sources, with varying capabilities, needs to accommodate the various types of information and differing access protocols for the base information sources. The Superimposed Pluggable Architecture for Contexts and Excerpts (SPARCE) defines a collection of architectural abstractions, placed between superimposed and base applications, to demarcate and revisit information elements inside base sources and provide access to content and context for elements inside these sources. SPARCE accommodates new base information types without altering existing superimposed applications. In this paper, we briefly introduce several superimposed applications that we have built, and describe the conceptual model each superimposes. We then focus on the use of context in superimposed applications. We describe how SPARCE supports context and excerpts. We demonstrate how SPARCE facilitates building superimposed applications by describing its use in building our two, quite diverse applications.