Reified Literals: A Best Practice Candidate Design Pattern for Increased Expressivity in the Intelligence Community

Reifying literals clearly increases expressivity. But reified literals appear to waste memory, slow queries, and complicate graph-based models. We show where this practice can be comparable to unreified literals in these respects and we characterize the cost where it is not. We offer examples of how reification allows literals to participate in a variety of relations enabling a marked increase in expressivity. We begin with a case study in reified person names, and then extend this analysis to reified dates and simple reified scalar values. We show benefits for name matching and temporal analysis such as would be of interest to the Intelligence Community (IC). We then show how these same sorts of analyses can drive or inform any decision as to whether to reify literals.