Chapter 6 - Using code models to instantiate data models

Abstract

This chapter will introduce a particular graph model which we call “syntagmatic” graphs, applicable to both natural-language parse-structures and to modeling computer code. We explore some properties of queries over these graphs and briefly consider features of a virtual machine whose capabilities would include operations forming the foundation of syntagmatic-graph query-evaluation. We argue that syntagmatic-graph queries are structurally related to other software-engineering constructions, such as aspect-oriented programming, and discuss implementation strategies for static code analysis and runtime-reflection, in the sense of exposing procedures to scripting and remote-procedure interfaces. We then introduce a schematic figure we call the “semiotic saltire,” which informally classifies programming concerns into four engineering “aspects” supplemental to core type-implementations, namely GUI design, serialization, (database) persistence, and “procedural models,” the latter of which includes details such as runtime reflection and remote-procedure call capabilities, as well as annotative procedural pre- and post-conditions. We conclude by introducing certain benefits of hypergraph-based representations for data persistence.