The unified approach embodied in data flow frameworks has inspired several tool­kits that partially automate the construction of solvers for various data flow problem classes. Unfortunately, existing toolkits have fairly limited application. Although most can handle both classical and more advanced analyses, the full breadth of flow analysis forms-particularly those arising in new applications of the technique-lies outside the range of any single system.

This dissertation extends and reformulates the traditional approach, in order to support the automatic generation of solvers for multisource data flow analysis prob­lems. In this very general class, a data flow problem is modeled by a directed graph in which more than one type of edge may be defined and information about the type of an edge considered along with the flow value it carries. While this describes, roughly, all forms of flow analysis, attempts to unify the members of the multisource family have so far held only theoretical interest. The approach we present here thus offers a significant increase in flexibility to existing flow analysis tools.

Our method is based on a new approach to user-level specification of data flow problems. Existing approaches require instantiation of the four parameters of a data flow framework: value lattice, function space, flow graph, and local semantic func­tional. All such approaches presume a fixed global semantics, an assumption that fails in the general multisource case. Instead, we propose a domain-specific language, which harkens back to the earliest "flow equation" forms, allowing us to view the global abstract semantics itself as a fifth parametric component.

We then leverage this language-theoretic view to develop a new technique for the automatic generation of efficient solvers. There are several improvements to the basic iterative solution strategy that exploit properties of the flow graph model in order to discover advantageous orderings and avoid redundant computation in the global solution. All rely on a known relationship between the structure of the flow graph and the flow of information between nodes. To adapt these to the multisource case, we present a new method for discovering the information flow relation induced by arbitrary global semantic specifications.