STRONGLY TYPED FLOW-DIRECTED REPRESENTATION TRANSFORMATIONS

We present a new framework for transforming data representations in a strongly typed intermediate language. Our method allows both value pro ducers (sources) and value consumers (sinks) to support multiple repre sentations, automatically inserting any required code. Specialized rep resentations can be easily chosen for particular source/sink pairs. Th e framework is based on these techniques: 1. Flow annotated types enco de the ''flows-from'' (source) and ''flows-to'' (sink) information of a flow graph. 2. Intersection and union types support (a) encoding pre cise flow information, (b) separating flow information so that transfo rmations can be well typed, (c) automatically reorganizing flow paths to enable multiple representations. As an instance of our framework, w e provide a function representation transformation that encompasses bo th closure conversion and inlining. Our framework is adaptable to data other than functions.