Options
2016
Conference Paper
Titel
Combinatory synthesis of classes using feature grammars
Abstract
We describe a method for automatically transforming feature grammars into type-specifications which are subsequently used to synthesize a code-generator for a product of a given feature selection. Feature models are assumed to be given in the form of feature grammars with constraints, and we present a generic type-theoretic representation of such grammars. Our synthesis method is based on an extension of previous work in combinatory logic synthesis, where semantic types can be superimposed onto native APIs to specify a repository of components as well as synthesis goals. In our case, semantic types correspond to feature selections. We use an encoding of boolean logic in intersection types, which allows us to directly represent logical formulas expressing complex feature selection constraints. The novelty of our approach is the possibility to perform retrieval, selection and composition of products in a unified form, without sacrificing modularity. In contrast to constraint based methods, multiple selections of a single feature can coexist.