A parallel framework for computational science

Parallel languages based on skeletons allow the programmer to abstract from implementation details, reducing the development time of the parallelizations of large applications. Unfortunately, these languages use to restrict the set of parallel patterns that can be used. The parallel functional language Eden extends the lazy functional language Haskell with expressions to define and instantiate process systems. These extensions also make possible to easily define skeletons as higher-order functions. By doing so, skeletons can be both defined and used in the same language, using a high level of abstraction. Due to these facts, the advantages of skeleton-based languages are kept in Eden, while we do not inherit the restrictions they have, as the set of skeletons can grow as needed. Moreover, in our approach the sequential code of the programs can be written in any language supporting a COM interface.