We propose to work on the refinement within the tuple space paradigm. Tuple space and mu]tiple tuple space languages, based on this paradigm, are well suited to model cooperation. More generally, they are suited to program large distributed reactive systems. They support concurrency providing both syntactic constructs to define autonomous entities of computation, i.e. the processes, that can be executed in true parallelism and mechanisms for process synchronization and cooperation. Accesses to the tuple spaces are defined by rules, that are composed of a query on the tuple space followed by its update. Queries and updates are compositions of read, remove, or write primitive statements.
In particular, we want to extend the current results on refinement to the languages TAO and Pbti, which are developed in Lisbon and Pisa, respectively: both these languages are based on the multiple tuple spaces model, and permit to declaratively define computations as combinations of rules. Indeed, they represent two examples of languages instantiating in a similar way the tuple space paradigm, in the direction of enhancing the expressive power, and allowing sensible distributed implementation.