Expressing global priorities by best-matching

The relative expressive power of global priorities against the local priorities induced by a best-matching communication paradigm is investigated. Two specific languages are taken as referential cases: FAP and COWS. FAP is a finite fragment of asynchronous CCS with explicit priorities. COWS is a process calculus with naming that was specifically devised for modelling the orchestration of web services and where best-matching serves the purpose of giving precedence to service instances over service definitions. Two main results are shown, one on the negative and the other on the positive side. First, we prove the impossibility of encoding FAP into COWS if the translation has to meet a few specific and reasonable requirements. Second, we define an encoding of FAP into COWS that does not respect the above requirements. We show however that it still possesses desirable properties.