Oversimplifying the message in the paper by Moshe and Sumit, in their opinion the plethora of process semantics is due to "semantic underspecification". To overcome this underspecification, they propose to develop process semantics based on the following principles.
- Principle of Contextual Equivalence: Two processes are equivalent if they behave the
same in all contexts, which are processes with “holes”.
- Principle of Comprehensive Modeling: A process description should model all relevant
aspects of process behaviour.
- Principle of Observable I/O: The observable behaviour of a tested process is precisely
its input/output behaviour.
As the authors state in the paper "In conclusion, this paper puts forward an, admittedly provocative, thesis, which is that process-equivalence theory allowed itself to wander in the “wilderness” for lack of accepted guiding principles. The obvious definition of contextual equivalence was not scrupulously adhered to, and the underspecificity of the formalisms proposed led to too many interpretations of equivalence.While one may not realistic expect a single paper to overwrite about 30 years of research, a more modest hope would be for a renewed discussion on the basic principles of process-equivalence theory."
What do you think?