From ACT-ONE to Miranda, a Translation Experiment

It is now almost universally acknowledged that the data language ACT-ONE associated with the formal description technique LOTOS is inappropriate for the purpose of OSI formal description. In response to this the LOTOS restandardisation activity plans to replace ACT-ONE with a functional language. Thus, compatibility between ACT-ONE and the replacement data language becomes an issue. In response to this, we present an experimental investigation of backward compatibility between ACT-ONE and the new LOTOS data language. Specifically, we investigate translating ACT-ONE data types into the functional language Miranda. Miranda has been chosen as it is a widely used functional programming language and it is close in form to the anticipated new data language. This work serves as a ``verification of concept'' for translating ACT-ONE to the E-LOTOS data language. It identifies the bounds on embedding ACT-ONE in a functional data language. In particular, it indicates what can be translated and what cannot be translated. In addition, the paper reveals pertinent issues which can inform the E-LOTOS work. For example, which constructs are needed in E-LOTOS in order to support the class of data type specifications typically made in the LOTOS setting? We conclude with a number of specific recommendations for the E-LOTOS data language.