When is an encoding good? Full abstraction and other criteria. Discussion introduced and moderated by Daniele Gorla and Uwe Nestmann. Summary. One of the hottest topic for the EXPRESS audience is the definition of the criteria that make an encoding 'good': this is a crucial issue to accept an encoding or to prove separation results. Here are some specific problems we will discuss. What is the the relevance of operational correspondence in proving separation results? What kind of renamings should an encoding preserve? Is it mandatory to map parallel composition homomorphically?