Failures, Finiteness, and Full Abstraction

For a simple CCS-like language a trace semantics and a failure semantics are presented. The failure semantics is shown to be fully abstract with respect to the trace semantics if and only if the set of internal actions is infinite.