The concept of branching time in the semantics of concurrent
systems is well known and well understood. Still a formal definition
of what it means for a model or equivalence to respect branching time
has never explicitly be given. This note proposes such a definition.
Additionally the opportunity is taken to voice an old but poorly
understood argument for using branching time semantics instead of
models or equivalences that are fully abstract with respect to some
notion of observability.
Rob van Glabbeek