%T How to Make a Process Invisible
%A  Neil C.C. Brown
%E Peter H. Welch,  S. Stepney,  F.A.C Polack,  Frederick R. M. Barnes,  Alistair A. McEwan,  G. S. Stiles,  Jan F. Broenink,  Adam T. Sampson
%B Communicating Process Architectures 2008
%X Sometimes it is useful to be able to invisibly splice a
   process
into a channel, allowing it to observe (log or
   present to a GUI)
communications on the channel without
   breaking the synchronous
communication semantics.
   occam\-π\[rs]s extended rendezvous when reading
from a
   channel made this possible; the invisible process could
   keep
the writer waiting until the real reader had accepted
   the forwarded
communication. This breaks down when it is
   possible to have choice
on outputs (also known as output
   guards). An extended rendezvous
for writing to a channel
   fixes this aspect but in turn does not
support choice on the
   input. It becomes impossible to keep your
process invisible
   in all circumstances. This talk explains the
problem, and
   proposes a radical new feature that would solve it.