[Further work is required here]
Preliminary discussions with the LAr and TileCal groups suggest
two possible scenarios for doing calibrations. Typically both
involve a scan over a range of values of some parameter.
This scan can be done in a single run. In this case the system
has to operate in a run/pause loop with the parameter value
being changed during the pauses. Alternatively it can be a series
of runs, each of which uses a single value of the parameter.
The multiple runs scheme may be controlled externally by some
script. Otherwise there are not many implications for the run
control hierarchy or transition actions.
The single run scenario however will require synchronisation
between pauses of the trigger (handled in the TTC crate
after a specified number have been taken)
and changes of the parameter values in other crates.
The Paused state (and Pause/Resume) commands can be used for
this purpose.
In this scheme, there would be a separate process polling eg
the TTCvi to count triggers. This process would then send
Pause and Resume commands to the Root run controller at the
appropriate times. This ``calibration sequencer'' process
would be started and stopped by the normal run Start and
Stop commands.