For starting up a process control system I am used to this routine:

1. Install the various subsystems.
2. Perform hook-up of the subsystems.
3. Post-installation testing of the subsystems: this is usually basic things like testing of connections for integrity and continuity, no functional testing.
4. Pre-commission the subsystem: functional checks, particularly against interfaces, but may also include leak and continuity tests.
5. Pre-commission the system: end to end functional check of the system, i.e. very few, if any, forced or simulated inputs.
6. Commission the system: this ends up being very little because all the work was done in pre-commissioning the system. Hence the question. What should be in pre-commissioning and what should be in commissioning? Is pre-commissioning at the subsystem level and commissioning at the system level? Our commissioning would normally also include displacing storage or transportation fluids from process and utility lines, replacing them with the working fluids, but most of this would be outside control system scope.
7. Startup of the process: operations take over here and follow their own procedures for opening valves and controlling the subsequent behavior of the process.

This routine is used for bringing on the control system for a reasonably complex process on a fairly large scale. The control system comprises everything from the HMI through to hydraulic actuating devices, remote PLC based I/O systems, pneumatic control of hydraulic and utility supply systems, interfaces to process shutdown systems.

The general definition seems to be that commissioning is the work required to bring a system to the point where it can be used. This may include leak tests, continuity tests and functional tests.