Programs can do things very reliably. Good programmers can keep the programs running without d'oh moments... :-)
Yes, the trick is that they must maintain their master data correctly and completely, otherwise it does not work and are sent mail alerts to remind them that the cost center is not assigned to a profit center etc as being the reason for it not working.
They only have themselves to blame, are given sufficient informative error / warning messages and alert mails about data errors which the processing is waiting for.
This year we are going to replace testing completely with automated change workflows from simulations of changes in production and via SOLMAN route them back to the corresponding development system to transfer the simulated change (failure) and record it into a transport request automatically. The user does not even notice anything or the error nor the change.
What lengths folks go to, to avoid tickets... :-)
Cheers,
Julius
