Back in the day my father was a service tech for NCR retail systems. For one of the lines they shipped out every system fully loaded with ram and disk - however how much their customers got to use depended on jumper configurations which the service techs would set and verify. I think IBM does the same thing with their mainframes - they ship out fully loaded but you can only use what you pay for, the rest is used as hot stand-by for hardware failures.
The company I now work for does something similar. We sell our product at what is essentially max capacity, capping memory/disk space/CPU per the contract. When the customer needs the extra capacity or wishes to upgrade, all that’s required is some settings changes and a restart.
I remember someone telling me that the printer in my primary school was only a black printer. When they wanted to upgrade to colour, a technician came down and flipped a dip switch
Former NCR tech here. I can confirm the parent’s comment - we’d routinely have to validate/ensure jumpers based on entitlement. As for recovery, I worked for NCR once they regained their independence from ATT and they were still running themselves into the ground. By the time I left (around the era of the 7878) things improved slightly.
