[delayed]

Old habits die hard.

There was a time when the company I was working for was selling service contracts for DEC TZ30 tape drives to DEC itself because their own internal service organization was unwilling or unable to support them.

The application was software distribution, most likely to customers they were contractually obligated to support, and no newer drives were able to write the required tape format, so upgrading to newer drives wasn't an option.

It's both time tested and at risk of ripping at the seams. Floppy disks and the mechanisms still wear. And because neither are still made, the chances of a zero population scenario happening is greater every year.

It's better to test a replacement while you still have the time, than to risk rushing something else.

Floppy disks aren't manufactured anymore. At some point there aren't going to be any.

The overall reliability of a system is time invariant... its either reliable, or not.

The fact that it has gone x days without an incident doesnt upgrade or downgrade its reliability. The engineering and processes are function in an equation that calculates mean time to failure. Life-critical processes have numbers placed to this... like a failure every 50 years or 100 years... or 100000 years, and you work towards solving that.

Systems are dynamic, not static. There are secondary consequences and forgotten/ignored opportunity costs. Sure, the existing system has proven workarounds, though labor and attention intensive. When a floppy disk fails it won't directly create disaster. But these legacy requirements do block the more cohesive and thorough introduction of additional safety mechanisms that increase redundancy and reduce the scope of possible human error - such as new sensor hardware and their automation, latency improvements, and information synthesis/notification clarity.

Here's a software analogy - required support for very old OS versions often interferes with the scope, simplicity, and security of new feature work, and even if the work is still feasible at that point it is slowed down greatly, eating up attention and and time that could go towards addressing other important issues. That accumulates over time, and cumulative loss of improvements that could have otherwise been made has a real effect in unprevented disasters and wasted controller availability.