(okok, it's Saturday so can we have this as a top level reply?)
It definitely is, not just to see what technologies were in use, but what an exemplary career in that field looked like, so thanks for sharing. It's just that the first thing that went through my head was this joke, since it's really something that isn't too far fetched situation wise (see that one longer reply).
Fun story: About 6 years ago, my uncle was still working at a big car manufacturer. They have bounties for when employees come up with ideas to optimize some workflow or cut costs somewhere. He handed in a suggestion to stop computing big reports on one of their mainframes that were meant for being printed out but weren't actually for quite some time, so basically just went to /dev/null, while the generation process took quite some time. There was quite some confusion first since the IT people in charge insisted no such thing was going on, so they had him show them where this job was actually wired in and where the output went. In the end they decided to remove it and paid him some reward. How did he know? He coded that job in the 80s.
I guess your dad would have quite some interesting stories too.
* My uncle was actually a maths teacher, but right when he finished college there was absolutely no demand, so a friend already working there got him on board. He learned cobol while trying to figure things out there, reading books day and night...
Sorry about that. I have this theory on dementia that part of the cause (stress among other things) is a change from an intellectually demanding career or a life involving thinking and a change to either retirement or not having to do much thinking. The delta in other words.
So my question is does any of this relate to your father? Did he remain active in thinking past his retirement?
The one common reason why all replacement attempts falter if not fail. No one knows all the touch points. No one knows exactly what they really do. Oh its easy to say "that machine is the warehouse inventory and related" but on a grand scale that is a fine description. it is when you get into all the custom interfaces, all the different systems relying on it, then it becomes a whole different ball game.
I full expect the z, i, and p, machines to all be there in the next ten to fifteen years for the same reason they are here now. it gets really expensive to replace something when you don't even know what you truly have. when you do find out it can be far more economical to keep what you got but be more diligent where new apps reside.