Old nuclear power plants, old large buildings with climatic systems, old military installations/machines, old (and not written off) product machines in companies, banks (COBOL) … These are the first things that are attached to old equipment and associated with software. However, I’m also sure I forgot a lot.
Upgrading software is often a huge expense without any obvious benefit. At least, in the minds of the people/management who have to pay and/or justify these costs.
I don’t deny that there have been enough people very attached to their specific software package on their computers at home for a very long time. But for companies, there are financial consequences that were not taken into account in your statement “If you’re still operating as a company with software that’s over 30 years old, you’re doing it completely wrong.”
Itanium’s longevity is solely due to contractual obligations on the part of Intel and its customers. So Intel had to develop further and the customer had to buy this processor in relatively small numbers. It was cheaper to continue operating in this way than to cancel the current contract. No one was really happy with the Itanium processor line.
Personally, I am of the opinion that after 10 years the hardware and software should be replaced. This is long enough to get a good overview of the sensible and nonsensical hardware and software developments. In other words, progressive insight. If you use a shorter period, you are playing a big role with no real trend. With a longer period you are very far from the facts.
“Coffee buff. Twitter fanatic. Tv practitioner. Social media advocate. Pop culture ninja.”