Maybe I'm thinking about computing as an end itself and just too insecure to acknowledge its application aspect. Because it seems to me that when the application becomes the end-goal then the engineering loses its value.
You'll see stories all the time about machine shops with MS-DOS programs still in use, for example. But consider the facts, a problem was solved, that solution has worked consistently for decades, any new solution is likely to create problems, so of course it's going to stay around forever. Just don't hook it to the internet, and it will be fine. 8)
Once a business need is met, you just want to keep the friction as low as possible.