It doesn't mean you can't build a new system to replace the old one. Just that you should not mess with the one running perfectly good, directly.
Innovating for the sake of innovating is just pointless (albeit trendy), improvement, progress should be the words used. The start menu of win8 or the taskbar of win 11 are just obvious examples where "innovation" comes at the price of a regress in the user experience. I could mention hundreds of cases where this is applicable to specific applications, the application used by companies and professionals to be productive. That said, many entities that decide to stick with the old thing, do that not because they like to be prisoners of the good old times, but because they already *are* prisoners. Prisoners of a bankrupt company, of a retired coder, of a old SW that hasn't an obvious modern replacement. Migrating an old app based on a proprietary DB after maybe two decades of usage (and debugging, and refinement) could be a gigantic task, w/o any guarantee that the new shiny thing will work better (or even will do the same job). Last but not least, some may feel more comfortable feeling prisoners of time rather than begin blackmailed by monthly fees, and dystopian practice that are so common in today's IT.
I was underlying the problem of freedom prevention when you wait too much, including the ability to continue to use legacy softwares. This planned obsolescence is still too often organized even today. To get back on Windows XP/Windows Server 2003 Setup, I have shared some ideas in post #16 with some paths to explore.