Was a very memorable time, great research and great people <3
Ted Nelson's Xanadu is a great example - it's core idea was that information is never deleted. This was literally baked into design - the main data storage is "append only". So sounds good on paper, but how was this supposed to be practical, especially with the high cost of storage back then?
Or that "unified document" with a workspace - sounds great on paper, until you start to think about more general use cases. A friend gives you few dozens of files on floppy (or you download from BBS), and you want to pass them to next friend. How do you do this easily without concept of "files"?
So there is no surprise that industry ignores "visionaries". It's easy to make a prototype and show it on presentation, but it's hard to make something actually usable.
The funny thing is that given how crap things have gotten it doesn't seem like it would be very hard at all to architect a, "radically improved" version of modern computer interfaces. We even have LLMs to help facilitate parts of the system that might have been historically difficult to implement. Why not instead of building a, "modern" Windows or Mac OS you made a, "useable" version which was optimized to run on anything or could run with any stalling on a modern computer? I don't want Windows 11; I want Windows 25' I want it to work orders of magnitude faster than Windows 98 rather than using Moore's Law to create something that can, "do more with more resources but averages out to roughly the same experience as previous generations."
We're still effectively using the same computers we were using when I was a kid in the 1990s.