A poignant story from 1994: https://user.eng.umd.edu/~blj/funny/requium.html It gets me to reflect that I'd have dismissed it 10 years ago, focusing on the superficial complaints and utterly missing the deeper complaint: of a community of astronomers losing something that used to work.

VMS had a great deal going for it. Just as one example, files were named in a conventional way, like

FOO.TXT

, but with an extra piece of metadata at the end:

FOO.TXT;1

. Whenever one changed a file it would create a new file with the number at the end incremented:

FOO.TXT;2

. A directory listing would show all versions. One could open any version by including the version metadata in the filename, and opening a file without a specific version number would just open the most recent version. One could also delete specific versions as desired while keeping other versions around, like a simple revision control mechanism built into the filesystem. Other things, like fault tolerant clustering, load shedding, disk quotas, and so on, also worked very well. TBF, there are many reasons one might consider the victory of Unix over VMS a worse-is-better thing. OTOH, the VMS C compiler was terrible (leading me to write hundreds of thousands of lines of VAX Pascal) and the command line language (DCL) lacked the grace of using pipes under Unix.