Indeed; what happened to systems like DR DOS, BeOS, and AmigaOS is a disgrace. DR DOS is still "supported" in the sense that you can pay for a copy from the ostensible rights holder, but the company doesn't appear to provide any actual development or support services, and it looks like they haven't issued a press release since 2004. DR DOS source is also out there, but only licensed for noncommercial use.
As far as I can tell, it's owned by DR DOS Inc. which also had a d/b/a "DeviceLogics" that now seems to be defunct (inasmuch as devicelogics.com now redirects to drdos.com). However, the status isn't exactly obvious to me.
The only source release I know of that is still readily downloadable is at drdos.net, which is version 7.01.
Imagine a fork of various Apple OS's that took skeuomorphic interfaces to their logical conclusion!
We need a name for this: an open set of parallel computing universes. Maybe OpenCandide: The Best of All Possible Worlds. (Ironic misinterpretation of the phrase intentional.)
I'm not sure if you were implying otherwise or not, but: spatial metaphor file managers are not skeuomorphic. Skeuomorphism strictly encompasses non-functional interface elements, such as faux-leather stitching on Apple's old address book application. The spatial metaphor Finder in OS 9 was 100% functional.
The way it remembered all of the details (size, layout, position, display mode) of each window and the way each window had an unbreakable 1 to 1 connection to a folder made it a joy to use. It allowed the user to identify what window belonged to a folder at a glance and to be able to set up a whole workspace arrangement and have it be remembered automatically by the Finder.
It would be interesting to have a system where getting a software patent required publishing working code under public domain. Then when the patent expires, it's open for all.
Which is arguably the point: the patent term should cover the likely amount of time that commercial exploitation is most advantaged by a government granted temporary monopoly. Obsolescence is pretty clearly a point where commercial exploitation is (severely) disadvantaged. The goal for patents was that for that temporary leg up (and today's 20 years starts to feel a lot less temporary) the company had to make all the information and science and schematics public domain. It's not a bad idea that if software/software algorithms are patentable the actual code should fall under that public domain disclosure. The balancing act that remains is the question of how much code is "schematics" and how much of it is "final product" and how much protection products deserve versus schematics...