Superb. Excellent sleuthing and development, and well written-up too.
I supported a few Apricot machines in production in the late 1980s, and Sirius too, although I had forgotten the strange keyboard layout. It always was a better design than IBM's PC, or XT or AT come to that.
It had long seemed to me that if Apricot and the other non-PC-compatible DOS vendors had just been able to hang on in there until later in the Windows era than the fairly bad Windows 1 that they'd have suddenly had a much better chance. This work sort of serves as an existence proof: given Windows 2, an 8086-based Apricot is suddenly much more compatible with way more mainstream PC software than it was running DOS.
Apricot did survive, of course. The only SCO UNIX…
… a huge tower server on castors with a built-in UPS as well as 5.25" drive bays. Before we provisioned it with UNIX and deployed it at the customer's site, we put DOS and Castle Wolfenstein on it, and me and 2 colleagues played Wolftenstein while trundling it up and down a (very smooth) corridor. Its built-in UPS was beefy enough to drive a colour VGA monitor as well as the computer, so with the screen balanced on the system unit, 1 colleague rolled the server while another colleague rolled an office chair with the player sitting on it.
This machine showed Apricot again backing the wrong horse: it's the highest-end x86 IBM Microchannel machine I ever worked on.
… SCO Xenix was -- and for compatibility had to be -- built with MS C, not AT&T C. So every copy of every SCO OS meant SCO had to pay a lot of royalties.
I supported a few Apricot machines in production in the late 1980s, and Sirius too, although I had forgotten the strange keyboard layout. It always was a better design than IBM's PC, or XT or AT come to that.
It had long seemed to me that if Apricot and the other non-PC-compatible DOS vendors had just been able to hang on in there until later in the Windows era than the fairly bad Windows 1 that they'd have suddenly had a much better chance. This work sort of serves as an existence proof: given Windows 2, an 8086-based Apricot is suddenly much more compatible with way more mainstream PC software than it was running DOS.
Apricot did survive, of course. The only SCO UNIX…
https://www.os2museum.com/wp/sco-unix-3-2-0f-limping-along/
… machine I ever installed was an Apricot VX/FT server…
https://ardent-tool.com/Apricot/vxft/
… a huge tower server on castors with a built-in UPS as well as 5.25" drive bays. Before we provisioned it with UNIX and deployed it at the customer's site, we put DOS and Castle Wolfenstein on it, and me and 2 colleagues played Wolftenstein while trundling it up and down a (very smooth) corridor. Its built-in UPS was beefy enough to drive a colour VGA monitor as well as the computer, so with the screen balanced on the system unit, 1 colleague rolled the server while another colleague rolled an office chair with the player sitting on it.
This machine showed Apricot again backing the wrong horse: it's the highest-end x86 IBM Microchannel machine I ever worked on.
https://en.wikipedia.org/wiki/Micro_Channel_architecture
MCA was better than PC ISA or VL-bus and for some things better than EISA:
https://en.wikipedia.org/wiki/Extended_Industry_Standard_Arc...
SCO UNIX was also arguably SCO backing the wrong horse too. I learned Unix on the older SCO Xenix:
https://en.wikipedia.org/wiki/Xenix
… which was smaller, simpler, and faster. SCO UNIX was more "official" but not better in any useful way.
However, Xenix had serious issues, some of which Charlie Stross recently documented in a comment on my blog:
https://liam-on-linux.dreamwidth.org/97149.html#comments
… SCO Xenix was -- and for compatibility had to be -- built with MS C, not AT&T C. So every copy of every SCO OS meant SCO had to pay a lot of royalties.