https://github.com/microsoft/MS-DOS

Now, I'm not saying that this should be thoroughly rifled though by many projects, but I am saying there were a lot of games that at the very least, ran on Dos 3.

The description says that just DOS 4.0 is newly "freed", as 1.25 and 2 have been available for ten years.
But unless it'll improve the existing emulators in some notable way, which seems unlikely, it's not exactly notable.
Now if they'd "free" Win 3.1, that'd open a whole (albeit brief) generation of games to being made legally available again...

I did not know DOS was written primarily in assembly.

Well, anything higher than assembly would be quite insane. Not only was it targeting anywhere from 4-40 Mhz across a broad swathe of platforms, not only would including anything higher would not only create dependencies, but also it would take up precious space when you had to communicate an OS that might have to live in under 160 kb, with nothing for storage but other 160 kb disks.

A lot of software was written in raw assembly, especially if you wanted more performance.

Most early console games were also written in assembly. There's a handful of BASIC games for 8bit machines (Questron comes to mind), but doing assembly was like 400x faster in speed than using the tools available. Course it depended on what you needed.

I did a bunch of assembly language programming when i was a teen (12-13), you tended to put arguments into your registers and use pusha/popa as calls, this was partially space and speed rather than pushing everything onto the stack, among other things. And you didn't have libraries, so exe files weren't a thing (at that time), instead you used com files which skipped a header and your whole program was under 64k (actually it was probably much smaller than that, unless you somehow needed multiple memory segments, but those were actually fairly rare. You'd be surprised how small you can crunch a program down in raw assembly.

I'm sure it makes sense. I had just never thought about it. I was aware of Unix having been ported to C around 1973 so I had assumed (not thinking too hard about it) that a more recent operating system would be likely in C too. But of course Unix was probably running on better hardware at the time. And maybe back then C was uncommon outside of Unix. I'm unfamiliar with what the programming landscape was like back then.

Yes, Unix was running on "better" hardware, but another factor is that the way Unix ran itself in very interesting ways to get around the limitations of the mainframes, minicomputers, and timesharing operations. AT&T's tech archives has fascinating (youtube) videos on the matter, and I'm going to link one here: Beware, it does get somewhat technical.

If / when Microsoft decides to open source MS-DOS 6.22, that will potentially be a big help for DOS emulation too.

That or anything from Windows 1-3; since they were more DOS augments than a proper OS that could crash and die.

I've got a paperback book i bought which was programming in the unix environment, 16bit too. Several API operations involved sending multiple 16bit arguments to add up to the 32 or higher; Though those were more for seek, read/write lengths, time, etc. While fascinating, being in 32bit systems i kinda just bought the book for novelty sake and only glanced at it. It was still being sold for like $40 back in 2005, where the owner refused to drop the price. (then i got it on discount for being an employee for a while).

Bout time since DOS itself was a well known ripoff of another OS that made PC work as we know it possible. Too bad when the idiot released it to the market after suing MS he wanted to WAY overcharge vs. DOS.

Quick & Dirty or Control/Program Monitor?