It sounds like you're not accounting for the fact that we have zero full-time employees, and hundreds of such bugs.
I understand that there are ~5 developers, all working part time. There's no question that limits what can be done.
At the same time I think 'hundreds of ->such<- bugs' is much worse than the reality. There ARE hundreds of bugs, but not all of them will show up on the way to a functioning desktop and installation of a browser using the Applications Manager. Thinking about what I've been experiencing -- ROS can't be installed at all on my Dimension 3000 due to file system troubles and on the Latitude D610 I'm using it often fails somewhere in second stage of install or soon after the first subsequent boot -- I think more in terms of no more than three real bugs and wouldn't be surprised to learn that it was just one that affects the file system. Particularly something that might be invisible or bypass-able if running in a virtual machine rather than on real hardware.
Timing dependent? These machines were among the fastest off-the-shelf-at-Wal-Mart of their day -- 2005 or so.
ROS often asks for a file system integrity check when booted. When it is allowed to do the check there are always problems, sometimes just a slight error in reporting of free space, often much worse. If the file system is corrupted, death will follow, though the apparent manner may vary.
I just tried a boot of the distro of Oct. 16th: It wanted a file system check, I let it go ahead, it reported shared clusters, it truncated PAGEFILE.SYS, then BSOD'ed out. Several repeats from the distro CD over a couple of hours got me to various modules reported as not made for ROS or corrupted, then to can't download an intact Firefox 48 (via App Mgr) and installing it from a copy downloaded on another machine succeeded but hangs on invocation.
When ROS can pass that test -- install, boot, install a browser, and look at a web site or two -- on most machines close to every time as long as you stay on the narrow path, it can be promoted as "Try this out!" That'll bring more resources. It's an EXCELLENT job, absolutely natural to use and will sell itself. Right now it's only for true fanatics, of which I am definitely one.
And this is one of the reasons I advocate open-sourcing of Windows. Windows will get more backwards compatibility and some performance enhancements, ReactOS will get more users and testers.
From the manufacturer's point of view 'open source' is a loser: You can't really sell it because you didn't make all of it. Microsoft makes a lot of money selling software.
Back in the day when the machines were big and few (hence VERY expensive) the money could be made there. Users often had microfiche and source code for the OS and we often debugged problems for the maker. But the OS could be a FREEBEE because we paid REALLY BIG $ for the machine on which it ran.
Which was maybe 1000 times slower than the one on which I type this. The operations were bigger so direct comparison isn't possible but that's the ballpark.
I can't see a route to paying MS (or any major company) very much other than for software. If computing moves into the cloud the software will be there with the data -- still not ours. The only hope is efforts like ROS (and Firefox and Linux of course) that are open source developed and ultimately carried forward by foundations (like Mozilla) that organize and distribute the result. So I don't see how open source Windows could happen.
But ROS definitely could. Damn I wish I'd heard about it at the start. 20 years younger some different choices would have been available.