In this day and age of net threats, who on earth is going to expose something to the internet, whereby all processes including the kernel run in a single memory address space (i.e. no concept of rings or protected mode)?
It will be a potential security nightmare. I was quite impressed and excited until I saw that news about the memory map. :(
We used an older version of vxWorks in my previous job that ran that same way, and just chasing bugs down (that crashed the system hard or even overwrote the onboard flash memory - due to infinite recursion) was a nightmare. I believe current versions of vxWorks have proper memory management/protection....

I used OS/2 1.3 - Warp between 1993 and 2000 both as a user and a developer with it as a target platform. Although at the top when I switched to Warp (1994), it was streets ahead of Windows (with the exception of NT 3.1/3.5 - but they had heavy resource consumption for the time!), there were still major problems:

1) The SIQ - Truly horrible - just as for Windows 3.0/3.1, it was just far too easy to get the whole system to lockup (basically all PM based apps used a single system input queue, thus if any blocked for long........)
2) Hardware support, though much improved with Warp was still very iffy, especially back in the days of OS/2 2.1, I remember setting up the netware drivers on my desktop - sheets of typed up A4, lots of config.sys hacking etc.
3) Even back then, the moment Windows 95 appeared (irrespective of it's technical merits), the GUI LOOKed ugly compared to Windows 95s.

It was fast and efficient though, I'll say that for it - a kernel written in assembler, rather than C, but that was probably the very same reason that it was inherently non portable apart from the briefly seen PowerPC version and the briefly living OS/2 2.1 SMP ("Special version"). I don't believe they even supported SMP on anything except that OS/2 2.1 build (i.e. they dropped it again for OS/2 Warp 3 and Warp 4 - maybe I'm wrong).