Yeah. And I managed to take it to the next level with dynamic executable loading and SVGA drivers.

But why not join forces with cosmos-instead of many half done OS we could have one functional? Were there any particular reasons?

For couple of years I contributed to Cosmos and fixed some bugs in MOSA. But my design idea was much different from them. So I decided to start with a new project. And seems like it worked!

Can you elaborate on differences?

Back when MOSA got created, we were aiming for a pure approach. For example, we were trying to use as few intrinsics as possible and a mostly architecture agnostic approach.

Furthermore we also focussed on Mono compatibility (that was way before CoreCLR came out).

Source: Ex co-creator of MOSA

I wonder whether now it can switch to core to native compiler (supposedly soon available)

This "why not join forces to have one good one" applies to all those "let fork it and do it better" projects....like systemd, KDE, Gnome etc (yes, they are not all forks, but you get the idea). This is one drawback of open source but also its strength....

True, but those are functional systems. While we don't have a serious C# OS.