To implement some of my ideas properly, you'd have to start with another OS from scratch. But this is not always bad anyway. Linux is still gathering a following. From my personal experience, it seems that it has mostly gathered followers from two camps, the anti-Microsoft league for one, and the "techie" that finds the positives outweigh the negatives when it comes to system administration for another. Something to bear in mind is that really good operating systems can be ignored by the public. BeOS (at least when it first came out) was an ultra realtime OS, really good, and cross platform... yet still didn't take off. AmigaOS (the new one) probably won't take off either. I think this is because they don't offer enough difference to existing OSs, and they don't let you carry on using the old OS at the same time (at least not simply), and even if they do (AmigaDEV), they just don't give anything back for being involved... no killer apps. nor any killer ability.
So, if you haven't got any killer apps, and you aren't 100% compatible with existing apps of some sort, then you've got a great big steep hill to climb. You aren't gonna be able to catch up with the leaders that without a boost, as the leaders are still climbing. This problem is insurmountable (heh), so don't try.
Ok, given these theories, why has Linux been reasonably successful? Two reasons, it has killer apps, and it was born of the original operating system, so there were thousands of people already trained for the OS.
Killer apps? what killer apps?
Linux is based on server code, so you can say that Linux's killer apps are things like SMTP hosting daemons, generally better http servers, and remember at the beginning, who was it hosting all the web content? unix machines, not windows machines. Unix machine users (typically the geeky and the admins of websites and ISPs), were very quick to adopt the ludicrously cheap Linux, because they knew what they were getting. Other people who had started to hate Microsoft started to move away from Windows, towards Linux, but the difference was normally to great, so they sprang back after initially dipping their toes in the *nix waters. I'm afraid I'm one of these cowards. I stick with windows because its what i know, not because its good. I know myself that if anything goes wrong with a windows machine, i can usually fix it. Its easy to fix because its a simple system. *nix is more complex in appearance to me, so i don't trust myself to be able to fix my machine in times of trouble. That's why i don't move. There two apps I love that are entrenched in Windows (or Mac): Microsoft Visual Studio .NET, and Adobe Photoshop. Obviously, games on Linux are also a little limited in comparison, but I am a software developer, so .NET is actually more important to me.
Anyway, basically, Linux is creeping in because it can't be got rid of, not without losing quite a considerable amount of productivity amongst network "people".
So, as long as "Some" people can't live without your OS, you can get enough of a foothold to start creeping in. If you have overlapping killer apps, then there shouldn't be a problem with migration, and if you actually have some killer features, you should be able to "clean up", even if you're incompatible with other OSs.
So, can my ideas "clean up"? I shall run down the list of things that make up my OS.
Are my ideas windows or Linux compatible? No for windows, they could be for Linux. I haven't the foggiest how long it would take to write a windows compatible version of my OS, but i know it would take a very long time. To make my OS Linux compatible would be a lot simpler, and for most of the ideas, the OS would be possible to implement IN Linux. A mate told me that it might be possible to rewrite the filesystem, the kernel, and a window manager of Linux, so it behaved exactly the way I want it to. I hope it is true.
Are my feature ideas "killer"? They may be killer in fact, but i don't think they appear to be killer, the idea for a user interface i have might be killer though, and that might also be enough to interest a migration ready crowd. The database filesystem idea is bandied about too much to be taken seriously, and file instance applications sounds strange, not brilliant. Document centric editing could be a good start, but that is not the OS, but the way the Apps are written.
Does my OS deliver a necessity? No, obvious really, but unless the apps are built to complement my OS, even the basic needs will not be met. If some basic applications are built, Text editor, Image viewer, Compiler, and a Media station (WMP/WinAMP), then it could build from that, from them. Maybe given some time we could develop the killer apps necessary. These killer apps would probably be impossible to build outside the OS, and all the better for it.