A recipe for coming up with good ideas is looking at something that has been around for a while and asking "What would it be like if it had been invented now?".
And if all else fails, this will be a wonderful learning experience.
I know this is a cliche, but before writing comments like this ask yourself if a) you are unnecessarily discouraging somebody for no reason and b) you would have said the same thing about successful projects when they were first started.
Finally, the one piece of advice I'd give on this project is that you shouldn't focus too much on the overly ambitious goals and don't worry if you can't accommodate for all of them just yet. The likelihood of achieving the next milestone is somewhat proportional to the number of milestones you have achieved. Get to v0.01 and take it from there.
a.) No, I don't think so. Considering the number of half-started kernels on OSdev.org, the creators of which incredibly often were convinced that they were going to create the newest awesomest kernel everyone is going to use, trying to get the OP to take a good look at his skillset seems like a proper thing to do.
b.) I would have, if the creators seemed completely ignorant about the basics of what they were trying to do.
This is not to say that I would tell the OP to give up on OS development - I personally love it. But I think a serious reality check is in order, and because we get people similar to the OP on the OSdev forums an awful lot, sometimes this frustrates me.
Unix
and asking "What would it be like if it had been invented now?".
creat(2) would have the second 'e'.
I would have mentioned it, but I was speaking as if Unix were invented today. Plan 9 is an almost textbook second-system effect; it seems like you have to have invented a Unix first before even starting on something like Plan 9.
Doomed = not commericailize-able? That doesn't mean its without value to the person behind it if they learn and derive satisfaction from the experience.
With the obvious caveats that you write as much of your code to be cross-platform as possible, and be careful with endianness.
What does Microsoft use?
Nothing is more flexible that command line tools as it lets you do scripting(programming your own environment).
It does not matter if you want to use "already made by someone else windows, text, menus and buttons", but it is essential when you make (or manage what other people's make) everything.
However, my impression is most of them edit code in MSVC, build with the DDK tools, and debug with windbg/ntsd/kd like most other driver developers.