Ok, well the thing is you guys seem to be a bit over-paranoid. I'm not going to "reinvent the wheel", like I said I'm going to be build upon the Linux Kernel, or at least thats my intention. For example at the end of my learning I want to be able to develop an an OS for an HTC or for my Laptop USING THE LINUX KERNEL of course.
That makes sense. Sceadwian conflated "OS" and "kernel": Linus T. created the Linux *kernel* (not by accident) but Richard Stallman and others created the rest of the system which, combined with the kernel, provides an OS. (I'm sure you already know all that
). And that's not even getting into all the distros on top of the base GNU/Linux OS, and all the OSes built on the Linux kernel which aren't even GNU/Linux at all.
I think cr0sh asked a lot of good questions, the answers to which will probably be important. Pending those answers:
I started with Slackware around 1994/95 and it wasn't really that bad. Of course I was in university in Finland studying comp. sci. at the time so that probably helped as everybody I knew was pretty much doing the same thing. As far as what you're doing, Slackware wouldn't be a bad place to start, actually.
Have you considered using a VirtualBox or similar VM environment to host your tests? That would be safer than subjecting the actual hardware to a new system. Especially the first few times you (try to) boot it up.
If you're not familiar with initramfs or similar, and the boot sequence in general, you probably should get familiar with it. Much will be dependent upon the actual hardware you are wanting to support, though. Also, something like
RedBoot would be good to get familiar with.
Have you messed with custom/minimized Linux installs on embedded hardware before? The one I'm most familiar with is the system used by the
TS-7400 module, but there are others. Might be a good idea to take a look at how they do things. And check out Android's architecture.
What is the ultimate scope of your plan? It's very important to decide on this--and to spell it out in painful detail just so we don't make unrelated suggestions.
I also agree with cr0sh that you probably should ready & study LFS. You said that you don't want to "spend a month reading it", but then you went on to say that you understand the project will take a significant amount of time. I'd just like to point out that laying a good foundation in your understanding is not a waste of time at all--and compared to the amount of time you will spend on the project as a whole, a month is likely to be a vanishingly short period, and thus well worth it. You'd probably waste well more than that figuring out all the stuff the book would have told you to start with.
Anyway, that's my $0.02 CDN. Without a LOT more information, there's really not a lot of concrete advice that can be given.
Cheers & good luck,
Torben