Doesn't matter what you're making, OS or not, the choice of language *certainly* carries repercussions throughout a project. Sure Linux is doing fine with C. So what? It could probably be doing a lot better with D.
I'm not saying that the programming language is not relevant at all. Rather, other issues outweigh the choice of the programming language by far. Just take driver support as an example. That's actually what is discouraging most people from using Linux over Window. You can't do anything with an OS that makes parts of your hardware as useful as a brick.
I'd only say that there are some "key" features of a language, that actually matter for something like a kernel. For example, as you said, a kernel written in assembler is very hard to port to another architecture. And you wouldn't write a kernel in Visual Basic (although some folks are writing kernels in languages like Java, C# or even Python).
But D is as good as C/C++ in this regard. when it comes to real life issues, D is probably even a bit worse, because of tool-chain issues. Would your D-OS ever run on, say, PowerPC?
Also, I should emphasize that I never said D would or wouldn't "shake up the OS market", just that the potential was there, whether it be *if* a new OS was built ground-up in D or *if* an existing one was ported. My main point was just that D could certainly improve the overall development process of whatever OS used it, allowing things to advance faster, be more reliable, etc., and thus potentially give it a real leg up.
Maybe. Note that the Linux developers refused to use C++, although the C++ advocates came up with similar arguments. Sure, the languages provide some nice features, which make life easier. But again, what would D help when writing device drivers? Or when figuring out a good locking hierarchy? It doesn't matter that much, you have to deal with much larger problems.
If all the carpenters are building houses with wooden hammers, and Joe Shmoe comes along with his metal hammer, well, he may succeed or he may fail, but he would certainly have that extra advantage, and thus have at least the potential to "shake things up".
That's a bad comparison, because it's very simple to switch a hammer. Also, the metal hammer would be nicer to use than the wooden one, but it'd split in two parts if you strained it too much. Even if you're careful. Some would build complicated, abstract works of art, using a new method called "nail mixin". They'd need at least a dozen hammers until the artwork is finished. The result would blow up in an explosion from time to time for unknown reasons.
