I would argue the point about the kernal needing a rewrite.
You would, or you are?
The nt 5.1/5.2 kernals(xp/2003) are stable, reliable, and known quantity. Aside from being 7 years old, what do we really lack with the XP kernal?
MS-DOS was very stable. Why did we move on? As Crash points out, stability is not a terminating value. It is a necessary value.
I dislike kernels that have more code than I require to run my system, and I dislike any system that requires rebooting for non-kernel code.
I have several reasons for my criticisms.
First, the dangers of placing all of your code into the kernel address space has been very well documented. The ease with which Windows OSes can be hacked can be laid in part at the feet of the design of the kernel and kernel-space. Buffer overflow exploits should not be so easy!
Contrast this with the linux kernel, whose address space is much more tightly regulated, and you can see where I find my objections.
Second, MS' kernel is not modular and this is just plain silly for a pre-compiled kernel. Unlike any open source product, you can't compile Windows to run optimally on your system. You are handed a binary and told 'this is Windows' and the code is exactly
the same whether you are running a quad-core 64-bit Phenom or a 1Ghz single-core 32-bit Duron with a fraction of the on-die cache.When you combine that with the fact that you can't load modules you require or unload modules you do not require
, and you don't have a modern OS in my opinion.
Imagine how much faster Crysis would be if you could strip the kernel down to just what it needs to run the game at run-time and then re-load the modules you need for web surfing when you want to brag about your score. Imagine how much faster your PC experience would be if you could recompile your entire system so that it is optimized for your hardware.
IT is a known quantity, reasonably secure(and it depts know how to secure it further), compatable, and works with new hardware.
It will be replaced in 3-5 years, as every OS is, so this is not an argument worth presenting.