Published at LXer:

"The news of Con Kolivas, a Linux kernel developer, quitting that role, along with an interview in which he explains why, could and should make loud noises around the Free Software community which is often touting GNU/Linux as the best operating system one could use, and not just because of freedom you have with it. In the interview he says certain things which should cause tectonic shifts in the mindset that we have all been having. Why didn't we realize these things before?"

Read More...

Well that's interesting indeed.

but I'm not very sure whether the "poor" performance of Desktop programs is related to the kernel itself.
the desktop is nothing but an interface between the user and the kernel so I suggest that the Desktop software and everything that's related to it has to be enhanced and not the kernel.

Furthermore , I think that "forking" a new kernel would be utter suicide , because one has to take into consideration that 90% of the computer world runs windowze which means that 90% doesn't even know what linux/unix even is ...and that's basically because the linux world could be more compared to a spaghetti jungle.
if we look at the current linux world we see a serious level of over-spreading because of so many distros being around one just gets confused . because at the end not everybody wants to waste time learning how to do simple tasks on each and every other distro .
unification is needed not a continues and self-destructive thinning of the community!

so my conclusion is that another fork would be a turn off for developers and users as well and a death sentence to *nix as a popular OS. (isn't that exactly what microsoft wants? linux to be pushed to the margin?)

Cheers

The question of forking the kernel should not be decided based on its effect on Linux's popularity, but rather whether or not it could truly improve performance. In Kolivas's interview, he admitted that it was difficult to evaluate whether or not any specific patch of his made a significant difference.

I do think that a forked kernel which addressed desktop considerations first, included real time preemption, and trimmed server specific elements would be an interesting project. Rather than a "turn-off for developers," I think it would generate a healthy competition which would improve a lot of things.

Granted, it would probably further confuse outsiders who don't understand why there are so many distro and desktop environment choices, but those choices are the life blood of Linux OS development. Any innovation that proves it's value on one platform quickly spreads to others as well.

I don't think I'd like to see a fork. Of course, this is excellent criticism though, and hopefully it will spur some change in the kernel development model.

Why would he need to fork the kernel? There are already dozens and dozens of "forks" on the GIT archive at kernel.org and he could just create another one. I'm sure if he created a trimmed-down and enhanced desktop kernel that is based on the code of the current kernel, it would attract more than a little bit of attention.