On Mon, 9 Sep 2002 18:31:31 -0500 (CDT)
Daniel Taylor <dante at plethora.net> wrote:

> Patches are applied at the whim of the core developers and Linus.
> 
> Dave Miller and Alan Cox are pretty good at tracking meritorious
> patches and helping keep them on target, but there is a definite
> barrier to getting patches into the kernel.
> 
> In some ways this is a good thing. I've seen more different scheduler
> patches fall by the wayside because nobody understood them, including
> the original authors. In some ways it is not so good, as the number of
> interface changes in recent years has begun to outweigh functionality
> improvements.
> 
> I think the model is due to fall apart in the next few years, I don't
> really know what the current state is as I got fed up and dropped out
> about 4 years ago.
> 

If this is the case, and to help to boost performace better, wouldn't it
be easier/better to develop more towards today's processors/architecture
than to keep "legacy" systems in as well?  Just a thought...

Shawn