(off topic) Greg KH talk

[Philip Edelbrock]

I had the opportunity to sneak out from work early to watch a talk that 
Greg KH gave to (mostly) undergrads at the local university (Oregon 
State U).  Mostly it was about the evolution of the kernel development 
and maintenance when things were moved to the bitkeeper system.  Also 
how there may never be a formal development kernel (aka 2.7) unless 
some very hairy, complex, yet desired patch is submitted that can't be 
introduced incrementally.

His slides of this talk are available here:

http://www.kroah.com/linux/talks/oscon_2004_kernel_devel_talk/

His young daughter handed out copies of instructions on how to submit a 
patch to a maintainer and he said 'now you have no excuse to submit a 
patch'. :')

He noted that the current system is both very unorthodox yet very 
efficient.  A 'web of trust' has been formed amoungst patch submitters 
and maintainers, which also can be called a 'path of shame' when a 
patch causes problems and is traced back down to it's source through 
the maintainers.  2.2 patches *per hour* excepted in the 2.6.0-2.6.7 
releases.  Linus and maintainers and very happy with the current flow 
of development, and it shows.  No other OS supports as many devices as 
Linux does.


Phil