I suggest syncing Linus's current tree up with VGER and moving
everything to VGER right now and until 2.0.0 is released. When 2.3
development starts, maybe Larry's new system will be up and running and
Linux development can be moved to it. If I remember right, it will be
like CVS with a "permissions tree" where change sets "flow" up the tree
through a series of increasingly-trusted-by-Linus folks (I am assuming
that at least David Miller and Alan Cox would sit directly under the root
node which would be Linus) and eventually when the changes were vetted by
Linus they would go into the kernel proper.
Let me use GGI as an example of the way I envision this
tiered-authority thing being organized (all names are examples only of
course):
Linus (whole kernel manager)
Alan (sound, video, networking manager)
Martin or whoever (video and console manager)
Geert (fbcon/fbdev manager)
Steffen Seeger (KGI/kgicon driver manager)
Jan Kneschke (S3 driver-family manager)
Teunis Peters (S3 Virge driver manager)
David (SPARC, etc manager)
xxx (arch/sparc manager)
yyy (sparc32 CPU support)
zzz (sparc64 CPU support)
...etc...
You get the idea. By the time any change sets percolated all the
way up to Linus, they would have had their design and stability vetted by
many people and Linus would be able to toos them into the kernel (if they
were uncontroversional) without even looking at them. The problem right
now is distribution of authority. Linus needs to (and most of us *want
him to) be able to keep his final say over what goes into the kernel, but
at the same time we need a way to take the load of scanning through every
little bullshit change to a buffer length in a SCSI driver somewhere off
his back. If Linus could take fewer small change sets from fewer, more
trusted people less often, I think his sanity would come back.
Another benefit of this system would be that anyone could work on
anything they wanted to, even if Linus didn't care for what they were
doing (like the GGI project) and their work would still be in the system
and easily accessible to others. As their work progressed, the parts of
it that gained acceptance would rise faster. Thus, the collective wisdom
of the Linux developer community would be harnessed into a multi-level
quality control filter. With Linus as the last filter in the pipeline.
Isn't that pretty much what we all want?
Jon
--- 'Cloning and the reprogramming of DNA is the first serious step in becoming one with God.' - Scientist G. Richard Seed
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/