Canonical Upstream Contributions WAS: More than half of Windows machines are INFECTED with malware

[Samuel Thurston, III]

Since it seems windows has left the discussion it's time for a topic change.

On Wed, Oct 7, 2009 at 2:30 PM, David Gerard <dgerard at gmail.com> wrote:
> 2009/10/7 Samuel Thurston, III <sam.thurston at gmail.com>:
>
>> Is the desktop Canonical's focus?  the X.org contribution factor would
>> be relevant here... Canonical is almost non-existent in this space.  I
>> don't know how much contribution Canonical makes to the gnome project.
>>  Picking desktop apps that work well together and setting them as part
>> of your distro's default installation doesn't strike me as all that
>> much of a community-oriented contribution.
>
>
> The important thing Ubuntu does is to regard a new user having
> problems as being a reportable bug. Contrast to Debian, where the (IMO
> appropriate) response is more likely "RTFM."
>
> Then there's things like the 100 Paper Cuts project. The business of a
> distro is the whole stack. Making it all work together really nicely
> is very important and not easy, and Canonical does spend its money
> doing just that.


Heh, I agree with that.

But, if you look at other commercial support companies and their
distros, e.g. Novell/SuSE & Redhat, it appears as though there's a lot
more of the commercial money & energy being poured into the
development effort.  Those guys do a nice job of integration and
polish too.  They also have strong community efforts that go upstream.

I guess what I'm not sure about is where the line between the Ubuntu
efforts and the Canonical efforts is drawn.  Maybe Shuttleworth's
money is flying at Ubuntu, but the resultant contributions are not
being counted as part of the Canonical effort because the contributors
aren't actually on Canonical payroll.

>
> When someone asked for more lovin' for Kubuntu, Mark Shuttleworth did
> say he spent a considerable sum every year on it. (Can't find cite,
> sorry.)

Don't worry, I accept your proposition uncontested. :)

>
> I think Greg K-H missed the important point: while layering is a good
> thing, being a jobsworth and deprecating all  tasks except your own is
> not. Look at his Linux Driver project, where they say "we support all
> the devices!" and when users say "but my stuff doesn't work", they
> answer "well, that's userland." That's the start of the job of
> supporting the devices, but it certainly is not finishing the job of
> supporting the devices. The kernel is important, but it's not anywhere
> near the whole thing.

No, but face it, without the kernel you've got an awfully tough row to
hoe.  The other commercial interests at play in the space are
obviously throwing some serious resources at kernel and other
"plumbing" efforts, upon which the rest of the beast sits.

To put it another way: I am by no means a kernel expert, but normally
to build a custom kernel, i go download 2.6.31.2 (current stable) from
kernel.org, jump through the normal set of hoops (make xconfig, etc.),
and reboot. If i do this on my 9.04 testing box, a bunch of stuff will
be broken  because the repository sources from which the stock kernels
are built have a set of "ubuntu patches" applied.  unless i can figure
out from my stock .config what those patches are and find 2.6.31.2
compatible patches for them, I'm going to have problems.

The same goes for applying external patches to the repository sources:
because they're already patched, adding in foreign patches creates
problems, unless they explicitly support ubuntu.

The response from people on the Ubuntu forums tends to be "don't worry
about it, just use the stock kernel"

But my question is: why aren't these patches being taken up by
mainline? Are they not being submitted, are they submitted but
rejected, or is it just a matter of timing?