Debian Common Core Alliance

[Magnus Blomfelt]

I will defend DCC a bit here, as I think it's a good but
not perfect project.
Links are provided on separate lines to be more readable.

Mark Shuttleworth wrote:
> This is simply my view on what will work, and what won't.

> 1. The Name.
Totally agree.
The best thing for DCC would be to become a Debian subproject.
"we are in the process of applying to make the DCC part of
 an official Debian subproject"
source:
http://www.dccalliance.org/faq.html

> 2. The Commonness. The DCC distro doesn't use the Debian kernel, and it
> modifies key pieces of the infrastructure like the linking system and
> core system libraries. So it's not really Debian at heart. There is now
> discussion of modifying many, many more packages, for example to use the
> Ubuntu X.org packages rather than the Sarge XFree86 packages.
Correct but very incomplete.
1. The DCC is not a distro.
   source:
   http://www.dccalliance.org/faq.html
2. The DCC uses a binary compatible Debian Sid kernel with 7 patches added
   and will remain compatible with stable Debian.
   source:
   http://lists.dccalliance.org/pipermail/dcc-devel/2005-December/000406.html
3. The DCC only modifies the linking system and core libraries
   when it is absolutely nessecary to be LSB compliant.
   Debian Etch (the next stable Debian) will be LSB compliant
   and then there will no longer be any need to modify as much.
4. Only Debian can be Debian at heart, neither DCC or Ubuntu.
   That is unless DCC becomes a Debian sub project.
5. The X.org packages in DCC are backports from Debian Etch.
   source:
   http://lists.dccalliance.org/pipermail/dcc-devel/2005-December/000397.html
   The DCC does not contain the whole of X.org, only the parts required for
   LSB certification. This makes it possible for the individual distributions
   to choose the X server to use.

> 3. The Style. The DCC distro is really aimed at shoehorning an
> LSB-compatible environment on top of Sarge. I think if LSB was a goal
> for Sarge then that could have been achieved directly in Sarge, not in
> this hybrid fashion. If we were to do LSB for Ubuntu, it would be done
> directly rather than as a compatibility layer.
LSB wasn't a goal of Sarge (the current stable Debian)
but it is a goal of Etch (the upcoming stable Debian).
Jeff Licquia of Progeny have been collaborating with Debian on
the experiences and code from DCC.
source:
http://lists.debian.org/debian-lsb/2005/09/msg00008.html
Since the DCC follows Debian stable it will eventually implement LSB the
same way as Etch will and the difference between DCC and Debian's core
will become smaller and smaller.

> 4. The Premise. The vision behind DCC, which is indeed compelling, is
> that it would provide a common platform for certification, and that the
> distros that make up the DCC would all ship exactly that same core. But
> it strikes me that this approach has never worked in the past. In fact,
> every distro ALWAYS modifies elements of the core, and with good reason.
The reason distros modify the core is because they don't see a reason not
to. There now is a reason and a lot of distros have chosen to use a common
core and add functionality as modules instead.
1. Of the 25 most popular Linux distributions on distrowatch.com
   - 9 of 25 are based on Debian
   - 4 of 9 Debian based distros are in the DCC Alliance
2. Progeny released their DCC-based distro in 2005 and LinEx, Linspire,
   Sun Wah and Xandros will release their DCC-based distros in 2006.
   source about Linspire, Progeny and Xandros releases:
   http://lists.dccalliance.org/pipermail/dcc-devel/2005-December/000397.html
   source about LinEx release:
   http://lists.dccalliance.org/pipermail/dcc-devel/2005-December/000415.html
   source about Sun Wah:
   http://lists.dccalliance.org/pipermail/dcc-devel/2005-December/000408.html
3. A lot of distributions outside the DCC Alliance are also planning to
   base their systems on DCC and some even wishes to join the Alliance.
   1. Symphony OS, which has a DCC based beta:
      http://www.symphonyos.com/forum/index.php?showtopic=460
   2. the German company bitbone:
      http://lists.dccalliance.org/pipermail/dcca-discuss/2005-December/000002.html
   3. the Portugese distribution Alinex:
      http://lists.dccalliance.org/pipermail/dcc-devel/2005-November/000384.html
      http://lists.dccalliance.org/pipermail/dcc-devel/2005-December/000427.html
   4. Muriqui Linux, currently based on DCC member Progeny seems interested:
      http://lists.dccalliance.org/pipermail/dcc-devel/2005-December/000480.html
   5. Linux Loco, based on Progeny and gnuLinEx will most likely be DCC
based as well.
      http://distrowatch.com/table.php?distribution=loco
4. About 50 distributions are based on Knoppix, and it's reasonable to
beleive some
   of them will be DCC based.
None of the mentioned distros seems to have any problem with the premise.

>And while we would love that not to be the case, the truth is that the
>reasons to specialise outweigh the benefits of homogeneity. Much as it
>may be compelling, the common core idea is ultimately flawed, because
>it's not just the bits at the core that matter. An ISV that certifies an
>OS is not just certifying the pieces on which it's app depends. It's
>also certifying the *environment* which it will support, which includes
>installer, documentation, even packaging. Screenshots, for example,
>won't be useful unless they reflect all of the certified OS's, and
>that's not possible in the DCC model.
Of course it's possible, the DCC model in fact improves the situation for
ISVs.
If they want to certify to the whole OS, they first certify for DCC and
then to
the other parts of that OS. When certifying for the next OS, the DCC part is
already done. Of course not only certifying gets easier, but also developing
and bug fixing.

>There have been several examples of places where this idea has failed.
>United Linux is one.
The idea failed because there were some differencies the members couldn't
agree on. By choosing Debian as base these things are already sorted out.

>Now, that said, it's great that Debian-derived commercial distros have a
>forum and a mailing list on which to discuss problems they run into, and
>I think the DCC project has served as a good touch point for bringing
>some of these hidden issues to the fore. So I'm in no way opposed to the
>DCC or its members, and in fact there are some neat places where we are
>collaborating with the DCC on source code, which is where I think the
>collaboration works best. The DCC kernel and Ubuntu kernel will be very
>similar if not identical in future DCC releases, and I expect that
>collaboration will spread to other parts of the system such as X, ACPI etc.
This is great news and I hope this source level collaboration will continue
and extend.

>The essential driver of free software is collaboration to achieve common
>goals. And that works best at the source code level.
The DCC Alliance seems to be aware of this and has therefore decided to
only collaborate at binary code level where it is necessary. Having a common
binary core does not impose that many restrictions on a Linux distribution
and it can still be bleading edge. The DCC will be up to date by providing
driver and security updates that do not break binary compatibility.
Distro makers can add their own functionality to the kernel by using modules.
When it comes to the desktop environment every distro need to adapt it to
their kernel and libs anyway, so simply using the DCC won't make it any
harder.
Having a common core, in my opinion, makes it easier to collaborate at the
source code level.

>Inasmuch as the DCC has promoted better collaboration between its members,
>it's a success.
>Mark

I completely agree with you on the name issue and hope it will be resolved.
On the other hand the commonness, the style, the premise, the certifying
and the
possible collaboration failure are no issues to a growing group of
distributions.
If you still think they are, I would be really interested in reading about
your further comments. I really believe in Ubuntu as it is my primary OS and
think both Ubuntu and the DCC Alliance would benefit from Ubuntu's
participation.

Magnus