Re: Timeline for official Debian cloud images for the Stretch release
On Sat, Mar 26, 2016 at 12:29 PM, Charles Plessy <firstname.lastname@example.org> wrote:
> Hi Richard,
> here are quick answers, but I will then run out of time and I am not sure
> to be able to continue the discussion on that pace in the next few days
> (which may be a good thing).
> Le Sat, Mar 26, 2016 at 08:48:22AM -0700, Richard Hartmann a écrit :
>> all teams are well within their delegations. If the DPL had felt that
>> to not be the case, I am sure he would have updated them accordingly
>> or created a new one.
> I will not second guess the wills of the DPL, and anyway, a new DPL will be
> elected very soon. If you think that formal management is necessary (and I do
> not), can you ask him for setting up a delegation ? If not, can the trademark
> team refrain from nominating managers ?
>> Our main concern is that we need a way to protect our users from
>> unexpected behavior in a somewhat automated and scalable manner, and
>> keep those results and their records out in the open.
> The usual way is to report issues, open bugs on the BTS, and work on a
> solution. Sorry, but here I see at worse fears, uncertainty and doubt, or at
> best ambitions to have a more unified system, but no concrete problem or
> complains. You are using a very heavy hammer, without having participating to
> problem reporting and problem solving before on this mailing list. And
> ambitions are great but somebody's ambitions should not become somebody else's
>> Your main concern seems to be that you will be forced to conform to an
>> arbitrary rule-set enforced top-down.
> My main concern is that what is being done is a strong departure from what I
> felt is a core Debian value: the principle that the one who does should usually
> be the one who decides.
>> As a specific proposal in light of your concern, might I suggest you
>> simply join the discussion about the test set?
> Sure. Unless I am mistaken, it has not started, isn't it ?
> There are also discussions that did not yet lead to a conclusion, in particular
> on the use of cloud-init backports. I am not happy to see people comming out
> of the blue and informing us that they self-appointed as decision makers on
> issues that include whether we can use a backport or not. We have been doing
> this for a while, and if people have a different opinion on what should be
> done, the best way forward is to step in the maintainance of the cloud-init
> package (in coordination with Ubuntu, which is our Upstream here), and work out
> how to get Stable updates, etc.
> To summarise it with a strong image, you come to a discussion with a gun, and
> after aiming at peoples heads, you say "let's do some great work together". If
> you are interested in the cloud images, can't you participate as a regular
> developer, rather than as a person who has the power to pull out the plug from
> other people's projects ? I feel that we are treated as potentially dangerous
> contributors who need to be controlled to prevent some mess to happen and
> damage Debian's name. This is highly demotivational.
I am sending this email as someone who's been involved with
debian-cloud since the beginning, and also as the first volunteer to
join what is now the Debian TM (trademark) team.
I'm going to give a quick history, so you understand where we came
from, and what we are trying to accomplish. It's not at all dire.
Initially the TM team was formed to offload some routine TM work from
the DPL, who had been handling all TM work alone. (mostly related to
answering emails.)The TM team didn't have any authority for approving
TM requests, as the team wasn't a standalone delegated team.
As such the "approvals" for the AWS and Google cloud images were made
by the DPL at the time, who consulted with the newly formed
debian-cloud mailing list. This list was and still is a great way to
allow experimentation and collaboration in building a variety of cloud
images by a wide ranging group of involved parties. This early
experimentation paved the way for a more "official" process and the
set of standards that we are working through now. That said
debian-cloud, isn't really a traditional Debian team, and is more of
what I would consider a shared interest list/group. I think it works
best this way, as we are very open to experimentation, with a low
barrier of entry for new people to come and start hacking. This list
is also a *GREAT* place to get a wide variety of interested opinions
related to all aspects of building cloud images. (Whether technical,
philosophical, or other.)
Eventually the TM team became a standalone delegated team on Feb 17,
2015 . As of that date, the TM team had the primary responsibility
for making decisions regarding Debian Trademarks, that previously had
been made by the DPL.
This meant, when folks wanted to publish "Debian blessed" images for
Microsoft Azure, the TM team was involved in the approval process,
rather than the DPL.
A number of similar requests came in at around the same time, and it
became clear that we needed some sort of policy to guide these
Ultimately we wanted the following:
1) A Debian project policy for official cloud images, that we could
point people requesting TM approval to.
2) Policy creation must be done in public, and include the input of
all interested parties. (debian-cloud, cd-team, and others)
a) debian-cloud, because this was where those doing the work to
date had been collaborating
b) cd-team because they were the existing team that indisputably
shipped "Official Debian" "images", and they had recently announced a
desire to support the building of "official" cloud images.
3) Policy enforcement ownership by a technical Debian team that would
be responsible for quality control of Debian "product" (Trademark team
can not and should not be responsible for technical work or policies,
and should be focused on the legal end of things.)
It is our hope that the people working on the AWS and GCE images now,
will participate in this discussion of standards, realizing that what
Steve has proposed, is simply that, a starting point for discussion.
(Any discussion needs to start somewhere.) We also expect that the
great work being done now, will continue to be done by those doing it
now. (Perhaps even as new members of the cd-team.)
I'll be happy to continue this discussion on or off-list, if you have
further concerns or questions.
 - https://lists.debian.org/debian-devel-announce/2015/02/msg00011.html
> Charles Plessy
> Tsurumi, Kanagawa, Japan