Re: [RFR] Press release about Debian on public clouds
Le Fri, Mar 15, 2013 at 02:16:44PM +0100, Stefano Zacchiroli a écrit :
> On Tue, Mar 12, 2013 at 02:26:50PM +0100, Francesca Ciceri wrote:
> > attached the draft of the upcoming press release about Debian on
> > public clouds: it'd be great if you could review it. Particularly,
> > I'd need two things:
> >
> > 1) the correct links for Amazon and Azure images
> > 2) a quote from representatives of both Amazon and Azure.
>
> Thanks everyone for the feedback given in this thread. Based on it, I've
> prepared a new draft, which I hope could be consensual.
>
> If you have further comments, please consider joining patches to them, so that
> the feedback loop could be quicker (i.e. you wouldn't have to wait 2 days for
> me to come up with an alternative proposal that addresses your comments :-)).
>
> Note that there are also very technical points that anyone on this list could
> help with, e.g. the most appropriate image links.
Hi Stefano,
I would like to express my dissatisfaction again, without friendly words as it
seems that you are determined to push a press release without asking first if
it would be helpful or not. This press release comes completely top-down, with
no public record to figure out who thought that now is the moment to send a
press release, and more importantly: why ?
If that press release gives us bad reputation of releasing half-baked images,
the damage will be bigger for the people working on cloud images than for the
press team, who will always find other projects to communicate about. The
do-ocracy model does not apply there. Regardless if we send patches, your
press release impacts our work.
So please discuss first before pushing public communication with tight
deadlines, and consider whether there is a common benefit for sending a press
release now.
As the maintainer of the euca2ools package, please let me remind the current
situation there:
- We consider Free our toolchain to produce AWS images because we can use
euca2ools as a drop-in replacement for Amazon's ec2-api-tools.
- Its version in Squeeze is completely outdated. I do not think that we can
use it to produce AWS images.
- There is and there will be no backports for Squeeze, and backports-sloppy
is not open yet.
- In Wheezy, euca2ools is not very fresh as well. I upload new upstream
releases to Experimental because the use of Unstable is discouraged
during the Freeze. Packages.debian.org indicates packages in Experimental
to be "rc-buggy" no matter their bug status, so I do not expect casual
users to try these packages.
- Only after python-boto's upstream publishes a new release, we will have
a Free solution to deploy our AWS images accross availability zones. For
the moment, we depend either on James to do it on the Amazon side, or
we can use the latest non-Free ec2-api-tools since a couple of days.
Altogether, if our users want to use a Free toolkit to interact with the EC2
from our EC2 images, they will be quite frustrated. Equally frustrating will
be their experience with the images, that are work in progress. For instance
they do not offer facilities such as cloud-init, commands such as 'killall'
are not installed, instance storage is not available by default, ...
All of this is work in progress, we are getting towards having much better
images. So I wonder, why sending a press release while we are at the middle of
our work ? Why not asking first if we feel ready and discuss on what would
be the best timing for both us and the press team ?
First contact with a new product is very important. Whatever we write,
whatever patch we send you, what people will remember is that Debian produces
minimal images that require extensive adaptations, and it will be difficult to
correct this reputation later (think about the bad reputation of
network-manager for instance).
Maybe I overestimate what "Press Release" mean, but I think that for the
moment, debian-devel-announce would be better if the goal is to call for help.
But to be blunt, I think that the biggest blocker is not the lack of manpower,
but the lack of Wheezy release, so if it can be accelerated by press releases
or posts on debian-devel-announce (for instance by indicating which bugs are
total blockers as opposed as bugs solvable by removing the affected packages),
that would have a stronger impact on the work on could images.
Sorry again to be unfriendly, but I hope that it helps to be clear.
Cheers,
--
Charles
Reply to: