[maemo-developers] Package promoting

[From: Attila Csipa]

On Wed, Sep 22, 2010 at 3:09 PM, Polyvertex <polyvertex at gmail.com> wrote:

>
> I think the maintainer should always have the ability to promote the
> package to extras by himself and take the risk of being under fire
> from users if he promote a very bad/bugged release and did not took
> enough time to resolve issues. This may also be added to the Karma
> system if this is not already done.
>

Generally, that was the setup with the N8x0 and Maemo4.x, so we have tried
that - IMO this works better (occasional developer frustration
notwithstanding) from a user standpoint, so let's not backpedal on that - do
everything to make it simpler for developers, but do not sacrifice users in
the process. Reflashing might be trivial for some, but a serious endeavour
for others.

> A secret weapon is in the making - KISStester, which would allow easier
> > feedback for users, and it also acts as a REMINDER for people who are
> using
> > software that is in dire need of testing (the problem is that people
> > *download* things from testing but forget/can't be bothered to come BACK
> to
> > vote on it).
>
> Well... As both of us said : most of the time, users don't vote. This
> is why the promoting process is way too constraining : it does not
> take users behavior into account at all...
>
>
We'll see the effects when KISStester lands. The point is that ofter people
don't even *know* they are downloading the version from extras-testing
(since H-A-M does not show that info - and they might have installed it
before it was promoted to extras-testing), they are forgetful, etc. So it's
not like 99.999% of users don't care, it's just the process that is awkward
for most. If we as much as double user response by flashing a 'hey, you have
used this app from testing for N days now, care to give some feedback ?' in
notifications (or 'you have uninstalled an app from extras-testing, was
there something wrong with it ?'), that would already be big help. At that
point the biggest concern will be educating people of leaving *good*
feedback, but if we have bad results, we can always disable the vote-part
and just leave comments - that can be very helpful, too.

PS also for feedback - I often see for example problems being reported on
talk - that goes to show that people are not comfortable with leaving
comments within the testing process, so silence in testing does not
necessarily mean all is well (but it does mean the process needs refining).

Best regards,
Attila
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.maemo.org/pipermail/maemo-developers/attachments/20100922/0cecba2b/attachment.htm>