[Mageia-sysadm] ML request for QAteam

[Michael Scherer]

Le jeudi 31 mars 2011 à 21:33 +0200, Romain d'Alverny a écrit :

> > I think we can all agree that the canonical list of member of a team is
> > in a ldap group for all teams ( because that's what we did so far for
> > web application integration, catdap integration, sympa integration,
> > epoll integration, etc ).
> 
> Yes, but we discussed (several months ago) too to remind that it's not
> because one is (or is not) a team member that it would necessarily be
> subscribed to the team list; that should not be a bijective
> relationship.

Yup, so we agree that this 2 different things.

> > So if the qa team is the list of people in the ldap group mga-qa,
> > it is confusing to refer to a 2nd group ( the list of people who
> > subscribed to the ml, and who would be ) as qateam.
> 
> There's the informal team list, where non-peers can subscribe and
> participate. That's the mailing-list (communication/collaboration
> goals).
> 
> And there's the formal team, made of peers only. That's the LDAP group
> (credential/recognition goal).
> 
> These are two distinct sets that don't match exactly with each other.

Yes, again we agree on the fact this is 2 different things.

> >> (yes, it comes out too from years of practice of this naming scheme at
> >> Mandriva where, well, these were internal list names/aliases - but
> >> here it would be public as team's lists are :-p ).
> >
> > On mandriva qateam alias/ml, who was behind  ?
> >
> > While I was not here so I am maybe wrong, I suspect the aliases matched
> > exactly the internal organisation.
> 
> Far from that. Cross-teams information

That do not really fully answer to the question, ie who was there ?
And how was the teams named in Mandriva ?


> > Here, I do not know exactly because Damien ask for a public list, and to
> > me that mean a list were anybody can subscribe and post, but maybe he
> > mean a list that anybody can read ( something like council ).
> >
> > So in the hypothesis that the list is not restricted to members of the
> > team, the name would be misleading.
> 
> We have already mageia-webteam at mageia.org. That's the team discussion
> list, and it is open to everyone (lurkers included). Only a little
> subset of its subscribers are team peers. And that is perfectly fine.

This one is more a exception than the norm if I look on
https://www.mageia.org/mailman/ 
( and well, I think we also agreed to use web@ when migrating to sympa :
https://www.mageia.org/pipermail/mageia-webteam/2011-January/000124.html ).

In the case of webteam, this is not obvious that the list is a
discussion list when you just take the name, as it sound more like a
alias for the team itself.

> If we were to move the Web team ml to ml.mageia.org (I guess that will
> happen some day) it should not strictly tie ml membership to team
> membership - these are two different things.

So we agree that's 2 different things. 

> > I would also add that it seems that people on the wiki
> > ( http://mageia.org/wiki/doku.php?id=qateam ) have a different view of
> > QA than what is written as a description ( and the purpose is the same
> > as others team like debian one  http://wiki.debian.org/qa.debian.org/
> > and ubuntu https://wiki.ubuntu.com/QATeam/ ).
> >
> > And the ratio of people saying "automation tool" vs "test packages" on
> > http://mageia.org/wiki/doku.php?id=qateam is quite clear ( 1 people for
> > automation, 53 for "testing" ).
> 
> But that should not prevent the QA team to be further designed and
> built for both practices (and still bear the "QA" name, as that's what
> it is about).

I think I was not clear. We agreed that the list and the team are
different. 

The problem is not the qa naming, as it is likely already too late to
change people perception as I explained. But that's not a big issue
because this could be done later with enough education.

The problem is that the name "qateam" already design something different
because of the team suffix.

> Testers are testers, but maybe not necessarily meant to be QA team
> peers that _organize_ and coordonate things - it's not because it does
> not stand out yet that a QA team can't be setup at all as such (look
> at marketing and communication teams - they are not there yet, but
> that doesn't prevent to work for it slowly).

Well, I agree that the lack of formal team is not a issue, that was not
my point ( nor the one that I wanted to make )

But the list of market and comm are not called marketingteam or
communicationteam especially since the team do not exist yet. 

We also said that we would not create 1 list for each team to work in
isolation ( see
https://www.mageia.org/pipermail/mageia-discuss/20100929/001538.html ),
as we want to have cross team collaboration.
And so one of the way to say that a list is not to be reserved for a
team is to remove the "team" from the name.

You didn't comment on my suggestion about distro-test so I suspect that
it is not fine. So let's do a 2nd proposition, what about :

qa-discuss at ml

It say this is about qa ( qa )
It is clearly a discussion list, open to everybody ( -discuss )
It is separate from the idea of team ( which is a separate concept as we
both said ). 

Moreover, qa is IMHO a better prefix for a potential qa-reports@ ( for
automated reports, like sysadmin-reports for puppet/xymon report ), for
qa-private@, or qa-security@, or
qa-updates@ or anything for a private list if needed as reminded by
Damien ( so we can do the vast majority of the work in the open and
still keep sensible issues separate, thus having both benefit ). 

And we could later have qa-announce@, qa-commits@, etc.

-- 
Michael Scherer