[all][tc] Formalizing cross-project pop-up teams
Doug Hellmann
doug at doughellmann.com
Thu Feb 7 20:27:39 UTC 2019
Adam Spiers <aspiers at suse.com> writes:
> Doug Hellmann <doug at doughellmann.com> wrote:
>>Adam Spiers <aspiers at suse.com> writes:
>>>Thierry Carrez <thierry at openstack.org> wrote:
>>>>Adam Spiers wrote:
>>>>>[...]
>>>>>Sure. I particularly agree with your point about processes; I think
>>>>>the TC (or whoever else volunteers) could definitely help lower the
>>>>>barrier to starting up a pop-up team by creating a cookie-cutter
>>>>>kind of approach which would quickly set up any required
>>>>>infrastructure. For example it could be a simple form or CLI-based
>>>>>tool posing questions like the following, where the answers could
>>>>>facilitate the bootstrapping process:
>>>>>- What is the name of your pop-up team?
>>>>>- Please enter a brief description of the purpose of your pop-up team.
>>>>>- If you will use an IRC channel, please state it here.
>>>>>- Do you need regular IRC meetings?
>>>>>- Do you need a new git repository? [If so, ...]
>>>>>- Do you need a new StoryBoard project? [If so, ...]
>>>>>- Do you need a [badge] for use in Subject: headers on openstack-discuss?
>>>>>etc.
>>>>>
>>>>>The outcome of the form could be anything from pointers to specific
>>>>>bits of documentation on how to set up the various bits of
>>>>>infrastructure, all the way through to automation of as much of the
>>>>>setup as is possible. The slicker the process, the more agile the
>>>>>community could become in this respect.
>>>>
>>>>That's a great idea -- if the pop-up team concept takes on we could
>>>>definitely automate stuff. In the mean time I feel like the next step
>>>>is to document what we mean by pop-up team, list them, and give
>>>>pointers to the type of resources you can have access to (and how to
>>>>ask for them).
>>>
>>>Agreed - a quickstart document would be a great first step.
>>>
>>>>In terms of "blessing" do you think pop-up teams should be ultimately
>>>>approved by the TC ? On one hand that adds bureaucracy / steps to the
>>>>process, but on the other having some kind of official recognition can
>>>>help them...
>>>>
>>>>So maybe some after-the-fact recognition would work ? Let pop-up teams
>>>>freely form and be listed, then have the TC declaring some of them (if
>>>>not all of them) to be of public interest ?
>>>
>>>Yeah, good questions. The official recognition is definitely
>>>beneficial; OTOH I agree that requiring steps up-front might deter
>>>some teams from materialising. Automating these as much as possible
>>>would reduce the risk of that.
>>
>>What benefit do you perceive to having official recognition?
>
> Difficult to quantify a cultural impact ... Maybe it's not a big
> deal, but I'm pretty sure it makes a difference in that news of
> "official" things seems to propagate along the various grapevines
> better than skunkworks initiatives. One possibility is that the TC is
> the mother of all other grapevines ;-) So if the TC is aware of
> something then (perhaps naively) I expect that the ensuing discussion
> will accelerate spreading of awareness amongst rest of the community.
> And of course there are other official communication channels which
> could have a similar effect.
>
>>>One challenge I see facing an after-the-fact approach is that any
>>>requests for infrastructure (IRC channel / meetings / git repo /
>>>Storyboard project etc.) would still need to be approved in advance,
>>>and presumably a coordinated approach to approval might be more
>>>effective than one where some of these requests could be approved and
>>>others denied.
>>
>>Isn't the point of these teams that they would be coordinating work
>>within other existing projects?
>
> Yes.
>
>>So I wouldn't expect them to need git repositories or new IRC
>>channels.
>
> Never? Code and documentation doesn't always naturally belong in a
> single project, especially when it relates to cross-project work.
> Similarly, if (say) Monasca, Vitrage, and Heat all need an IRC channel
> in which to collaborate on a specific topic, it seems fairly clear
> that none of #openstack-{monasca,vitrage,heat} are optimal choices.
What's wrong with #openstack-dev?
> The self-healing SIG has both a dedicated git repository (for docs,
> code, and in order to be able to use StoryBoard) and a dedicated IRC
> channel. We find both useful.
>
> Of course SIGs are more heavy-weight and long-lived so I'm not
> suggesting that all or even necessarily the majority of popup teams
> would need git/IRC. But I imagine it's possible in some cases, at
> least.
Right, SIGs are not designed to disappear after a task is done in the
way that popup teams are. If a popup team is going to create code, it
needs to end up in a repository that is owned and maintained by someone
over the long term. If that requires a new repo, and one of the existing
teams isn't a natural home, then I think a new regular team is likely a
better fit for the task than a popup team.
--
Doug
More information about the openstack-discuss
mailing list