From: Chris Dent <cdent...@anticdent.org>
Reply: OpenStack Development Mailing List (not for usage questions)
Date: October 12, 2016 at 09:10:45
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] PTG from the Ops Perspective - a few short notes
> The reason, as I understood it, for having a separate event is not
> because all those users are such a horrible distraction but because
> the sales and marketing efforts are (inherent in themselves and also
> in the way that the corporate patrons want the devs to participate
> in those efforts at the cost of the design summit-ing).
> For most of us, as Dmitry points out, getting to summit even once a
> year is a real challenge. Getting to 4 events, which is what will be
> required to maintain full engagement with both the planning and
> development of a project (not to mention cross project concerns), will
> either be impossible or require a significant change in the financial
> commitment that the patron companies are making. I think we'd all like
> to see that change, but I think we all also know that it's not very likely
> at this time.
So, are there people who already go to both Summits and both Midcycles for
projects? They're already attending 4 events. In effect you're now changing it
from 2 (ostensibly) significantly cheaper and 2 expensive events to 4
moderately expensive events. While it might not average out perfectly, it
probably comes close.
That said, the people that already attend all 4 are either in very significant
roles in that project team (PTL, etc.) or influential enough in their company
to secure that funding. There are people that will only attend a fraction of
those 4 though as it already is. That said, these projects *are* teams. I think
they could strategically determine who should go to which event because the
people who already have significant roles will likely be able to get that
funding and attend all events. The rest should be able to determine who can
represent the project at each as funding becomes a topic for each employer.
For Glance and other projects, I have been absent from many events for a
variety of reasons and have been able to rely on my fellow team members to
represent the project appropriately and communicate effectively what happened
(including feedback received). If there isn't that level of trust in a team, I
wonder if that team is really working effectively together today.
> That leaves us in a bit of a bind. It feels very much like we are
> narrowing the direct feedback pipeline between operators and users. It
> also feels like we are strengthening distinctions between various
> types of contributors, based mostly on the economic power their
> corporate patrons are willing to wield for them. And finally it
> feels like we are making it harder for more people to be engaged in
> cross project work.
Again, I'd argue that teams need to determine the priorities for team members
at events to partake in different efforts. Maybe teams will prioritize having
cross-project liaisons who can travel at the PTG. Meanwhile those who are
driving specific efforts and features may end up at the summit. Presumably the
team is already communicating and there will be some overlap so having disjoint
set of team members at the different events will not be harmful. Ostensibly,
having a cheaper PTG will mean that more developers can get together for those
cross-project efforts that would never have happened at a mid-cycle (unless
mid-cycle organizers worked to co-locate them).
> I'm certain that none of these things are the desired outcomes.
> What can we do to clarify the situation and remedy the issues?
I agree that these worst case scenarios were not the desired outcomes. I think
these outcomes are also the product of a mindset that "I" need to attend all of
the events otherwise "I" won't be as involved as "I" want/need to be. I would
challenge that mindset and remind everyone with that mindset that they need to
consider the entirety of the team and the opportunities this might open up for
different team members. This could decentralize some of the responsibilities
only a few people currently hold which will only benefit the team. If the split
of the PTG and Summit act as a forcing function for that, I'm especially glad
for this decision.
As a matter of transparency, I've been involved with OpenStack in various ways
for 2+ years and have only attended 1 midcycle and 1 summit. Both were
productive, but I have been limited by "real life" most of the time. I think
some of the teams I work with would find it hard to argue that they suffered
from my absence or that I've been less effective due to it.
Yes, for those of us who love to travel on some one else's dime this will be
painful, but I think it will ultimately prove good if teams remain optimistic
(and perhaps clever).
OpenStack Development Mailing List (not for usage questions)