Open Stack

On Thu, 2013-12-12 at 10:40 -0500, Doug Hellmann wrote:
> On Thu, Dec 12, 2013 at 8:24 AM, Thierry Carrez <thierry at openstack.org>wrote:
> 
> > Hi fellow members of the TC,
> >
> > We now have a set of 10 potential names for the J release [1], and need
> > to pick the best method to select the winner. Historically we held a
> > Launchpad poll (~openstack group) over a refined set of 4-5 options. The
> > problem is, we don't really use the Launchpad ~openstack group anymore
> > (used to be the subscribers to the openstack ML) and the group was
> > therefore closed... so that's no longer the "less worse" option.
> >
> > A few proposed solutions:
> >
> > 1. "Public" CIVS vote, accessible to everyone with the link
> > Easy to set up, but vote is limited to one per IP address, which can
> > block legitimate voters while encouraging ballot stuffing.
> >
> > 2a. Private CIVS vote, link sent to openstack-dev subscribers
> > A bit painful to set up (need to extract the 2926 subscribers emails,
> > then feed them to CIVS by batches of less than 1000), but at least it's
> > not gameable. One difference is that receiving a private vote email will
> > result in more participation than the only-announced-on-ML Launchpad poll.
> >
> > 2b. Private CIVS vote, link sent to openstack list subscribers
> > More painful to set up (8573 members), and CIVS advises against being
> > used to run above "a couple thousands voters". That said it's probably
> > the closest to the people who voted in the past (general list).
> >
> > 3. Launchpad poll over ~openstack
> > The OpenStack group is still there, so technically we could still refine
> > the list to 4-5 candidates and set up a poll there. It's not as good as
> > Condorcet though, and would be more like "the people who used to vote in
> > previous names" than "the people who should vote now".
> >
> > 4a. Screw popular voting, let's do a TC members condorcet
> > Easy option. Might be seen as TC power landgrab, and it was funny to let
> > the "people" decide.
> >
> > 4b. Screw voting, let's pick Jekyll and be done with it
> > A variant of the previous option in case of consensus.
> >
> > I'm open to other suggestions :)
> >
> 
> It seems like part of the problem is with the tooling. Do we need this vote
> to be as secure as PTL or TC elections? Would a tool like survey monkey be
> easier to work with for this case?

Yeah, I'm kind of thinking the same - I'd prefer this to be open to the
widest possible audience and I'm not too worried about gaming.

So, I'd tend to prefer (1) ... except the IP address limitation is
unfortunate.

Mark.