[Pulp-dev] PUP Process: "obvious consensus"

[Brian Bouterse]

I asked about some of these governance questions to a group of community
managers from several open source projects that I meet with weekly. They
said that if you don't have a BDFL (Pulp does not) the other very popular
model is the lazy consensus model. I think lazy consensus is the spirit of
pup1. I asked for some examples and they pointed me at the CentOS
governance model [0][1].

Also @daviddavis and I were talking and codifying the problem as what value
should X be if X are the number of +1s required to pass a decision with
zero -1 votes (vetos)? The CentOS governance model sets X = 0 by stating
"There is no minimum +1 vote requirement". I'm also advocating for X=0 for
the reasons I wrote in my earlier email. Practically speaking, I don't
think an X=1, or X=2 will prevent many proposals that would have also
passed with X=0.

Regardless of the X value, we should continue the discussion so we can
arrive at a decision on both pup1 and pup3. Thanks for continuing the convo.

[0]:
https://www.centos.org/about/governance/appendix-glossary/#consensus-decision-making
[1]: https://www.centos.org/about/governance/voting/

-Brian


On Tue, Jun 13, 2017 at 11:46 AM, Ina Panova <ipanova at redhat.com> wrote:

> And if we would remove all 'shades of grey' and go back just to +1 and -1
> where people would need to make their mind up *clearly* which would lead
> stronger arguments of doing or not doing this.
>
>
>
> --------
> Regards,
>
> Ina Panova
> Software Engineer| Pulp| Red Hat Inc.
>
> "Do not go where the path may lead,
>  go instead where there is no path and leave a trail."
>
> On Tue, Jun 13, 2017 at 5:30 PM, David Davis <daviddavis at redhat.com>
> wrote:
>
>> In this model of where only -1 votes stop the PUP from passing, wouldn’t
>> it mean that there needn't be any consensus at all? In other words we could
>> effectively strike the language about consensus from PUP-1. This model
>> makes me worried that people other than those casting -1 won’t bother to
>> vote or participate since only -1 votes matter.
>>
>> I personally like the idea of having at least 30% that are +1 or +0. This
>> means that enough -0 votes can still block the vote, and also +0 votes goes
>> towards helping the PUP pass. Thus +0 and -0 would both matter. I think
>> this is a good compromise between the extremes of "broad buy-in" and
>> "default to change."
>>
>>
>> David
>>
>> On Tue, Jun 13, 2017 at 10:36 AM, Brian Bouterse <bbouters at redhat.com>
>> wrote:
>>
>>> We should (I thought we did) adopt a process that favors change and does
>>> not have a "broad buy-in requirement". Any change that doesn't harm the
>>> project should be allowed without broad buy-in. This empowers even a single
>>> individual to enact change. This makes Pulp better because:
>>>
>>> * Everyone is empowered. A single individual can have a meaningful
>>> impact.
>>> * Anyone can stop an idea that will negatively affect the project or
>>> community via veto.
>>> * We avoid the tyranny of the majority [0] or supermajority.
>>> * It avoids politics. If we start averaging, or counting votes
>>> for/against in an offsetting way, there will be politics. Counting votes
>>> for/against will create inequality because influential project members will
>>> likely see their ideas adopted but others won't. Having a "default to
>>> change and any core dev can veto" approach creates equality.
>>>
>>> Regarding how "obvious consensus" works with the "veto-or-it-passes"
>>> model, if there are zero -1 votes cast, that means no one wanted to stop
>>> the process. If no wants to stop it, and at least one is for it, then the
>>> most sensible thing to do is to pass it. Since someone took time to write
>>> the PUP there is obviously someone giving it a +1. If one person really
>>> wants to go to place X for dinner (aka a +1), and there are no
>>> counterproposals (aka a -1 with a suggestion) or strong preferences against
>>> (aka -0 or +0) then the group will probably go to place X for dinner by way
>>> of "obvious consensus".
>>>
>>> In summary, adopting a "default to accept or reject with even a single
>>> veto" system creates an equal system. A system where, a single individual
>>> can make a difference, and anyone can stop a bad idea from occurring. To
>>> @mhrivnak's point about a change not meeting a broad range of needs, I
>>> expect -1's to be cast in those cases, so this system is still very safe in
>>> terms of protecting the projects needs and interests.
>>>
>>> [0]: https://en.wikipedia.org/wiki/Tyranny_of_the_majority
>>>
>>> -Brian
>>>
>>> On Mon, Jun 12, 2017 at 7:53 PM, David Davis <daviddavis at redhat.com>
>>> wrote:
>>>
>>>> Not sure this is true. I actually abstained from voting on PUP-3
>>>> because I was somewhere between a +0 and a -0.
>>>>
>>>>
>>>> David
>>>>
>>>> On Mon, Jun 12, 2017 at 11:43 AM, Ina Panova <ipanova at redhat.com>
>>>> wrote:
>>>>
>>>>> Having at least one  +1 is not impartial approach just because the
>>>>> developer who , as you said, found the time for the research and writing
>>>>> down the proposal obviously will vote as +1 :)
>>>>>
>>>>>
>>>>>
>>>>> --------
>>>>> Regards,
>>>>>
>>>>> Ina Panova
>>>>> Software Engineer| Pulp| Red Hat Inc.
>>>>>
>>>>> "Do not go where the path may lead,
>>>>>  go instead where there is no path and leave a trail."
>>>>>
>>>>> On Mon, Jun 12, 2017 at 5:35 PM, Austin Macdonald <amacdona at redhat.com
>>>>> > wrote:
>>>>>
>>>>>> This reminds me of the concept of a "Do-ocracy".
>>>>>>
>>>>>> If developers take the time to research and write up a proposal, they
>>>>>> have "done". It seems completely reasonable to default to the opinion of
>>>>>> the people that cared enough to do the work. If it isn't the right
>>>>>> decision, then someone must actively block it, simple as that.
>>>>>>
>>>>>> I think the rule should be "PUP passes if we have at least one +1 and
>>>>>> no -1s".
>>>>>>
>>>>>> _______________________________________________
>>>>>> Pulp-dev mailing list
>>>>>> Pulp-dev at redhat.com
>>>>>> https://www.redhat.com/mailman/listinfo/pulp-dev
>>>>>>
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Pulp-dev mailing list
>>>>> Pulp-dev at redhat.com
>>>>> https://www.redhat.com/mailman/listinfo/pulp-dev
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> Pulp-dev mailing list
>>>> Pulp-dev at redhat.com
>>>> https://www.redhat.com/mailman/listinfo/pulp-dev
>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/pulp-dev/attachments/20170615/5b14db8f/attachment.htm>