You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@community.apache.org by Bertrand Delacretaz <bd...@apache.org> on 2024/10/31 14:04:13 UTC
Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
On Thu, Oct 31, 2024 at 1:57 PM Rich Bowen <no...@github.com>
wrote (on a GitHub page):
>
> I think possibly one of the points of confusion here is the use of the word "consensus",
> which has been a point of confusion for years, since that word has several substantially different uses....
We do have an ASF glossary at
https://www.apache.org/foundation/glossary, which we might improve and
which we should IMHO point to more often.
Or maybe create an additional glossary on the comdev website, and link
both pages for completeness.
-Bertrand
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org
Re: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by Jarek Potiuk <ja...@potiuk.com>.
Comment: I also have a natural inclination to think that consensus is a bit
different than "allow veto". But, well, technically that's it. But I think
there is a subtle difference between consensus and "I have the right to
veto" thing - especially that consensus is all about getting people united,
and veto is all about getting them conflicted.
IMHO veto should only be used in very specific situations where the cost of
handling potential veto are much smaller than potential gains from applying
it. I think personally that the cost of veto for cases where people are
involved and control can be exerted overy the whole project are far too big.
Maybe because veto has a very strong meaning in Polish and for Polish
people.
The "liberum veto" [1] was the "golden liberty" thing in the XVI th/ XVII
th/ XVII th century that every Polish-Lithuanian nobleman had the right to
- block any decision by any deciding body. And while historians universally
agree that "liberum veto" was great to build "constitutionalism" in Poland
- at first when it has been used mostly to show different opinions and
trigger discussion (and eventually was withdrawn), they also agree it's
been the main reason for its fall. Especially when neighbouring "predator"
countries (Russia especially, Prussia and Austria) learned that they can
use it to manipulate the system and started to "buy" liberum veto
from noblemen and started to be applied to more things that
originally intended.
Here an interesting passage from the Wikipedia article:
> the proceedings of the next few sejms, the veto was still occasionally
overruled, but it became gradually more accepted. Before 20 years had
passed, in 1669 in Kraków, the entire Sejm was prematurely disrupted on the
strength of the liberum veto before it had finished its deliberations by
the Kyiv deputy, Adam Olizar. The practice spiraled out of control, and in
1688, the Sejm was dissolved even before the proceedings had begun or the
Marshal of the Sejm was elected.
This later led to the First [2], Second [3] and Third [4] Partition of
Poland. Poland disappeared from the maps (but luckily not the hearts of
people) for 123 years after.
I see a lot of parallels here.
J.
[1] Liberum veto https://en.wikipedia.org/wiki/Liberum_veto
[2] First Partition of Poland -
https://en.wikipedia.org/wiki/First_Partition_of_Poland
[3] Second Partition of Poland -
https://en.wikipedia.org/wiki/Second_Partition_of_Poland
[4] Third Partition of Poland -
https://en.wikipedia.org/wiki/Third_Partition_of_Poland
On Thu, Oct 31, 2024 at 3:27 PM Rich Bowen <rb...@rcbowen.com> wrote:
>
>
> > On Oct 31, 2024, at 10:04 AM, Bertrand Delacretaz <
> bdelacretaz@apache.org> wrote:
> >
> > On Thu, Oct 31, 2024 at 1:57 PM Rich Bowen <no...@github.com>
> > wrote (on a GitHub page):
> >>
> >> I think possibly one of the points of confusion here is the use of the
> word "consensus",
> >> which has been a point of confusion for years, since that word has
> several substantially different uses....
> >
> > We do have an ASF glossary at
> > https://www.apache.org/foundation/glossary, which we might improve and
> > which we should IMHO point to more often.
> >
> > Or maybe create an additional glossary on the comdev website, and link
> > both pages for completeness.
>
>
> Thank you. I tend to forget that that’s there. While the definition of
> “consensus” there is at odds with how I understand the word, it does seem
> reasonable for me to agree to use that definition. I am willing to accept
> that my *opinion* is different from the *consensus* of the community as to
> the meaning of the word, and I’ll use it accordingly.
>
> I do continue to think that if a PMC votes unanimously-minus-one for a
> particular candidate, that’s a pretty good indicator that there’s
> dysfunction in that PMC. Ironically, if I now attempt to override that
> position, I would be exhibiting the very behavior that I am condemning.
>
> —
> Rich Bowen
> rbowen@rcbowen.com
>
>
>
Re: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by Rich Bowen <rb...@rcbowen.com>.
> On Oct 31, 2024, at 10:04 AM, Bertrand Delacretaz <bd...@apache.org> wrote:
>
> On Thu, Oct 31, 2024 at 1:57 PM Rich Bowen <no...@github.com>
> wrote (on a GitHub page):
>>
>> I think possibly one of the points of confusion here is the use of the word "consensus",
>> which has been a point of confusion for years, since that word has several substantially different uses....
>
> We do have an ASF glossary at
> https://www.apache.org/foundation/glossary, which we might improve and
> which we should IMHO point to more often.
>
> Or maybe create an additional glossary on the comdev website, and link
> both pages for completeness.
Thank you. I tend to forget that that’s there. While the definition of “consensus” there is at odds with how I understand the word, it does seem reasonable for me to agree to use that definition. I am willing to accept that my *opinion* is different from the *consensus* of the community as to the meaning of the word, and I’ll use it accordingly.
I do continue to think that if a PMC votes unanimously-minus-one for a particular candidate, that’s a pretty good indicator that there’s dysfunction in that PMC. Ironically, if I now attempt to override that position, I would be exhibiting the very behavior that I am condemning.
—
Rich Bowen
rbowen@rcbowen.com
Re: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by Dave Fisher <wa...@apache.org>.
> On Nov 1, 2024, at 1:13 PM, Jarek Potiuk <ja...@potiuk.com> wrote:
>
> I think Christopher nailed it. Now when I think of it the difference is
> that consensus is **not** a vote. It's a belief that we have an agreement
> and if someone disagrees, then we can start voting (or discussing).
>
> Maybe (proposal) renaming it to "unanimous vote" would be better. And keep
> consensus to "agreement with no vote even started".
>
> That would make much more sense IMHO. It still leaves the question "should
> PMC members be voted via unanimous or majority vote?" - but that sentence
> makes so much more sense if we use "unanimous" here.
Do we have consensus on when unanimous (actually no vetoes) vs. majority votes are to be used?
We probably have more discussion about when vetoes are and are not appropriate. They are often quite poisonous when used.
1. PMC votes on committers and committee membership there are probably more reasons for requiring unanimity.
2. In Technical choices unanimity is more often a Bad Idea™
If we are actually having a [DISCUSS] phase as a consensus reaching discussion where we actually listen to each other legitimate concerns will prevent a vote and no vetoes are needed.
I liked Jarek’s history lesson about vetoes in the Polish-Lithuanian Commonwealth.
If we MUST mention both majority and unanimous the specify majority first.
Best,
Dave
>
> J.
>
> On Fri, Nov 1, 2024 at 8:25 PM Christopher <ct...@apache.org> wrote:
>
>> I agree consensus is a core value, and I think that word makes sense to use
>> when describing the important practice of building consensus.
>>
>> However, I also agree that the term, used as a type of vote, as in
>> "consensus vote" is confusing and ambiguous. I would be in favor of
>> updating the docs around voting to avoid using "consensus" to describe a
>> kind of vote, but I would suggest keeping the word around to describe the
>> value and importance of consensus building, prior to and during a vote.
>>
>> On Thu, Oct 31, 2024, 21:56 Keith N. McKenna
>> <ke...@comcast.net.invalid> wrote:
>>
>>> On 2024-10-31 at 19:17, "Dave Fisher" <wa...@apache.org> wrote:
>>>>> On Oct 31, 2024, at 2:28 PM, Craig Russell <ap...@gmail.com>
>>> wrote:
>>>>>
>>>>> I would be in favor of removing "consensus" from our documentation.
>>>>>
>>>>> It has a general meaning of "group agreement" which is not how we use
>> it
>>> here.
>>>>>
>>>>> I would propose using "unanimous" instead of "consensus" as there
>> seems
>>> to
>>> be no ambiguity about unanimity.
>>>>
>>>> Consensus is a great term and is what we do. We don’t do unanimity, we
>>> find
>>> an acceptable path by working towards Consensus. We should avoid Votes if
>>> at
>>> all possible.
>>>>
>>>> To me Consensus is a Core value of the ASF and without it I have no
>>> interest
>>> in the Foundation.
>>>>
>>>> Best,
>>>> Dave
>>>>
>>>
>>> +1 from me
>>>
>>> Regards
>>> Keith
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
>>> For additional commands, e-mail: dev-help@community.apache.org
>>>
>>>
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org
Re: Re[2]: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by Jarek Potiuk <ja...@potiuk.com>.
I think Christopher nailed it. Now when I think of it the difference is
that consensus is **not** a vote. It's a belief that we have an agreement
and if someone disagrees, then we can start voting (or discussing).
Maybe (proposal) renaming it to "unanimous vote" would be better. And keep
consensus to "agreement with no vote even started".
That would make much more sense IMHO. It still leaves the question "should
PMC members be voted via unanimous or majority vote?" - but that sentence
makes so much more sense if we use "unanimous" here.
J.
On Fri, Nov 1, 2024 at 8:25 PM Christopher <ct...@apache.org> wrote:
> I agree consensus is a core value, and I think that word makes sense to use
> when describing the important practice of building consensus.
>
> However, I also agree that the term, used as a type of vote, as in
> "consensus vote" is confusing and ambiguous. I would be in favor of
> updating the docs around voting to avoid using "consensus" to describe a
> kind of vote, but I would suggest keeping the word around to describe the
> value and importance of consensus building, prior to and during a vote.
>
> On Thu, Oct 31, 2024, 21:56 Keith N. McKenna
> <ke...@comcast.net.invalid> wrote:
>
> > On 2024-10-31 at 19:17, "Dave Fisher" <wa...@apache.org> wrote:
> > >> On Oct 31, 2024, at 2:28 PM, Craig Russell <ap...@gmail.com>
> > wrote:
> > >>
> > >> I would be in favor of removing "consensus" from our documentation.
> > >>
> > >> It has a general meaning of "group agreement" which is not how we use
> it
> > here.
> > >>
> > >> I would propose using "unanimous" instead of "consensus" as there
> seems
> > to
> > be no ambiguity about unanimity.
> > >
> > >Consensus is a great term and is what we do. We don’t do unanimity, we
> > find
> > an acceptable path by working towards Consensus. We should avoid Votes if
> > at
> > all possible.
> > >
> > >To me Consensus is a Core value of the ASF and without it I have no
> > interest
> > in the Foundation.
> > >
> > >Best,
> > >Dave
> > >
> >
> > +1 from me
> >
> > Regards
> > Keith
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
> > For additional commands, e-mail: dev-help@community.apache.org
> >
> >
>
Re: Re[2]: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by Christopher <ct...@apache.org>.
I agree consensus is a core value, and I think that word makes sense to use
when describing the important practice of building consensus.
However, I also agree that the term, used as a type of vote, as in
"consensus vote" is confusing and ambiguous. I would be in favor of
updating the docs around voting to avoid using "consensus" to describe a
kind of vote, but I would suggest keeping the word around to describe the
value and importance of consensus building, prior to and during a vote.
On Thu, Oct 31, 2024, 21:56 Keith N. McKenna
<ke...@comcast.net.invalid> wrote:
> On 2024-10-31 at 19:17, "Dave Fisher" <wa...@apache.org> wrote:
> >> On Oct 31, 2024, at 2:28 PM, Craig Russell <ap...@gmail.com>
> wrote:
> >>
> >> I would be in favor of removing "consensus" from our documentation.
> >>
> >> It has a general meaning of "group agreement" which is not how we use it
> here.
> >>
> >> I would propose using "unanimous" instead of "consensus" as there seems
> to
> be no ambiguity about unanimity.
> >
> >Consensus is a great term and is what we do. We don’t do unanimity, we
> find
> an acceptable path by working towards Consensus. We should avoid Votes if
> at
> all possible.
> >
> >To me Consensus is a Core value of the ASF and without it I have no
> interest
> in the Foundation.
> >
> >Best,
> >Dave
> >
>
> +1 from me
>
> Regards
> Keith
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
> For additional commands, e-mail: dev-help@community.apache.org
>
>
Re[2]: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by "Keith N. McKenna" <ke...@comcast.net.INVALID>.
On 2024-10-31 at 19:17, "Dave Fisher" <wa...@apache.org> wrote:
>> On Oct 31, 2024, at 2:28 PM, Craig Russell <ap...@gmail.com> wrote:
>>
>> I would be in favor of removing "consensus" from our documentation.
>>
>> It has a general meaning of "group agreement" which is not how we use it
here.
>>
>> I would propose using "unanimous" instead of "consensus" as there seems to
be no ambiguity about unanimity.
>
>Consensus is a great term and is what we do. We don’t do unanimity, we find
an acceptable path by working towards Consensus. We should avoid Votes if at
all possible.
>
>To me Consensus is a Core value of the ASF and without it I have no interest
in the Foundation.
>
>Best,
>Dave
>
+1 from me
Regards
Keith
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org
Re: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by Dave Fisher <wa...@apache.org>.
> On Oct 31, 2024, at 2:28 PM, Craig Russell <ap...@gmail.com> wrote:
>
> I would be in favor of removing "consensus" from our documentation.
>
> It has a general meaning of "group agreement" which is not how we use it here.
>
> I would propose using "unanimous" instead of "consensus" as there seems to be no ambiguity about unanimity.
Consensus is a great term and is what we do. We don’t do unanimity, we find an acceptable path by working towards Consensus. We should avoid Votes if at all possible.
To me Consensus is a Core value of the ASF and without it I have no interest in the Foundation.
Best,
Dave
>
> And then "majority approval" which is "more +1 than -1 votes".
>
> So we would have these with the Apache three vote minimum:
>
> Procedural votes (new PMC members, new committers) require majority approval with at least three +1 votes.
>
> Code changes require unanimous approval with at least three +1 votes.
>
> While I really dislike changing history, "consensus" has been a bad term for so long...
>
> Craig
>
>
>> On Oct 31, 2024, at 07:04, Bertrand Delacretaz <bd...@apache.org> wrote:
>>
>> On Thu, Oct 31, 2024 at 1:57 PM Rich Bowen <no...@github.com>
>> wrote (on a GitHub page):
>>>
>>> I think possibly one of the points of confusion here is the use of the word "consensus",
>>> which has been a point of confusion for years, since that word has several substantially different uses....
>>
>> We do have an ASF glossary at
>> https://www.apache.org/foundation/glossary, which we might improve and
>> which we should IMHO point to more often.
>>
>> Or maybe create an additional glossary on the comdev website, and link
>> both pages for completeness.
>>
>> -Bertrand
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
>> For additional commands, e-mail: dev-help@community.apache.org
>>
>
> Craig L Russell
> clr@apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
> For additional commands, e-mail: dev-help@community.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org
Re: Consensus? we do have a glossary (was:Refactor of "new pmc member" doc...)
Posted by Craig Russell <ap...@gmail.com>.
I would be in favor of removing "consensus" from our documentation.
It has a general meaning of "group agreement" which is not how we use it here.
I would propose using "unanimous" instead of "consensus" as there seems to be no ambiguity about unanimity.
And then "majority approval" which is "more +1 than -1 votes".
So we would have these with the Apache three vote minimum:
Procedural votes (new PMC members, new committers) require majority approval with at least three +1 votes.
Code changes require unanimous approval with at least three +1 votes.
While I really dislike changing history, "consensus" has been a bad term for so long...
Craig
> On Oct 31, 2024, at 07:04, Bertrand Delacretaz <bd...@apache.org> wrote:
>
> On Thu, Oct 31, 2024 at 1:57 PM Rich Bowen <no...@github.com>
> wrote (on a GitHub page):
>>
>> I think possibly one of the points of confusion here is the use of the word "consensus",
>> which has been a point of confusion for years, since that word has several substantially different uses....
>
> We do have an ASF glossary at
> https://www.apache.org/foundation/glossary, which we might improve and
> which we should IMHO point to more often.
>
> Or maybe create an additional glossary on the comdev website, and link
> both pages for completeness.
>
> -Bertrand
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
> For additional commands, e-mail: dev-help@community.apache.org
>
Craig L Russell
clr@apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org
Re: Consensus? we do have a glossary
Posted by Shane Curcuru <as...@shanecurcuru.org>.
Bertrand Delacretaz wrote on 10/31/24 10:04 AM:
> On Thu, Oct 31, 2024 at 1:57 PM Rich Bowen <no...@github.com>
> wrote (on a GitHub page):
>>
>> I think possibly one of the points of confusion here is the use of the word "consensus",
>> which has been a point of confusion for years, since that word has several substantially different uses....
"Consensus" is a *concept* around group decision making; I hope we agree
on that. Personally, I'm -0.5 on any definitions of "consensus vote",
because the ASF meaning of "vote" tends to be very specific around
procedure, not the overall concept of "how we make some decision".
Consensus at the board level is simple to define, we should find a way
to add it to the glossary - noted that this is the Board process, and
PMCs often have different kinds of decisions. At every board meeting:
- We have a posted agenda with detailed reports.
-- Thus, the group deciding something (here, the board) has clear
information available with plenty of notice to all, giving plenty of
chances for questions or comments before the meeting/decision process.
- During the meeting, we review PMC reports, and perhaps discuss flags.
-- Thus, we know the group deciding something is all participating, and
is clear on what the questions to decide are.
- The Chair then says "are there any *objections* to accepting submitted
reports *by consensus*".
- If any Director objects, they state so, and we discuss to figure out a
way forward. This is similar conceptually to a veto, but the word
"veto" has special meaning elsewhere, so let's not use that one.
- If no Director objects (even if they're not completely happy with
FooBar's report, and will be asking questions later), then the Chair
asks the Secretary to record acceptance of the reports "by consensus".
> We do have an ASF glossary at
> https://www.apache.org/foundation/glossary, which we might improve and
> which we should IMHO point to more often.
Huh, I completely forgot about that. And I'll have changes to suggest,
especially for words that have common governance meanings outside the
ASF. "Consensus decision making" has a sizeable wikipedia page, along
with plenty of organizational governance meanings; for major concepts
like that we should work toward linking to a couple of outside
definitions, so we can explain any subtly different ASF connotations.
> Or maybe create an additional glossary on the comdev website, and link
> both pages for completeness.
Please, please let's not duplicate the glossary. We have enough trouble
with awareness of the existing one; duplicating something as simple as a
glossary of words is like copypasta code in the core of your event loop.
--
- Shane
Member
The Apache Software Foundation
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org