You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@hadoop.apache.org by Tsz Wo Sze <sz...@yahoo.com> on 2012/11/13 00:23:47 UTC
[DISCUSS] Clarify bylaws on PMC chair voting
Let's move the discussion to general@first.
Tsz-Wo
----- Forwarded Message 1/2 -----
From: Arun C Murthy <ac...@hortonworks.com>
Sent: Monday, November 12, 2012 3:15 PM
Eli,
If you are going to start a public vote on this, please
propose 'lazy majority', plus add a clause for STV or some such
mechanism for multiple nominations.
thanks,
Arun
On Nov 12, 2012, at 3:14 PM, Eli Collins wrote:
> I'll restart the vote on general.
>
> On Mon, Nov 12, 2012 at 2:53 PM, Robert Evans <ev...@yahoo-inc.com> wrote:
>> You are right we should move this to general@
>>
>> On 11/12/12 4:47 PM, "Mattmann, Chris A (388J)"
>> <ch...@jpl.nasa.gov> wrote:
>>
>>> Guys,
>>>
>>> The bylaws are a public document for Hadoop, no? Why is this VOTE
>>> happeningon private@ then?
>>>
>>> Cheers,
>>> Chris
>>>
>>> On Nov 12, 2012, at 4:16 PM, Robert Evans wrote:
>>>
>>>> +1
>>>>
>>>> On 11/12/12 4:11 PM, "Eli Collins" <el...@cloudera.com> wrote:
>>>>
>>>>> Here are the current bylaws for voting on the PMC chair:
>>>>>
>>>>> "The chair of the PMC is rotated annually. When the chair is rotated
>>>>> or when the current chair of the PMC resigns, the PMC votes to
>>>>> recommend a new chair using lazy consensus, but the decision must be
>>>>> ratified by the Apache board."
>>>>>
>>>>> Per the thread on nominating a chair, let's clarify that this means we
>>>>> propose the candidate with the most binding +1s and no -1s. Ie the
>>>>> following change:
>>>>>
>>>>> site $ svn diff
>>>>> Index: main/author/src/documentation/content/xdocs/bylaws.xml
>>>>> ===================================================================
>>>>> --- main/author/src/documentation/content/xdocs/bylaws.xml (revision
>>>>> 1408466)
>>>>> +++ main/author/src/documentation/content/xdocs/bylaws.xml (working
>>>>> copy)
>>>>> @@ -132,8 +132,10 @@
>>>>>
>>>>> <p>The chair of the PMC is rotated annually. When the chair is
>>>>> rotated or when the current chair of the PMC resigns, the PMC
>>>>> - votes to recommend a new chair using lazy consensus, but the
>>>>> - decision must be ratified by the Apache board.</p></li>
>>>>> + votes to recommend a new chair using lazy consensus. If there
>>>>> + are multiple candidates, the candidate with the most binding
>>>>> + +1 votes and no binding vetoes is selected. The decision must
>>>>> + be ratified by the Apache board.</p></li>
>>>>> </ul>
>>>>> </section>
----- Forwarded Message 2/2 -----
From: Arun C Murthy <ac...@hortonworks.com>
Sent: Monday, November 12, 2012 11:56 AM
I don't think a veto makes sense...
On Nov 12, 2012, at 11:55 AM, Eli Collins wrote:
> That's my understanding, and is how Ian ran the vote last year.
>
> On Monday, November 12, 2012, Robert Evans wrote:
>
>> Under the section for the Project Management Committee the last sentence
>> states that "When the chair is rotated or when the current chair of the
>> PMC resigns, the PMC votes to recommend a new chair using lazy consensus,
>> but the decision must be ratified by the Apache board." I am not really
>> sure how to apply lazy consensus to votes between multiple choices, but
>> Aaron's suggestion sounds as good to me as any. In my opinion a -1 would
>> mean I really don't want this person to be the chair, and just like in any
>> other veto it would require a valid explanation as to why. This feels to
>> be consistent with lazy consensus. However, I don't think a -1 will be an
>> issue in practice. If others think this is a problem we can update the
>> bylaws to make it more explicit how nominations for, voting on, and
>> rotation of the PMC chair happens. It will just take a week for the
>> bylaws vote and then we can start the vote for the PMC chair afterwards.
>>
>> --Bobby Evans
>>
>> On 11/12/12 12:57 PM, "Arun C Murthy" <acm@hortonworks.com <javascript:;>>
>> wrote:
>>
>>> I don't think the concept of '-1' makes sense... it should just be a
>>> straight vote?
>>>
>>> If we have more than two, we need to go STV?
>>>
>>> Arun
>>>
>>> On Nov 12, 2012, at 10:49 AM, Aaron T. Myers wrote:
>>>
>>>> The bylaws say that the vote for changing the PMC chair is by lazy
>>>> consensus. How will the vote work given that we have two candidates?
>>>>
>>>> Perhaps it should be: "the candidate who gets the most +1 votes with no
>>>> -1votes" ?
>>>>
>>>> --
>>>> Aaron T. Myers
>>>> Software Engineer, Cloudera
>>>>
>>>
>>> --
>>> Arun C. Murthy
>>> Hortonworks Inc.
>>> http://hortonworks.com/
>>>
>>
--
Arun C. Murthy
Hortonworks Inc.
http://hortonworks.com/
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Robert Evans <ev...@yahoo-inc.com>.
That sounds good to me.
On 11/15/12 8:44 PM, "Konstantin Shvachko" <sh...@gmail.com> wrote:
>The tiebreaker can be resolved by the current PMC chair.
>Or left for the board to choose.
>
>Thanks,
>--Konst
>
>On Thu, Nov 15, 2012 at 1:12 PM, Tsz Wo Sze <sz...@yahoo.com> wrote:
>> Owen's proposal sounds good in general. There are slight variances of
>>STV. I guess Owen probably means the one used in Apache board voting
>>(http://wiki.apache.org/general/BoardVoting). We should add a link to
>>their wiki in our bylaws.
>>
>>
>> How about tiebreaker? What if there are only two candidates and they
>>get exactly the same number of votes?
>>
>>
>> Tsz-Wo
>>
>>
>>
>>
>> ________________________________
>> From: Robert Evans <ev...@yahoo-inc.com>
>> To: "general@hadoop.apache.org" <ge...@hadoop.apache.org>
>> Sent: Tuesday, November 13, 2012 12:10 PM
>> Subject: Re: [DISCUSS] Clarify bylaws on PMC chair voting
>>
>> Vinod,
>>
>> I don't see what the PMC Chair does has any barring on how we select
>>them.
>> Yes I agree that a -1 will not be an issue. That is why I said
>>"However,
>> I don't think in practice it really matters if we allow for vetoes or
>> not." I too am +1 for Owen's suggestion, but I would like to see a vote
>> thread with the exact diff of the change to the bylaws.
>>
>> --Bobby
>>
>> On 11/13/12 12:47 PM, "Vinod Kumar Vavilapalli"
>><vi...@hortonworks.com>
>> wrote:
>>
>>>
>>>+1 to Owen's suggestion.
>>>
>>>Bobby, recall that PMC Chair is (just) a representative who communicates
>>>with the board on behalf of the PMC, and not any sort of "leader" (See
>>>http://www.apache.org/dev/pmc.html#chair); all the project decisions are
>>>driven by the PMC collectively. Given that, one should not expect
>>>vetoes
>>>at all in this vote.
>>>
>>>Thanks,
>>>+Vinod
>>>
>>>On Nov 13, 2012, at 7:25 AM, Robert Evans wrote:
>>>
>>>> The current bylaws state that the PMC chair recommendation to the
>>>>apache
>>>> board should be based off of lazy consensus. That means that any PMC
>>>> member can -1(veto) a candidate so long as they give a valid reason
>>>>with
>>>> the veto. The validity of the reason for the veto if challenged can be
>>>> confirmed by another PMC member. I am fine with the proposal to use
>>>>STV.
>>>> However, I don't think in practice it really matters if we allow for
>>>> vetoes or not. If someone really feels strongly enough to veto a
>>>> candidate, they would also feel strongly enough make their reason
>>>>known
>>>> during the voting and discussion on the candidate. If the reason is
>>>>valid
>>>> enough to withstand a challenge I would suspect it would also be valid
>>>> enough to influence any voting process we set up. I don't care what
>>>> voting process we use, I just care that the bylaws are clarified to
>>>>pick
>>>> one that can handle one or more candidates.
>>>>
>>>> -- Bobby
>>>>
>>>> On 11/12/12 5:53 PM, "Owen O'Malley" <om...@apache.org> wrote:
>>>>
>>>>> Thanks, Nicholas.
>>>>>
>>>>> I think the vote for PMC chair should be a straight majority vote
>>>>>with
>>>>>STV
>>>>> used in the case of more than 2 choices. Using +1 and/or -1's when
>>>>>voting
>>>>> in a multiple choice seems confused and likely to cause more problems
>>>>>than
>>>>> it solves.
>>>>>
>>>>> -- Owen
>>>>
>>>
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Konstantin Shvachko <sh...@gmail.com>.
The tiebreaker can be resolved by the current PMC chair.
Or left for the board to choose.
Thanks,
--Konst
On Thu, Nov 15, 2012 at 1:12 PM, Tsz Wo Sze <sz...@yahoo.com> wrote:
> Owen's proposal sounds good in general. There are slight variances of STV. I guess Owen probably means the one used in Apache board voting (http://wiki.apache.org/general/BoardVoting). We should add a link to their wiki in our bylaws.
>
>
> How about tiebreaker? What if there are only two candidates and they get exactly the same number of votes?
>
>
> Tsz-Wo
>
>
>
>
> ________________________________
> From: Robert Evans <ev...@yahoo-inc.com>
> To: "general@hadoop.apache.org" <ge...@hadoop.apache.org>
> Sent: Tuesday, November 13, 2012 12:10 PM
> Subject: Re: [DISCUSS] Clarify bylaws on PMC chair voting
>
> Vinod,
>
> I don't see what the PMC Chair does has any barring on how we select them.
> Yes I agree that a -1 will not be an issue. That is why I said "However,
> I don't think in practice it really matters if we allow for vetoes or
> not." I too am +1 for Owen's suggestion, but I would like to see a vote
> thread with the exact diff of the change to the bylaws.
>
> --Bobby
>
> On 11/13/12 12:47 PM, "Vinod Kumar Vavilapalli" <vi...@hortonworks.com>
> wrote:
>
>>
>>+1 to Owen's suggestion.
>>
>>Bobby, recall that PMC Chair is (just) a representative who communicates
>>with the board on behalf of the PMC, and not any sort of "leader" (See
>>http://www.apache.org/dev/pmc.html#chair); all the project decisions are
>>driven by the PMC collectively. Given that, one should not expect vetoes
>>at all in this vote.
>>
>>Thanks,
>>+Vinod
>>
>>On Nov 13, 2012, at 7:25 AM, Robert Evans wrote:
>>
>>> The current bylaws state that the PMC chair recommendation to the apache
>>> board should be based off of lazy consensus. That means that any PMC
>>> member can -1(veto) a candidate so long as they give a valid reason with
>>> the veto. The validity of the reason for the veto if challenged can be
>>> confirmed by another PMC member. I am fine with the proposal to use
>>>STV.
>>> However, I don't think in practice it really matters if we allow for
>>> vetoes or not. If someone really feels strongly enough to veto a
>>> candidate, they would also feel strongly enough make their reason known
>>> during the voting and discussion on the candidate. If the reason is
>>>valid
>>> enough to withstand a challenge I would suspect it would also be valid
>>> enough to influence any voting process we set up. I don't care what
>>> voting process we use, I just care that the bylaws are clarified to pick
>>> one that can handle one or more candidates.
>>>
>>> -- Bobby
>>>
>>> On 11/12/12 5:53 PM, "Owen O'Malley" <om...@apache.org> wrote:
>>>
>>>> Thanks, Nicholas.
>>>>
>>>> I think the vote for PMC chair should be a straight majority vote with
>>>>STV
>>>> used in the case of more than 2 choices. Using +1 and/or -1's when
>>>>voting
>>>> in a multiple choice seems confused and likely to cause more problems
>>>>than
>>>> it solves.
>>>>
>>>> -- Owen
>>>
>>
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Eli Collins <el...@cloudera.com>.
Hey Nicholas,
See my comment on the vote thread to update the bylaws. Agree,
pointing to http://wiki.apache.org/general/BoardVoting is the right
thing. Doug pointed out to me offline that there's a project (Apache
Steve) for STV that they'll be switching to (currently only OpenTSV
IIUC).
On Thu, Nov 15, 2012 at 1:12 PM, Tsz Wo Sze <sz...@yahoo.com> wrote:
> Owen's proposal sounds good in general. There are slight variances of STV. I guess Owen probably means the one used in Apache board voting (http://wiki.apache.org/general/BoardVoting). We should add a link to their wiki in our bylaws.
>
>
> How about tiebreaker? What if there are only two candidates and they get exactly the same number of votes?
>
>
> Tsz-Wo
>
>
>
>
> ________________________________
> From: Robert Evans <ev...@yahoo-inc.com>
> To: "general@hadoop.apache.org" <ge...@hadoop.apache.org>
> Sent: Tuesday, November 13, 2012 12:10 PM
> Subject: Re: [DISCUSS] Clarify bylaws on PMC chair voting
>
> Vinod,
>
> I don't see what the PMC Chair does has any barring on how we select them.
> Yes I agree that a -1 will not be an issue. That is why I said "However,
> I don't think in practice it really matters if we allow for vetoes or
> not." I too am +1 for Owen's suggestion, but I would like to see a vote
> thread with the exact diff of the change to the bylaws.
>
> --Bobby
>
> On 11/13/12 12:47 PM, "Vinod Kumar Vavilapalli" <vi...@hortonworks.com>
> wrote:
>
>>
>>+1 to Owen's suggestion.
>>
>>Bobby, recall that PMC Chair is (just) a representative who communicates
>>with the board on behalf of the PMC, and not any sort of "leader" (See
>>http://www.apache.org/dev/pmc.html#chair); all the project decisions are
>>driven by the PMC collectively. Given that, one should not expect vetoes
>>at all in this vote.
>>
>>Thanks,
>>+Vinod
>>
>>On Nov 13, 2012, at 7:25 AM, Robert Evans wrote:
>>
>>> The current bylaws state that the PMC chair recommendation to the apache
>>> board should be based off of lazy consensus. That means that any PMC
>>> member can -1(veto) a candidate so long as they give a valid reason with
>>> the veto. The validity of the reason for the veto if challenged can be
>>> confirmed by another PMC member. I am fine with the proposal to use
>>>STV.
>>> However, I don't think in practice it really matters if we allow for
>>> vetoes or not. If someone really feels strongly enough to veto a
>>> candidate, they would also feel strongly enough make their reason known
>>> during the voting and discussion on the candidate. If the reason is
>>>valid
>>> enough to withstand a challenge I would suspect it would also be valid
>>> enough to influence any voting process we set up. I don't care what
>>> voting process we use, I just care that the bylaws are clarified to pick
>>> one that can handle one or more candidates.
>>>
>>> -- Bobby
>>>
>>> On 11/12/12 5:53 PM, "Owen O'Malley" <om...@apache.org> wrote:
>>>
>>>> Thanks, Nicholas.
>>>>
>>>> I think the vote for PMC chair should be a straight majority vote with
>>>>STV
>>>> used in the case of more than 2 choices. Using +1 and/or -1's when
>>>>voting
>>>> in a multiple choice seems confused and likely to cause more problems
>>>>than
>>>> it solves.
>>>>
>>>> -- Owen
>>>
>>
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Tsz Wo Sze <sz...@yahoo.com>.
Owen's proposal sounds good in general. There are slight variances of STV. I guess Owen probably means the one used in Apache board voting (http://wiki.apache.org/general/BoardVoting). We should add a link to their wiki in our bylaws.
How about tiebreaker? What if there are only two candidates and they get exactly the same number of votes?
Tsz-Wo
________________________________
From: Robert Evans <ev...@yahoo-inc.com>
To: "general@hadoop.apache.org" <ge...@hadoop.apache.org>
Sent: Tuesday, November 13, 2012 12:10 PM
Subject: Re: [DISCUSS] Clarify bylaws on PMC chair voting
Vinod,
I don't see what the PMC Chair does has any barring on how we select them.
Yes I agree that a -1 will not be an issue. That is why I said "However,
I don't think in practice it really matters if we allow for vetoes or
not." I too am +1 for Owen's suggestion, but I would like to see a vote
thread with the exact diff of the change to the bylaws.
--Bobby
On 11/13/12 12:47 PM, "Vinod Kumar Vavilapalli" <vi...@hortonworks.com>
wrote:
>
>+1 to Owen's suggestion.
>
>Bobby, recall that PMC Chair is (just) a representative who communicates
>with the board on behalf of the PMC, and not any sort of "leader" (See
>http://www.apache.org/dev/pmc.html#chair); all the project decisions are
>driven by the PMC collectively. Given that, one should not expect vetoes
>at all in this vote.
>
>Thanks,
>+Vinod
>
>On Nov 13, 2012, at 7:25 AM, Robert Evans wrote:
>
>> The current bylaws state that the PMC chair recommendation to the apache
>> board should be based off of lazy consensus. That means that any PMC
>> member can -1(veto) a candidate so long as they give a valid reason with
>> the veto. The validity of the reason for the veto if challenged can be
>> confirmed by another PMC member. I am fine with the proposal to use
>>STV.
>> However, I don't think in practice it really matters if we allow for
>> vetoes or not. If someone really feels strongly enough to veto a
>> candidate, they would also feel strongly enough make their reason known
>> during the voting and discussion on the candidate. If the reason is
>>valid
>> enough to withstand a challenge I would suspect it would also be valid
>> enough to influence any voting process we set up. I don't care what
>> voting process we use, I just care that the bylaws are clarified to pick
>> one that can handle one or more candidates.
>>
>> -- Bobby
>>
>> On 11/12/12 5:53 PM, "Owen O'Malley" <om...@apache.org> wrote:
>>
>>> Thanks, Nicholas.
>>>
>>> I think the vote for PMC chair should be a straight majority vote with
>>>STV
>>> used in the case of more than 2 choices. Using +1 and/or -1's when
>>>voting
>>> in a multiple choice seems confused and likely to cause more problems
>>>than
>>> it solves.
>>>
>>> -- Owen
>>
>
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Robert Evans <ev...@yahoo-inc.com>.
Vinod,
I don't see what the PMC Chair does has any barring on how we select them.
Yes I agree that a -1 will not be an issue. That is why I said "However,
I don't think in practice it really matters if we allow for vetoes or
not." I too am +1 for Owen's suggestion, but I would like to see a vote
thread with the exact diff of the change to the bylaws.
--Bobby
On 11/13/12 12:47 PM, "Vinod Kumar Vavilapalli" <vi...@hortonworks.com>
wrote:
>
>+1 to Owen's suggestion.
>
>Bobby, recall that PMC Chair is (just) a representative who communicates
>with the board on behalf of the PMC, and not any sort of "leader" (See
>http://www.apache.org/dev/pmc.html#chair); all the project decisions are
>driven by the PMC collectively. Given that, one should not expect vetoes
>at all in this vote.
>
>Thanks,
>+Vinod
>
>On Nov 13, 2012, at 7:25 AM, Robert Evans wrote:
>
>> The current bylaws state that the PMC chair recommendation to the apache
>> board should be based off of lazy consensus. That means that any PMC
>> member can -1(veto) a candidate so long as they give a valid reason with
>> the veto. The validity of the reason for the veto if challenged can be
>> confirmed by another PMC member. I am fine with the proposal to use
>>STV.
>> However, I don't think in practice it really matters if we allow for
>> vetoes or not. If someone really feels strongly enough to veto a
>> candidate, they would also feel strongly enough make their reason known
>> during the voting and discussion on the candidate. If the reason is
>>valid
>> enough to withstand a challenge I would suspect it would also be valid
>> enough to influence any voting process we set up. I don't care what
>> voting process we use, I just care that the bylaws are clarified to pick
>> one that can handle one or more candidates.
>>
>> -- Bobby
>>
>> On 11/12/12 5:53 PM, "Owen O'Malley" <om...@apache.org> wrote:
>>
>>> Thanks, Nicholas.
>>>
>>> I think the vote for PMC chair should be a straight majority vote with
>>>STV
>>> used in the case of more than 2 choices. Using +1 and/or -1's when
>>>voting
>>> in a multiple choice seems confused and likely to cause more problems
>>>than
>>> it solves.
>>>
>>> -- Owen
>>
>
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Vinod Kumar Vavilapalli <vi...@hortonworks.com>.
+1 to Owen's suggestion.
Bobby, recall that PMC Chair is (just) a representative who communicates with the board on behalf of the PMC, and not any sort of "leader" (See http://www.apache.org/dev/pmc.html#chair); all the project decisions are driven by the PMC collectively. Given that, one should not expect vetoes at all in this vote.
Thanks,
+Vinod
On Nov 13, 2012, at 7:25 AM, Robert Evans wrote:
> The current bylaws state that the PMC chair recommendation to the apache
> board should be based off of lazy consensus. That means that any PMC
> member can -1(veto) a candidate so long as they give a valid reason with
> the veto. The validity of the reason for the veto if challenged can be
> confirmed by another PMC member. I am fine with the proposal to use STV.
> However, I don't think in practice it really matters if we allow for
> vetoes or not. If someone really feels strongly enough to veto a
> candidate, they would also feel strongly enough make their reason known
> during the voting and discussion on the candidate. If the reason is valid
> enough to withstand a challenge I would suspect it would also be valid
> enough to influence any voting process we set up. I don't care what
> voting process we use, I just care that the bylaws are clarified to pick
> one that can handle one or more candidates.
>
> -- Bobby
>
> On 11/12/12 5:53 PM, "Owen O'Malley" <om...@apache.org> wrote:
>
>> Thanks, Nicholas.
>>
>> I think the vote for PMC chair should be a straight majority vote with STV
>> used in the case of more than 2 choices. Using +1 and/or -1's when voting
>> in a multiple choice seems confused and likely to cause more problems than
>> it solves.
>>
>> -- Owen
>
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Robert Evans <ev...@yahoo-inc.com>.
The current bylaws state that the PMC chair recommendation to the apache
board should be based off of lazy consensus. That means that any PMC
member can -1(veto) a candidate so long as they give a valid reason with
the veto. The validity of the reason for the veto if challenged can be
confirmed by another PMC member. I am fine with the proposal to use STV.
However, I don't think in practice it really matters if we allow for
vetoes or not. If someone really feels strongly enough to veto a
candidate, they would also feel strongly enough make their reason known
during the voting and discussion on the candidate. If the reason is valid
enough to withstand a challenge I would suspect it would also be valid
enough to influence any voting process we set up. I don't care what
voting process we use, I just care that the bylaws are clarified to pick
one that can handle one or more candidates.
-- Bobby
On 11/12/12 5:53 PM, "Owen O'Malley" <om...@apache.org> wrote:
>Thanks, Nicholas.
>
>I think the vote for PMC chair should be a straight majority vote with STV
>used in the case of more than 2 choices. Using +1 and/or -1's when voting
>in a multiple choice seems confused and likely to cause more problems than
>it solves.
>
>-- Owen
Re: [DISCUSS] Clarify bylaws on PMC chair voting
Posted by Owen O'Malley <om...@apache.org>.
Thanks, Nicholas.
I think the vote for PMC chair should be a straight majority vote with STV
used in the case of more than 2 choices. Using +1 and/or -1's when voting
in a multiple choice seems confused and likely to cause more problems than
it solves.
-- Owen