[tc] Gradually reduce TC to 11 members over 2020
Hi everyone,
The size of the TC is a trade-off between getting enough community representation and keeping enough members engaged and active. The current size (13 members) was defined by in 2013, as we moved from 5 directly-elected seats + all PTLs (which would have been 14 people) to a model that could better cope with our explosive growth. Since then, 13 has worked well, to ensure that new blood could come in at every cycle.
I would argue that today, there are far less need to get wide representation in the TC (we are pretty aligned), and less difficulty to enter the TC (there is more turnover). In 2019 OpenStack, 13 members is a rather large group. It is becoming difficult to find 13 people able to commit to a significant amount of time over the coming year. And it is difficult to keep all those 13 members active and engaged.
IMHO it is time to reduce the TC to 11 members, which sounds like a more reasonable and manageable size. We should encourage people to stop for a while and come back, rather than burn too many people at the same time. We should encourage more people to watch from the sidelines, rather than have a group so large that everyone that can be in it is in it.
This would not be a big-bang change, just something we would gradually put in place over the next year. My strawman plan would be as follows:
- Sept. 2019 election: no change, elect 6 seats as planned - Feb. 2020 election: elect 6 seats instead of 7 - Sept. 2020 election: elect 5 seats instead of 6 - Then elect 6 seats at every start-of-year and 5 at every end-of-year
That would result in TC membership sizes:
- U cycle (Q4 2019/Q1 2020): 13 members - V cycle (Q2/Q3 2020): 12 members - W cycle (Q4 2020/Q1 2021): 11 members - after that: 11 members
FWIW, I intend to not run for reelection in the Feb 2020 election, so nobody else has to sacrifice their seat to that reform for that election :)
Thoughts?
On Tue, 27 Aug 2019, Thierry Carrez wrote:
Thoughts?
Good idea, but I'd argue that 11 is still too many. 7 or 5 might be a better target. Getting there might be complicated but I think it would be useful by providing an opportunity to winnow the focus of the TC to what matters (whatever we might decide that is).
On Tue, Aug 27, 2019 at 6:30 AM Chris Dent cdent+os@anticdent.org wrote:
On Tue, 27 Aug 2019, Thierry Carrez wrote:
Thoughts?
Good idea, but I'd argue that 11 is still too many. 7 or 5 might be a better target. Getting there might be complicated but I think it would be useful by providing an opportunity to winnow the focus of the TC to what matters (whatever we might decide that is).
+1. I would vote for a reduction to 11, but I would like to see it smaller.
// jim
Jim Rollenhagen wrote:
On Tue, Aug 27, 2019 at 6:30 AM Chris Dent <cdent+os@anticdent.org mailto:cdent%2Bos@anticdent.org> wrote:
On Tue, 27 Aug 2019, Thierry Carrez wrote: > Thoughts? Good idea, but I'd argue that 11 is still too many. 7 or 5 might be a better target. Getting there might be complicated but I think it would be useful by providing an opportunity to winnow the focus of the TC to what matters (whatever we might decide that is).+1. I would vote for a reduction to 11, but I would like to see it smaller.
Yes, I thought about that. But if we reduce too fast, we might have too many incumbents going for reelection over a smaller amount of slots... and I wanted to make sure that we preserved opportunities for new blood to join, even during that transition.
Maybe 9 is a better sweet spot, though. Or we could just continue the gradual reduction over the next years.
On 8/27/19 2:16 PM, Thierry Carrez wrote:
Jim Rollenhagen wrote:
On Tue, Aug 27, 2019 at 6:30 AM Chris Dent <cdent+os@anticdent.org mailto:cdent%2Bos@anticdent.org> wrote:
On Tue, 27 Aug 2019, Thierry Carrez wrote:
> Thoughts?
Good idea, but I'd argue that 11 is still too many. 7 or 5 might be a better target. Getting there might be complicated but I think it would be useful by providing an opportunity to winnow the focus of the TC to what matters (whatever we might decide that is).
+1. I would vote for a reduction to 11, but I would like to see it smaller.
Yes, I thought about that. But if we reduce too fast, we might have too many incumbents going for reelection over a smaller amount of slots... and I wanted to make sure that we preserved opportunities for new blood to join, even during that transition.
Maybe 9 is a better sweet spot, though. Or we could just continue the gradual reduction over the next years.
9 holders of Rings of Power is what I would propose if I had a say in that :)
On Tue, 2019-08-27 at 14:38 +0200, Dmitry Tantsur wrote:
9 holders of Rings of Power is what I would propose if I had a say in that :)
My lore is a little rusty, wouldn't that only apply to humans? What about inclusiveness?
Regards, JP
PS: Sorry for those to become the Nazguls, I guess?
On 27/08/19 6:13 AM, Thierry Carrez wrote:
Hi everyone,
The size of the TC is a trade-off between getting enough community representation and keeping enough members engaged and active. The current size (13 members) was defined by in 2013, as we moved from 5 directly-elected seats + all PTLs (which would have been 14 people) to a model that could better cope with our explosive growth. Since then, 13 has worked well, to ensure that new blood could come in at every cycle.
I would argue that today, there are far less need to get wide representation in the TC (we are pretty aligned), and less difficulty to enter the TC (there is more turnover). In 2019 OpenStack, 13 members is a rather large group. It is becoming difficult to find 13 people able to commit to a significant amount of time over the coming year. And it is difficult to keep all those 13 members active and engaged.
IMHO it is time to reduce the TC to 11 members, which sounds like a more reasonable and manageable size. We should encourage people to stop for a while and come back, rather than burn too many people at the same time. We should encourage more people to watch from the sidelines, rather than have a group so large that everyone that can be in it is in it.
This would not be a big-bang change, just something we would gradually put in place over the next year. My strawman plan would be as follows:
- Sept. 2019 election: no change, elect 6 seats as planned
- Feb. 2020 election: elect 6 seats instead of 7
- Sept. 2020 election: elect 5 seats instead of 6
- Then elect 6 seats at every start-of-year and 5 at every end-of-year
That would result in TC membership sizes:
- U cycle (Q4 2019/Q1 2020): 13 members
- V cycle (Q2/Q3 2020): 12 members
- W cycle (Q4 2020/Q1 2021): 11 members
- after that: 11 members
TBH, I'd suggest we:
* Aim for 9 * Reduce by 2 each time so we don't end up with an even number * Start now when 4 people have announced they're stepping down (though in a perfect world this would have been the cycle when 7 seats are up for election)
FWIW, I intend to not run for reelection in the Feb 2020 election, so nobody else has to sacrifice their seat to that reform for that election :)
That is still my intention also, so nobody would have to sacrifice their seat even if we reduced to 9.
cheers, Zane.
On 2019-08-27 11:25:32 -0400 (-0400), Zane Bitter wrote: [...]
- Start now when 4 people have announced they're stepping down
[...]
With my election official hat on, this is rather short notice. We've announced the number of seats up for election and nominations start at the (UTC) end of today. While I agree reducing the size of the TC sounds worthwhile, realistically I think we need to announce at the start of the new term that it's the plan for the following cycle's election.
With my (soon to be former, at least for a cycle) TC member hat on, I suggest that we hold this election as planned but circulate the suggestion of electing fewer seats in coming cycles. Then once the new TC has been confirmed let them decide whether that should be the plan as soon into the new term as possible, and announce that decision so the community will be prepared for it in the next election.
On Tue, Aug 27, 2019 at 11:02 AM Jeremy Stanley fungi@yuggoth.org wrote:
With my (soon to be former, at least for a cycle) TC member hat on, I suggest that we hold this election as planned but circulate the suggestion of electing fewer seats in coming cycles. Then once the new TC has been confirmed let them decide whether that should be the plan as soon into the new term as possible, and announce that decision so the community will be prepared for it in the next election.
I completely agree with this, it is usually around election time (often just after) that we think about these things but making a change for the impending election is too quick. I like Thierry's suggested schedule and I like the target of 9. I do not have a strong preference for getting to 9 in 2 vs 4 cycles (2 at a time vs 1 at a time). As has been mentioned elsewhere (IRC?) having an even number should not be an issue, looking back I can recall only one vote where it might have been a problem and in that case since it was clear we did not have consensus the plan was dropped anyway. FWIW, StarlingX had an even number of members on its Technical Steering Committee (TC equivalent) for two cycles and it has not been a problem for basically the same reason, when driving for even close consensus you generally do not get ties.
dt
On Aug 27, 2019, at 12:27 PM, Dean Troyer dtroyer@gmail.com wrote:
On Tue, Aug 27, 2019 at 11:02 AM Jeremy Stanley fungi@yuggoth.org wrote:
With my (soon to be former, at least for a cycle) TC member hat on, I suggest that we hold this election as planned but circulate the suggestion of electing fewer seats in coming cycles. Then once the new TC has been confirmed let them decide whether that should be the plan as soon into the new term as possible, and announce that decision so the community will be prepared for it in the next election.
I completely agree with this, it is usually around election time (often just after) that we think about these things but making a change for the impending election is too quick. I like Thierry's suggested schedule and I like the target of 9. I do not have a strong preference for getting to 9 in 2 vs 4 cycles (2 at a time vs 1 at a time). As has been mentioned elsewhere (IRC?) having an even number should not be an issue, looking back I can recall only one vote where it might have been a problem and in that case since it was clear we did not have consensus the plan was dropped anyway. FWIW, StarlingX had an even number of members on its Technical Steering Committee (TC equivalent) for two cycles and it has not been a problem for basically the same reason, when driving for even close consensus you generally do not get ties.
Yes, I think it’s too close to the current election to change this now. Other than that, the only reason to take longer would be to ensure that we maintain a balance in the number of seats up for election each cycle.
Doug
On Tue, Aug 27, 2019 at 12:34 PM Dean Troyer dtroyer@gmail.com wrote:
On Tue, Aug 27, 2019 at 11:02 AM Jeremy Stanley fungi@yuggoth.org wrote:
With my (soon to be former, at least for a cycle) TC member hat on, I suggest that we hold this election as planned but circulate the suggestion of electing fewer seats in coming cycles. Then once the new TC has been confirmed let them decide whether that should be the plan as soon into the new term as possible, and announce that decision so the community will be prepared for it in the next election.
I completely agree with this, it is usually around election time (often just after) that we think about these things but making a change for the impending election is too quick. I like Thierry's suggested schedule and I like the target of 9. I do not have a strong preference for getting to 9 in 2 vs 4 cycles (2 at a time vs 1 at a time). As has been mentioned elsewhere (IRC?) having an even number should not be an issue, looking back I can recall only one vote where it might have been a problem and in that case since it was clear we did not have consensus the plan was dropped anyway. FWIW, StarlingX had an even number of members on its Technical Steering Committee (TC equivalent) for two cycles and it has not been a problem for basically the same reason, when driving for even close consensus you generally do not get ties.
I agree even numbers are not a problem. I don't think (hope?) the existing TC would merge anything that went 7-6 anyway.
// jim
dt
-- Dean Troyer dtroyer@gmail.com
--- On Wed, 28 Aug 2019 01:50:40 +0900 Jim Rollenhagenjim@jimrollenhagen.com wrote ----On Tue, Aug 27, 2019 at 12:34 PM Dean Troyer dtroyer@gmail.com wrote:On Tue, Aug 27, 2019 at 11:02 AM Jeremy Stanley fungi@yuggoth.org wrote:
With my (soon to be former, at least for a cycle) TC member hat on, I suggest that we hold this election as planned but circulate the suggestion of electing fewer seats in coming cycles. Then once the new TC has been confirmed let them decide whether that should be the plan as soon into the new term as possible, and announce that decision so the community will be prepared for it in the next election.
I completely agree with this, it is usually around election time (often just after) that we think about these things but making a change for the impending election is too quick. I like Thierry's suggested schedule and I like the target of 9. I do not have a strong preference for getting to 9 in 2 vs 4 cycles (2 at a time vs 1 at a time). As has been mentioned elsewhere (IRC?) having an even number should not be an issue, looking back I can recall only one vote where it might have been a problem and in that case since it was clear we did not have consensus the plan was dropped anyway. FWIW, StarlingX had an even number of members on its Technical Steering Committee (TC equivalent) for two cycles and it has not been a problem for basically the same reason, when driving for even close consensus you generally do not get ties.I agree even numbers are not a problem. I don't think (hope?)the existing TC would merge anything that went 7-6 anyway.I agree with the idea here to reduce the TC number. I think even number make more sense to have balance between both elections. We can reduce it gradually with target of 11 or 10 as of now and depends on future situations we can decide for more lower number.-gmann// jim
dt
-- Dean Troyer dtroyer@gmail.com
On Tue, 2019-08-27 at 12:50 -0400, Jim Rollenhagen wrote:
I agree even numbers are not a problem. I don't think (hope?) the existing TC would merge anything that went 7-6 anyway.
Agreed with that.
And because I didn't write my opinion on the topic: - I agree on the reduction. I don't know what the sweet spot is. 9 might be it. - If all the candidates and the election officials are ok with reduced seats this time, we could start doing it now.
It seems the last point isn't obvious, so in the meantime could we propose the plan for the reduction by proposing a governance change?
Regards, JP
Jean-Philippe Evrard wrote:
On Tue, 2019-08-27 at 12:50 -0400, Jim Rollenhagen wrote:
I agree even numbers are not a problem. I don't think (hope?) the existing TC would merge anything that went 7-6 anyway.
Agreed with that.
And because I didn't write my opinion on the topic:
- I agree on the reduction. I don't know what the sweet spot is. 9
might be it.
- If all the candidates and the election officials are ok with reduced
seats this time, we could start doing it now.
It seems the last point isn't obvious, so in the meantime could we propose the plan for the reduction by proposing a governance change?
To close on that:
It's too late to change for the current election, however if we don't get any new candidate, then the TC would mechanically get reduced to 11 already.
Based on how the election turns out, once it is over I'll propose a governance change to gradually transition to 9 or 11 members, which will affect future elections.
Cheers,
On Tue, Aug 27, 2019 at 11:26 AM Zane Bitter zbitter@redhat.com wrote:
TBH, I'd suggest we:
- Aim for 9
I also think 9 is likely a better number.
- Reduce by 2 each time so we don't end up with an even number
My biggest concern with this proposal is having the possibility of an even number of people sitting on the TC at any one given time.
- Start now when 4 people have announced they're stepping down (though
in a perfect world this would have been the cycle when 7 seats are up for election)
I feel like this is something that would have to be discussed before the board so they are aware in advance of starting to reduce the size of the TC. If the board says "nope", then I suspect we need to respect that.
FWIW, I intend to not run for reelection in the Feb 2020 election, so nobody else has to sacrifice their seat to that reform for that election :)
That is still my intention also, so nobody would have to sacrifice their seat even if we reduced to 9.
cheers, Zane.
On 2019-08-28 08:30:58 -0400 (-0400), Julia Kreger wrote: [...]
I feel like this is something that would have to be discussed before the board so they are aware in advance of starting to reduce the size of the TC. If the board says "nope", then I suspect we need to respect that.
[...]
I think it would be nice of us to give them a heads-up, but the OSF bylaws don't mandate any particular size for the OpenStack TC:
https://www.openstack.org/legal/technical-committee-member-policy/
Instead it only requires a supermajority of the sitting TC members to agree to change where it's specified in our charter:
https://governance.openstack.org/tc/reference/charter.html#tc-members
So the OSF board of directors doesn't actually have any direct control over the decision as far as I can tell.
To give a survey on the numbers of the valid commit contributors
We got
61 in Austin :) ... 1204 contributors in havana [1] ... 2629 in liberty 2991 in mitaka 3104 in newton (which is peek point) [2] 2581 in ocata 2452 in pike 1925 in queens 1665 in rocky 1612 in stein [3]
I believe the numbers of contributors should reflect on the numbers of TCs. If we try to compare the ratio (like stein cycle vs past six released cycles), I think to reduce to 9 (from 13) is acceptable numbers (also should be more than 7).
And for even number issue, I believe to have 1 cycle of even number TCs will not affect much. But if we can, odd number definitely helps to reduce conflict when we end up using a poll to resolved some issues (one TC can propose solutions but it takes all TCs can to resolve the conflict.).
[1] https://www.stackalytics.com/?release=havana&project_type=all&metric... [2] https://www.stackalytics.com/?release=newton&project_type=all&metric... [3] https://www.stackalytics.com/?release=stein&project_type=all&metric=...
On Tue, Aug 27, 2019 at 6:17 PM Thierry Carrez thierry@openstack.org wrote:
Hi everyone,
The size of the TC is a trade-off between getting enough community representation and keeping enough members engaged and active. The current size (13 members) was defined by in 2013, as we moved from 5 directly-elected seats + all PTLs (which would have been 14 people) to a model that could better cope with our explosive growth. Since then, 13 has worked well, to ensure that new blood could come in at every cycle.
I would argue that today, there are far less need to get wide representation in the TC (we are pretty aligned), and less difficulty to enter the TC (there is more turnover). In 2019 OpenStack, 13 members is a rather large group. It is becoming difficult to find 13 people able to commit to a significant amount of time over the coming year. And it is difficult to keep all those 13 members active and engaged.
IMHO it is time to reduce the TC to 11 members, which sounds like a more reasonable and manageable size. We should encourage people to stop for a while and come back, rather than burn too many people at the same time. We should encourage more people to watch from the sidelines, rather than have a group so large that everyone that can be in it is in it.
This would not be a big-bang change, just something we would gradually put in place over the next year. My strawman plan would be as follows:
- Sept. 2019 election: no change, elect 6 seats as planned
- Feb. 2020 election: elect 6 seats instead of 7
- Sept. 2020 election: elect 5 seats instead of 6
- Then elect 6 seats at every start-of-year and 5 at every end-of-year
That would result in TC membership sizes:
- U cycle (Q4 2019/Q1 2020): 13 members
- V cycle (Q2/Q3 2020): 12 members
- W cycle (Q4 2020/Q1 2021): 11 members
- after that: 11 members
FWIW, I intend to not run for reelection in the Feb 2020 election, so nobody else has to sacrifice their seat to that reform for that election :)
Thoughts?
-- Thierry Carrez (ttx)
on 2019/8/28 11:38, Rico Lin wrote:
I believe the numbers of contributors should reflect on the numbers of TCs. If we try to compare the ratio (like stein cycle vs past six released cycles), I think to reduce to 9 (from 13) is acceptable numbers (also should be more than 7).
And for even number issue, I believe to have 1 cycle of even number TCs will not affect much. But if we can, odd number definitely helps to reduce conflict when we end up using a poll to resolved some issues (one TC can propose solutions but it takes all TCs can to resolve the conflict.).
I think that's good suggestion.
regards.
On Wed, 2019-08-28 at 11:38 +0800, Rico Lin wrote:
To give a survey on the numbers of the valid commit contributors
We got
61 in Austin :) ... 1204 contributors in havana [1] ... 2629 in liberty 2991 in mitaka 3104 in newton (which is peek point) [2] 2581 in ocata 2452 in pike 1925 in queens 1665 in rocky 1612 in stein [3]
Is that ponderated with length of a cycle? Ocata was shorter IIRC.
Regards, JP
participants (13)
-
Chris Dent
-
Dean Troyer
-
Dmitry Tantsur
-
Doug Hellmann
-
Eliza
-
Ghanshyam Mann
-
Jean-Philippe Evrard
-
Jeremy Stanley
-
Jim Rollenhagen
-
Julia Kreger
-
Rico Lin
-
Thierry Carrez
-
Zane Bitter