[opensuse-project] From today's openSUSE board meeting: Who can vote for the openSUSE board?
We had tonight on the opensuse-project IRC channel a lifely discussion about the board election. The central question was who can vote for the openSUSE board. We haven't come to a real conclusion yet, see below for some of the comments that were discussed. We'd like to continue the discussion here on the mailing list and meet again in two weeks on the openSUSE-project IRC channel.
I took the following notes trying to summarize the key points of the discussion:
* Who can vote? - every individual can only vote once, so nobody should be allowed to fake several personalities and use them for voting
Suggestion: have a public list of voters to check that nobody is listed twice. The list will be closed at a certain time.
- basically three different proposals: + only members + anyone + members + non-members members vouch for (web-of-trust alike)
The idea here is to have the openSUSE community vote for the openSUSE board. The members are the smallest part of the community that has been appointed by the board, a member is somebody with a continued and substantial contribution to openSUSE. Having just members excludes some parts of the community. On the other hand, it makes fraud easier if everybody is allowed to vote - including people that have nothing to do with openSUSE and just show up to vote for their "friend".
* We should have an Election Committee that organizes the election. The members of the Election Committee cannot be nominated for the board, they have to oversee the election.
What are others doing?
Let's compare what other projects are doing, here's a first list of URLs: Debian: http://www.debian.org/vote/ Fedora: http://fedoraproject.org/wiki/Board/Elections
Andreas
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Andreas Jaeger wrote: [...] | * Who can vote? [...] | - basically three different proposals: | + only members | + anyone | + members + non-members members vouch for (web-of-trust alike)
Just to clarify the latter option: the idea is to have the following people being able to vote: - - members - - non-members who have a verified identity by the proxy of a member
Say A is a member, but X, Y and Z aren't. But A knows that X, Y and Z are - - real individuals, not fake identities used to fraud the election - - part of the openSUSE community, as contributors or users
A would notice the election committee (whoever that will be, to be discussed) that X, Y and Z are as defined above. And X, Y and Z would then be given a ticket to vote on the board election.
Of course, there's still a potential risk, that A would only do that for people he knows would vote the same way he would. Or X, Y and Z could all be fake identities of A. But by granting membership status, we put a lot of trust into the members, as we verify their contributions. And as active contributors, it's probably safe enough to think they wouldn't sabotage the elections, the board and, to a certain extent, the community.
So basically, we've identified 3 options, two extreme ones: 1) only members 2) anyone (with a publicly published and frozen list some time before the elections) And one in-between: 3) members + non-members who have been acknowledged as being real individuals + part of the community by members
It's remotely inspired from the web-of-trust principle used by techniques such as PGP/GPG.
PS: I'm not saying it's the best approach, just clarifying the idea behind it ;))
cheers - -- ~ -o) Pascal Bleser pascal.bleser@opensuse.org ~ /\ http://opensuse.org -- I took the green pill ~ __v FOSDEM::23+24 Feb 2008, Brussels, http://fosdem.org
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Mandag den 17. Marts 2008 23:06:26 skrev Pascal Bleser:
- non-members who have a verified identity by the proxy of a member
Say A is a member, but X, Y and Z aren't. But A knows that X, Y and Z are
- real individuals, not fake identities used to fraud the election
- part of the openSUSE community, as contributors or users
How about restricting it so that each member can only vouch for one non-member. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tuesday 18 March 2008 03:50:34 am Martin Schlander wrote:
Mandag den 17. Marts 2008 23:06:26 skrev Pascal Bleser:
- non-members who have a verified identity by the proxy of a member
Say A is a member, but X, Y and Z aren't. But A knows that X, Y and Z are
- real individuals, not fake identities used to fraud the election
- part of the openSUSE community, as contributors or users
How about restricting it so that each member can only vouch for one non-member.
We can also restrict this elections to members, but take care to have more contributors included as members, now, before elections, and later, as permanent policy. Members should have right to propose and be active in recruiting new members.
Online identities are similar to real, you can't have 10 of them and be significant contributor with all of them, so allowing to vote guys that are often present and contribute can be pretty much enough to ensure fair voting.
If someone has 2 identities that nobody knows that belong to the same person and significant contributions under both, is it really evil that such person can cast 2 votes. On the other side I doubt that someone with good credit in community will attempt to cast 2 votes and ruin his reputation if ever that become known.
One of things that can help is trust rating system currently discussed only for Build Service, but it should be extended to community. User directory exist, add some voting system to each user name and there is lesser problems when you need someone dependable. User can actively advertise his entry in user directory, or hide it, but it would be one central place where one can give opinion. Now it is only our memory that collects data on someone's activity, special skills, contributions.
On Monday 17 March 2008 22:28:08 Andreas Jaeger wrote:
- Who can vote?
The idea here is to have the openSUSE community vote for the openSUSE board. The members are the smallest part of the community that has been appointed by the board, a member is somebody with a continued and substantial contribution to openSUSE. Having just members excludes some parts of the community. On the other hand, it makes fraud easier if everybody is allowed to vote - including people that have nothing to do with openSUSE and just show up to vote for their "friend".
I think every community member should have the right to vote.
The board's main task is to facilitate communication. It certainly helps to have a broad backing by the community to actually do this. It also gives the board a strong legitimation, e.g. when representing the interests of the community to Novell.
So I would suggest to let everybody vote who has registered in the openSUSE user directory and has signed the Guiding Principles. That's a very inclusive way to define who has the right to vote, so it's likely to cover a big fraction of the community. We could also add an additional step and require people to register as a voter, but in general I think the simpler the system is the better.
To prevent fraud I think it's enough to forbid a person to register twice as a voter or to give access to the account to other persons in the terms of service. Of course technically it's possible to do that, but if somebody is trying to affect results of a vote this has to be done in such a massive way, that it's probably not too hard to detect that and to deactivate fraudulent accounts. I also don't think it's very likely that somebody tries to affect elections this way, as without the backing by the community being in the board isn't associated with any particular power or benefit.
On Mon, 2008-03-17 at 23:42 +0100, Cornelius Schumacher wrote:
To prevent fraud I think it's enough to forbid a person to register twice as a voter or to give access to the account to other persons in the terms of service. Of course technically it's possible to do that, but if somebody is trying to affect results of a vote this has to be done in such a massive way, that it's probably not too hard to detect that and to deactivate fraudulent accounts. I also don't think it's very likely that somebody tries to affect elections this way, as without the backing by the community being in the board isn't associated with any particular power or benefit.
Is it possible to implement a system by which the election software registers your IP address, so e.g.: if you have an account and vote once, it logs your IP address so if you log out on to another account, but you're still using the same IP address, it won't let you vote?
If not, the idea of allowing those who members confirm to vote is the next best alternative IMHO.
On Mon, 2008-03-17 at 18:04 -0500, Kevin Dupuy wrote:
On Mon, 2008-03-17 at 23:42 +0100, Cornelius Schumacher wrote:
To prevent fraud I think it's enough to forbid a person to register twice as a voter or to give access to the account to other persons in the terms of service. Of course technically it's possible to do that, but if somebody is trying to affect results of a vote this has to be done in such a massive way, that it's probably not too hard to detect that and to deactivate fraudulent accounts. I also don't think it's very likely that somebody tries to affect elections this way, as without the backing by the community being in the board isn't associated with any particular power or benefit.
Is it possible to implement a system by which the election software registers your IP address, so e.g.: if you have an account and vote once, it logs your IP address so if you log out on to another account, but you're still using the same IP address, it won't let you vote?
If not, the idea of allowing those who members confirm to vote is the next best alternative IMHO.
Voting by restricting to a single IP address would exclude anyone who works for the same company behind a NAT, a husband and wife (both ardent supporters of openSUSE) at home behind a NAT, and so on...
Bryen
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Mon, Mar 17, 2008 at 5:38 PM, Bryen suseROCKS@bryen.com wrote:
On Mon, 2008-03-17 at 18:04 -0500, Kevin Dupuy wrote:
On Mon, 2008-03-17 at 23:42 +0100, Cornelius Schumacher wrote:
To prevent fraud I think it's enough to forbid a person to register twice as a voter or to give access to the account to other persons in the terms of service. Of course technically it's possible to do that, but if somebody is trying to affect results of a vote this has to be done in such a massive way, that it's probably not too hard to detect that and to deactivate fraudulent accounts. I also don't think it's very likely that somebody tries to affect elections this way, as without the backing by the community being in the board isn't associated with any particular power or benefit.
Is it possible to implement a system by which the election software registers your IP address, so e.g.: if you have an account and vote once, it logs your IP address so if you log out on to another account, but you're still using the same IP address, it won't let you vote?
If not, the idea of allowing those who members confirm to vote is the next best alternative IMHO.
Voting by restricting to a single IP address would exclude anyone who works for the same company behind a NAT, a husband and wife (both ardent supporters of openSUSE) at home behind a NAT, and so on...
It also doesn't block anyone who wants to vote from home & work, or uses TOR to spoof different IPs, or has any other method of using multiple IPs...
IP addresses aren't very reliable as a way to validate identity or uniqueness.
Best,
Zonker
On 3/18/2008 at 01:04, Kevin Dupuy kevindupuy@bellsouth.net wrote:
On Mon, 2008-03-17 at 23:42 +0100, Cornelius Schumacher wrote:
To prevent fraud I think it's enough to forbid a person to register twice as
a
voter or to give access to the account to other persons in the terms of service. Of course technically it's possible to do that, but if somebody is trying to affect results of a vote this has to be done in such a massive
way,
that it's probably not too hard to detect that and to deactivate fraudulent accounts. I also don't think it's very likely that somebody tries to affect elections this way, as without the backing by the community being in the board isn't associated with any particular power or benefit.
Is it possible to implement a system by which the election software registers your IP address, so e.g.: if you have an account and vote once, it logs your IP address so if you log out on to another account, but you're still using the same IP address, it won't let you vote?
Even though this would be technically possible, you might exclude complete school campuses in some regions. Just imagine the first one from there has the possibility to vote, all the others, as most likely they use the same proxy / internet access will be identified to come from the same IP Address... and I'm not sure blocking out students will be a good start ;)
Dominique
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On 17/03/2008, Cornelius Schumacher cschum@suse.de wrote:
To prevent fraud I think it's enough to forbid a person to register twice as a voter or to give access to the account to other persons in the terms of service. Of course technically it's possible to do that, but if somebody is trying to affect results of a vote this has to be done in such a massive way, that it's probably not too hard to detect that and to deactivate fraudulent accounts. I also don't think it's very likely that somebody tries to affect elections this way, as without the backing by the community being in the board isn't associated with any particular power or benefit.
As we discussed in the meeting, it's not clear at all that it's easy to detect fraudulent accounts: it's pretty trivial to register several accounts in the user directory and to cast a vote with those.
In this case we get quite a severely negative scenario, if someone can skew the results of the election so easily (even if it's, say, 10 votes). So we're taking a big risk here -- and dangerously IMO -- if you want to gamble on the idea that it's unlikely for _anyone_ from _anywhere_ to have the idea of inappropriately skewing results.
At the moment I prefer the membership only method because it's the simplest and is only really problematic in theory if the board is corrupt and unfairly give out membership in order to further their cause. In fact, this is pretty much the last premise I would ever accept as a possibility in the argument. Perhaps making the membership process even more transparent could help here, and I'd be willing to listen to any ideas about that.
I prefer this to the third option mainly because of simplicity, and I'm not totally opposed to it depending on its execution.
Kind thoughts,
On Tuesday 18 March 2008 02:15:58 Francis Giannaros wrote:
At the moment I prefer the membership only method because it's the simplest and is only really problematic in theory if the board is corrupt and unfairly give out membership in order to further their cause. In fact, this is pretty much the last premise I would ever accept as a possibility in the argument. Perhaps making the membership process even more transparent could help here, and I'd be willing to listen to any ideas about that.
If we are going this route I think it's mandatory that the process of giving out membership status is transferred to a committee which is not the board itself. This would also help to free the board to do the tasks which it's primarily meant to do.
Cornelius Schumacher wrote:
On Tuesday 18 March 2008 02:15:58 Francis Giannaros wrote:
At the moment I prefer the membership only method because it's the simplest and is only really problematic in theory if the board is corrupt and unfairly give out membership in order to further their cause. In fact, this is pretty much the last premise I would ever accept as a possibility in the argument. Perhaps making the membership process even more transparent could help here, and I'd be willing to listen to any ideas about that.
If we are going this route I think it's mandatory that the process of giving out membership status is transferred to a committee which is not the board itself. This would also help to free the board to do the tasks which it's primarily meant to do.
The other main problem with members-only is that there are many (including myself) in a grey area where we are long time committed SuSE / OpenSuSE users but don't have the skills / time / whatever to make the huge contribution that most of those so far approved as members do. Are our views really not relevant?
Conversely, opening to all certainly leads to problems balancing prevention of multiple votes whilst ensuring free access for e.g. those behind small business NATs. No voting system ever gets this 100% - in the UK we've had some real scandals recently with postal voting being (allegedly) rigged, when the full force of the law is there to protect it. Someone determined to cheat *will* succeed, though they may be found out eventually.
So I think the solution is broadly as suggested above - to have a wider definition of 'membership', supported by a specialist committee, and to allow all these members to vote. The cream of the membership maybe get special status (though not for voting) - such as @opensuse.org email aliases. When people apply for membership they maybe need to supply some minimal evidence of identity to minimise the risk of bot applications, plus maybe some evidence of contribution (e.g. to lists / fora / source code...) but beyond this I think we have to go on trust. The hassle required to create more than a few cheat duplicates will mean that most wouldn't bother, and there is only a small risk of it affecting any results. Also perhaps memberships expire after ? a year ? if nothing is done to use them (such as voting / contributing to relevant fora, wikis...)
Just my half a groat's worth.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Cornelius Schumacher wrote: | On Tuesday 18 March 2008 02:15:58 Francis Giannaros wrote: |> At the moment I prefer the membership only method because it's the |> simplest and is only really problematic in theory if the board is |> corrupt and unfairly give out membership in order to further their |> cause. In fact, this is pretty much the last premise I would ever |> accept as a possibility in the argument. Perhaps making the membership |> process even more transparent could help here, and I'd be willing to |> listen to any ideas about that. | | If we are going this route I think it's mandatory that the process of giving | out membership status is transferred to a committee which is not the board | itself. This would also help to free the board to do the tasks which it's | primarily meant to do.
Do you mean right now, or now and for the future ? Because if it's also for the future, then that "committee which is not the board itself" has no legitimation at all, as it hasn't even been elected.
And let's not fool ourselves, the level of independence isn't all that high either. If someone who works for Novell is inside that committee for memberships, and 3 Novell employees are on the board, how independent is it at all ? We're one community, with certain people knowing others personally and not others, hence a high degree of independence will never be reached anyway.
So who will determine that committee ? - - the board ? not independent from the board then - - the members ? possibly, they should have the least interest as they're members already - - a vote open to all ? worst of options IMO, as non-members could have an interest to vote for people who put the bar a lot lower for becoming member) - - Novell ? pointless, Novell appointed the board
It really gets us to the same point again and again.
This distrust of the individuals who compose the board right now is something that puzzles me, tbh. We're in a bootstrapping scenario now, and certain trade-offs have to be made because we're always getting back to chicken/egg until we have an elected board. The question is always which trade-offs to choose as the least worst one. You really think that Francis, Stephan, Andreas, Federico and I are favouring people who we know would vote for us when we check and vote internally on their level of contribution to the project ? (and I'm explicitly putting names on our faces as the confidence that has to be granted to us is not about "The Board" as some black box driven by evil interests, but about the individuals that compose it)
Now if you mean for the future, pretty much the same problems apply. Shall we also vote for a membership committee ? And create two boards, more or less ? As the next board will be elected, won't they have enough legitimacy and trust to do that on their own ? (or at least determine a membership committee themselves)
cheers - -- ~ -o) Pascal Bleser pascal.bleser@opensuse.org ~ /\ http://opensuse.org -- I took the green pill ~ __v FOSDEM::23+24 Feb 2008, Brussels, http://fosdem.org
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Pascal Bleser schrieb: | Cornelius Schumacher wrote: | | On Tuesday 18 March 2008 02:15:58 Francis Giannaros wrote: | |> At the moment I prefer the membership only method because it's the | |> simplest and is only really problematic in theory if the board is | |> corrupt and unfairly give out membership in order to further their | |> cause. In fact, this is pretty much the last premise I would ever | |> accept as a possibility in the argument. Perhaps making the membership | |> process even more transparent could help here, and I'd be willing to | |> listen to any ideas about that. | | | | If we are going this route I think it's mandatory that the process of | giving | | out membership status is transferred to a committee which is not the | board | | itself. This would also help to free the board to do the tasks which it's | | primarily meant to do. | | Do you mean right now, or now and for the future ? | Because if it's also for the future, then that "committee which is not | the board itself" has no legitimation at all, as it hasn't even been | elected. | | And let's not fool ourselves, the level of independence isn't all that | high either. If someone who works for Novell is inside that committee | for memberships, and 3 Novell employees are on the board, how | independent is it at all ? We're one community, with certain people | knowing others personally and not others, hence a high degree of | independence will never be reached anyway. | | So who will determine that committee ? | - the board ? not independent from the board then | - the members ? possibly, they should have the least interest as they're | members already | - a vote open to all ? worst of options IMO, as non-members could have | an interest to vote for people who put the bar a lot lower for becoming | member) | - Novell ? pointless, Novell appointed the board | | It really gets us to the same point again and again. | | This distrust of the individuals who compose the board right now is | something that puzzles me, tbh. We're in a bootstrapping scenario now, | and certain trade-offs have to be made because we're always getting back | to chicken/egg until we have an elected board. The question is always | which trade-offs to choose as the least worst one. | You really think that Francis, Stephan, Andreas, Federico and I are | favouring people who we know would vote for us when we check and vote | internally on their level of contribution to the project ? | (and I'm explicitly putting names on our faces as the confidence that | has to be granted to us is not about "The Board" as some black box | driven by evil interests, but about the individuals that compose it) | | Now if you mean for the future, pretty much the same problems apply. | Shall we also vote for a membership committee ? And create two boards, | more or less ? As the next board will be elected, won't they have enough | legitimacy and trust to do that on their own ? (or at least determine a | membership committee themselves) | | cheers
I fully agree. I would just like to add, that the problem of board members choosing the members in order to be reelected can be circumvented by restricting the board members to serve not more than one or two terms in a row. I also find this problem to be unlikely, but this could be the best solution for this attack vector.
HTH Felix
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tuesday 18 March 2008, Pascal Bleser wrote:
Cornelius Schumacher wrote: | On Tuesday 18 March 2008 02:15:58 Francis Giannaros wrote: |> At the moment I prefer the membership only method because it's the |> simplest and is only really problematic in theory if the board is |> corrupt and unfairly give out membership in order to further their |> cause. In fact, this is pretty much the last premise I would ever |> accept as a possibility in the argument. Perhaps making the membership |> process even more transparent could help here, and I'd be willing to |> listen to any ideas about that.
-snip-
Now if you mean for the future, pretty much the same problems apply. Shall we also vote for a membership committee ? And create two boards, more or less ? As the next board will be elected, won't they have enough legitimacy and trust to do that on their own ? (or at least determine a membership committee themselves)
What about enlarging the group of voters step by step? The current board has been appointed by Novell. Why not elect the next board "only" out of the group of members? This definitely gives the next board more legitimation, is a step in the right direction and is simple to do. And as goal for the future set up a process/election commitee to enlarge the group of voters to the group of "long time committed SuSE / OpenSuSE users" like Richard (MQ) (we can't afford losing this group) and to "members + non-members members vouch"
Best M
cheers
Pascal Bleser a écrit :
- a vote open to all ? worst of options IMO, as non-members could have
an interest to vote for people who put the bar a lot lower for becoming member)
?
I think for the first vote, vote should be open to all the *volunteer* (every people that asks to be able to vote).
* right now the board *power* is very small. In fact the board can only suggest things to Novell or administrators. so the interest to be board member is low (and I'm sure the work is not low - we must thank the present board for doing that), in a near future, the board migh be given more power and be more attractive to ambitious people
* its possible (and necessary) to make public the voters list, so anybody can see it and eventually quote people with duplicate subscribtions.
* the guiding principles themselves where not approved by a vote, so we should not even ask to sign them.
You really think that Francis, Stephan, Andreas, Federico and I are favouring people who we know would vote for us when we check and vote internally on their level of contribution to the project ?
right now, I don't think. But this have been seen in many occasions (not only on web votes).
However, the main problem is probably not this. The board have a given idea of what membership must be. this idea can be challenged, but if the members only vote, they are probably of the same group. life is made like this: you always choose people near from you, so this kind of vote lead to less open mind (as a group)
I think we should see this vote as a way to *enhance* the community, both in quantity and in trust. I'm pretty sure most of the voters will have *more* interest in openSUSE after the vote than before. The more we can make vote, the more members we may have after that.
jdd
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
jdd wrote: | Pascal Bleser a écrit : |> - - a vote open to all ? worst of options IMO, as non-members could have |> an interest to vote for people who put the bar a lot lower for becoming |> member) | | ?
I meant that for a committee that would handle the membership status. I think Cornelius proposed having another group of people manage that, instead of the board.
My question was that in that case, how would that group of people be selected ? Voting ? Similar issues as with the board.
And if I may add, that group of people who'd grant the membership status (or not) would have more "power" than the board IMHO. Hence it needs even more legitimation.
| I think for the first vote, vote should be open to all the *volunteer* | (every people that asks to be able to vote).
That's one option. But you don't address the issue of fake votes.
My personal position atm (not that I wouldn't change my mind) is that fake votes make the vote illegitimate and is one of the worst issues we could have. If only one thing, then it's the fact that one individual only votes once. It's a question of fairness and equal rights to everyone who may vote. Someone with bad intentions (she would be cheating, after all) having a higher weight than others is not acceptable IMVHO.
Saying that "anyone may vote" is the preferred option, without a way to avoid or at least /strongly/ restrict the possibilities of cheating means that the fact to have a wide open process (which I agree is pretty much a good thing) is _more_ _important_ than avoid cheating.
It's really a question of balance. There is no silver bullet. We have to decide what is more important, or find a compromise (hence the 3 proposals in the original mail).
| * right now the board *power* is very small. In fact the board can only | suggest things to Novell or administrators. so the interest to be board | member is low (and I'm sure the work is not low - we must thank the | present board for doing that),
That is correct. Coolo is mostly arguing in favour of his "wide open" proposal by saying that the interest of cheating to be on the board should be very low, as it means a lot of work and not much "power" (if at all). This is clearly also my idea of the board: it is there to serve the community, not to drive it. When we start initiatives, it isn't to force something onto everyone, but to try to enhance the community (visibility, communication channels, functioning, etc...). In an ideal situation, even those aspects shouldn't come out of the board, but from the community in general.
| in a near future, the board migh be given more power and be more | attractive to ambitious people
I doubt that. The mission of the board is stated in the Guiding Principles: http://en.opensuse.org/Guiding_Principles#Governance
Just to be very clear about this: "The board of maintainers has the following tasks: * Act as a central point of contact * Help resolve conflicts * Communicate community interests to Novell * Facilitate communication with all areas of the community * Facilitate decision making processes where needed"
and: "The board should provide guidance and support existing governance structures, but shouldn't direct or control development, since community mechanisms exist to accomplish the goals of the project."
Whatever superpowers future boards may gain, they may *NOT* break the definition above. The Guiding Principles are pretty much our constitution. And the board may not break the constitution, at least not without a new version of the Guiding Principles, have it approved by the community, have people sign it again, etc... (a tedious process, nothing that may happen overnight).
This is also why people who are elected on the board *MUST* have signed the Guiding Principles: they must comply to the mission statement of the board they're part of.
| * its possible (and necessary) to make public the voters list, so | anybody can see it and eventually quote people with duplicate | subscribtions.
Indeed. It has been proposed already, and I haven't seen anyone object to having a public list of voters before the election.
| * the guiding principles themselves where not approved by a vote, so we | should not even ask to sign them.
See above, at least for people who run for the board.
But I think it also applies to voters. If you vote for a board, you first have to agree with the mission of the board... no ?
The Guiding Principles have been discussed widely on the -project list.
|> You really think that Francis, Stephan, Andreas, Federico and I are |> favouring people who we know would vote for us when we check and vote |> internally on their level of contribution to the project ? | | right now, I don't think. But this have been seen in many occasions (not | only on web votes). | | However, the main problem is probably not this. The board have a given | idea of what membership must be. this idea can be challenged, but if the | members only vote, they are probably of the same group. life is made | like this: you always choose people near from you, so this kind of vote | lead to less open mind (as a group)
Have a look at the current list of members. There are quite a lot of people there no one on the current board knows. We had to check the contributions for many applicants as we didn't know them. Which is a pretty cool side effect of the membership: you find people who are actively contributing to the community you never heard of, even after being in the community for quite some time. But I digress... :)
| I think we should see this vote as a way to *enhance* the community, | both in quantity and in trust. I'm pretty sure most of the voters will | have *more* interest in openSUSE after the vote than before. The more we | can make vote, the more members we may have after that.
Very true, but I'm not sure about the corollary. As said, having everyone who's in the community cast a vote would obviously be a good thing (although... the board mostly acts for active contributors -- not by definition, but in practice -- at least that's what I think, and in that case, it could be more logical that only members vote for the board.. well.. a bit difficult to explain in a few lines, and I don't want to hijack the thread). But I've yet to see mechanisms that would prevent cheating. And, to me, at this time, the most important aspect of an election is that there is no cheating. I'm open to suggestions though :)
cheers - -- ~ -o) Pascal Bleser pascal.bleser@opensuse.org ~ /\ http://opensuse.org -- I took the green pill ~ __v FOSDEM::23+24 Feb 2008, Brussels, http://fosdem.org
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On 18/03/2008, Cornelius Schumacher cschum@suse.de wrote:
If we are going this route I think it's mandatory that the process of giving out membership status is transferred to a committee which is not the board itself.
I'm kind of interested in why you think this, though. Do you think it's because selection of membership is open to abuse by the board, really?
Kind thoughts,
On Tuesday 18 March 2008 11:10:47 Francis Giannaros wrote:
On 18/03/2008, Cornelius Schumacher cschum@suse.de wrote:
If we are going this route I think it's mandatory that the process of giving out membership status is transferred to a committee which is not the board itself.
I'm kind of interested in why you think this, though. Do you think it's because selection of membership is open to abuse by the board, really?
If the board selects its own voters than this can create a potential conflict of interest in some cases. I think it would be better to avoid this from the beginning. I fully trust the current board to handle this conflict of interest in the right way, but why create a potentially problematic situation when it's easy to avoid that.
I also think that the time of the board can be spent better than on reviewing membership applications.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Cornelius Schumacher schrieb: | On Tuesday 18 March 2008 11:10:47 Francis Giannaros wrote: |> On 18/03/2008, Cornelius Schumacher cschum@suse.de wrote: |>> If we are going this route I think it's mandatory that the process of |>> giving out membership status is transferred to a committee which is not |>> the board itself. |> I'm kind of interested in why you think this, though. Do you think |> it's because selection of membership is open to abuse by the board, |> really? | | If the board selects its own voters than this can create a potential conflict | of interest in some cases. I think it would be better to avoid this from the | beginning. I fully trust the current board to handle this conflict of | interest in the right way, but why create a potentially problematic situation | when it's easy to avoid that.
And who should select the people who are eligible to vote? As yaloki already pointed out, just by appointing an election committee, the chicken / egg problem won't go away. Who will be in the committee? The members who have already been appointed by the board? Or just anybody? Who will control the committee? How do we trust the committee? There simply is no solution to that. We trust the existing board. Let use utilize this strength now. Once the board is elected, the problem is more or less gone anyways. Especially if you really restrict the board members to one term (which I strongly favor).
| | I also think that the time of the board can be spent better than on reviewing | membership applications. |
No, members are important to the community. It makes it stronger. And who is chosen to be a member is important in this sense. As the board is like a connection between community and novell, it is important to choose the members wisely.
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tuesday 18 March 2008 23:37:18 Felix-Nicolai Müller wrote:
And who should select the people who are eligible to vote? As yaloki already pointed out, just by appointing an election committee, the chicken / egg problem won't go away. Who will be in the committee? The members who have already been appointed by the board? Or just anybody? Who will control the committee? How do we trust the committee? There simply is no solution to that.
Even if it only would be a committee appointed by the board, it would reduce the problem as the people reviewing the membership applications would have less personal involvement.
We trust the existing board. Let use utilize this strength now. Once the board is elected, the problem is more or less gone anyways. Especially if you really restrict the board members to one term (which I strongly favor).
If you restrict the term to one term the problem is indeed reduced, but I think the question of reelectibility of board members should be independent of the question who is granting member status. There are good reasons why it would be good to have board members serve for more terms, e.g. to provide continuity, so restricting the options for that because of unrelated topics wouldn't be a wise solution.
No, members are important to the community. It makes it stronger. And who is chosen to be a member is important in this sense. As the board is like a connection between community and novell, it is important to choose the members wisely.
Members are important, but members shouldn't be chosen. The membership status should be granted according to transparent criteria. It would be fatal for the project, if the impression would be created that members are randomly chosen by a very small group of people.
* Cornelius Schumacher cschum@suse.de [03-19-08 11:15]:
If you restrict the term to one term the problem is indeed reduced, but I think the question of reelectibility of board members should be independent of the question who is granting member status. There are good reasons why it would be good to have board members serve for more terms, e.g. to provide continuity, so restricting the options for that because of unrelated topics wouldn't be a wise solution.
for a five member board, have one three year term and four two year terms with the two year terms occuring on alternating years, with the possibility of one of the two year electees being ?promoted? to fill the third year of the three year term in special circumstances.
On 3/18/2008 at 03:15, "Francis Giannaros" francis@opensuse.org wrote:
At the moment I prefer the membership only method because it's the simplest and is only really problematic in theory if the board is corrupt and unfairly give out membership in order to further their cause. In fact, this is pretty much the last premise I would ever accept as a possibility in the argument. Perhaps making the membership process even more transparent could help here, and I'd be willing to listen to any ideas about that.
I think the influence of the board on the voting mechanism can be somewhat reduced if any board member can be only 1 or maybe 2 consecutive periods in the position of a board member. Then we already rule out that they only 'grant membership' to users that would vote for them. Not being able to be re-elected makes this usecase void.
I would not go as far and block them completely after several periods, but having the board remixed once in a while is probably not bad neither.
Dominique
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On 17/03/2008, Andreas Jaeger aj@suse.de wrote:
We had tonight on the opensuse-project IRC channel a lifely discussion about the board election. The central question was who can vote for the openSUSE board. We haven't come to a real conclusion yet, see below for some of the comments that were discussed. We'd like to continue the discussion here on the mailing list and meet again in two weeks on the openSUSE-project IRC channel.
For reference, here's a meeting log thanks to jervine: http://en.opensuse.org/Board_election_proposals/meeting
Or in slightly more readable html: http://giannaros.org/suse/logs/board-election-meeting.log.html
Kind thoughts,
On 17/03/2008, Andreas Jaeger aj@suse.de wrote:
- basically three different proposals:
- only members
- anyone
- members + non-members members vouch for (web-of-trust alike)
My thoughts:
- Anyone
(Anyone at all, Or any identity that has signed the guiding principles)
This is so open to abuse that IMHO doing it this way is almost as bad as no elections at all. One can have no confidence that the result is indicative of anything.
This is the only method I object to.
- Only members
By far the simplest solution. Might also encourage people to apply for membership. This is how pretty much every other project does it.
The main criticism of this approach is that members are those who have a "continued and substantial" contribution to the project, wheras ideally any contributor should be allowed to vote, however small the contribution.
- members + non-members members vouch for (web-of-trust alike)
Allows nearly all contributors to vote if they wish. However, it is more open to abuse, particularly by members who would essentially have infinite votes if they wanted to abuse their position.
This option could work. But it requires a lot more planning. It would be best if the electoral role were publically viewable so that people can review those registered to vote to help reduce fraud. It seems almost like creating a second tier of membership though.
The biggest problem with this option is the increase in complexity. You have to consider all sorts of implementation details like who members are allowed to allow to vote. The bikeshedding on this topic is bad enough already, the more details there are to be decided the more difficult a decision is.
-- Benjamin Weber --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
-
Andreas Jaeger -
Benji Weber -
Bryen -
Cornelius Schumacher -
Dominique Leuenberger -
Felix-Nicolai Müller -
Francis Giannaros -
jdd -
Joe 'Zonker' Brockmeier -
Kevin Dupuy -
Martin Schlander -
Michael Loeffler -
Pascal Bleser -
Patrick Shanahan -
Rajko M. -
Richard (MQ)