From: Ian Kelling Date: Thu, 3 Apr 2025 15:02:33 +0000 (-0400) Subject: Add details and somewhat narrow the scope X-Git-Url: https://iankelling.org/git/?a=commitdiff_plain;h=cbb5c62aca09bf15ce6c66117313069a589fd1be;p=election-results-2025 Add details and somewhat narrow the scope --- diff --git a/osi-2025-marked-ballot-example.png b/osi-2025-marked-ballot-example.png new file mode 100644 index 0000000..93d05c0 Binary files /dev/null and b/osi-2025-marked-ballot-example.png differ diff --git a/osi-2025-unmarked-ballot-example.png b/osi-2025-unmarked-ballot-example.png new file mode 100644 index 0000000..3bb62cb Binary files /dev/null and b/osi-2025-unmarked-ballot-example.png differ diff --git a/petition.md b/petition.md index 0bccfd8..89ce0e9 100644 --- a/petition.md +++ b/petition.md @@ -1,54 +1,157 @@ -DRAFT ONLY: DO NOT SIGN THIS YET -================================ - -This petition is in **draft form only**. Please do not sign this yet. As you can -see, there are already pending merge requests for changes, big and small, to this -document. We started drafting this as a private group of small individuals, but -we decided to move to a transparent method ourselves and it's here, but it's -not ready to sign (even though there are actually pending sign merge request, I'll -get back to those folks when we're ready). - - -Petition to Open Source Initiative To Publish 2025 Board of Directors Election Results -====================================================================================== - -in the interest of community trust and transparency of process, we the -undersigned petition the Open Source Initiative (OSI) to release the -anonymized ballot data for OSI's 2025 Board of Directors election — -including votes for all candidates who received any nomination. - -The basis for our petition is that three candidates’ votes were not counted. -OSI states these candidates failed to follow a qualification procedure. -Multiple candidates have confirmed that they were informed of this -qualification procedure only after voting closed. - -[OSI's announcement](https://opensource.org/blog/announcing-the-new-directors-of-osi-board) -indicates these three candidates were removed from those already-cast ballots -before vote tabulation. (Please also -[see *Linux Weekly News*](https://lwn.net/SubscriberLink/1014603/ac0cfc0a74755501/) for -relevant press coverage of this election.) - -We are grateful for OSI's advisory election process. While only advisory, -the process includes community and builds trust with Affiliates and -Members. To retain that trust, the process must operate transparently and -respect the electorates. OSI's credibility is at risk if it violates its -mission — which includes “distributed peer review and transparency of -process” (as well as consensus building and good governance). Therefore, OSI -should candidly report the full voting results (by release of the anonymized -ballot data) and honestly and directly, rather than obliquely, describe the -differences between the election results and the actual Board appointments. - -The signed document is -[available on Codeberg](https://codeberg.org/OSI-Concerns/election-results-2025/src/branch/main/petition.md). -Supporters are encouraged to submit their signature through a pull -request. If you have a relationship with OSI (e.g., a Member (even a “Basic” -one), Former Director, etc.), pease indicate your relationship to OSI in your + +Petition to Open Source Initiative and To Publish 2025 Board of Directors Election Results +================================================================ + +We the undersigned petition the Open Source Initiative (OSI) to release +the complete and accurate results of the 2025 Board of Directors +Individual and Affiliate elections. + +In 2012, the OSI began running annual community elections, and for many +years, those elections have determined the majority of its board +of directors. In 2025, OSI published [overall election +information](https://web.archive.org/web/20250319054143/https://opensource.org/about/board-of-directors/elections) +and eventually announced that [OSI Affiliate organizations would +nominate and elect two +candidates](https://web.archive.org/web/20250323020823/https://opensource.org/about/board-of-directors/elections/affiliate), +and [OSI Individual members would nominate and elect one +candidate](https://web.archive.org/web/20250126161015/https://opensource.org/about/board-of-directors/elections/individual). Community +elections have always been advisory to the board, but OSI has +consistently written about board members as elected by Affiliates +members or Individual members, including calling them Individual +directors and Affiliate directors. + + +In 2025, the polls for both elections closed on March 17th. On March +21st, OSI posted what it claimed to be the "Complete election +results"[4], but unfortunately, they plainly are not complete and they +misrepresent the votes that were cast. Essentially, OSI altered the +ballots to erase 3 of the candidates and shifted voters ranked +preferences to fill in the gaps. We don't know if this is a very +unfortunate mistake with some good intentions, but we have seen multiple +OSI representatives suggesting there is nothing wrong. + +OSI has consistently stated[1] that the online elections are being run +with Scottish STV (Single Transferable Vote) rules, and included a link +to [OpaVote, detailing +STV](https://opavote.com/methods/single-transferable-vote#scottish-stv) +. OpaVote is a service which facilitates online elections and which OSI +used for the collection of votes. Understanding how OpaVote was used +helps understand exactly what went wrong. On OpaVote, anyone can sign up +as an "election manager", and they get access to self-service forms +which facilitate running an online election. For each of the 2 +elections, OSI logged into opavote.com as an election manager, uploaded +a list of voter's email addresses, a list of candidates, selected +Scottish STV rules, and clicked a button open the polls. OpaVote then +emailed each of the voters a unique ballot url on opavote.com. Here is +[a screenshot of an empty ballot](/osi-2025-unmarked-ballot-example.png) +of the OSI 2025 Individual election, and [a filled in +one](/osi-2025-marked-ballot-example.png). The only way to fill in the +ballot is by specifying a 1st preference and then optionally 2nd, and +then optionally a 3rd, etc. A sample ballot form from the OSI 2025 +Individual election [is +here](https://web.archive.org/web/20250327040900/https://opavote.com/en/vote/5182543794536448?p=1), +but beware: it requires running a nonfree javascript program. After the +polls close, OpaVote provides the election manager an option to publish +the voting results to the voters. OSI did not use that option. Instead, +it seems they downloaded the ballot data, altered the ballots, and used +this to publish altered election results. In OSI's results announcement, +it is stated that [that 3 candidates were not included in the final +tally because of a failure to sign an agreement with +OSI](https://opensource.org/blog/announcing-the-new-directors-of-osi-board). They +did not state which candidates had been removed from which +elections. What we quickly figured out is that in the individual +elections, Richard Fontana and Bentley Hensel were erased from final +ballots and apparently replaced with lower preference votes or if no +lower preference had been specified, then no preference at all. For the +Affiliates election, the same procedure was used to erase Bradley Kuhn +from the results. + +We content what hopefully seems obvious: the correct and true election +results include the actual preferences that voters marked on their +ballots. There are several points of evidence beyond common sense to +show this. The link that OSI [uses to describe STV](https://opavote.com/methods/single-transferable-vote#scottish-stv) +contains a set of rules, and the rules +are clear that altering ballots is not an allowed part of the +procedure. This is nothing unique to STV, this is widely accepted +practice in elections. OpaVote states that it follows the relevant +parts of the actual Scottish legislation +https://www.legislation.gov.uk/ssi/2007/42/made/data.xht?view=snippet&wrap=true +. That legislation is quite clear that there is no case in which a +voter's list of preferences may be altered, and the full election +results which must be announced are: + +> give public notice of– +> +> (i)the name of the candidates elected; +> +> (ii)the number of first and subsequent preference votes for each candidate; +> +> (iii)the numbers of ballot papers transferred and their transfer values at each stage of the count; +> +> (iv)the number of votes credited to each candidate at each stage of the count; +> +> (v)the number of non transferable ballot papers at each stage of the count; and +> +> (vi)the number of rejected ballot papers under each head shown in the statement of rejected ballot papers. + +Furthermore, we contacted OpaVote to find out their opinion of this +election. In an email, they stated "I understand your point; you want +the full results of the election to be published in the interest of +transparency." OpaVote cannot force OSI to publish the full results, but +they can see, just as we do, that contrary to [OSI's official +statement](https://opensource.org/blog/announcing-the-new-directors-of-osi-board), +the "complete election results" have not been published. + +According to the Scottish STV rules, even in the event of a candidate's +death, if the full election results have already been determined, they +must be announced. And if a candidate's death is known before an +election's results are determined, the election is immediately canceled, +the ballots are never counted, and the election is rerun. The three +candidates have told us that they were disqualified after the polls had +closed. At that point, the full results had already become available to +OSI. Whatever disagreement OSI has or had with the three candidates, the +voters have no part in it. We only ask OSI to live up to a simple and +basic obligation of any election. + +Some of us have heard from OSI what sounded like the theory that if OSI +had run an election without these 3 candidates, then voters would have +voted with the same preferences as the altered ballots, so the voters +haven't lost anything. We reject that idea. First of all, we've already +heard that additional candidates would have run if one of the existing +candidates had not. Secondly, there are several common and legitimate +voting strategies which do not conform to that assumption, and the +voters do not deserve to be punished for using them. The voters trusted +OSI to publish their true votes and have every right to feel that trust +has been broken. + +We also know that if the OSI board were to override the will of the +voters by invoking the advisory nature of the elections, that would be +legitimate. But at this point, we cannot agree with OSI's claim that any +candidate has been elected by the Individual or Affiliate voters until +we know the true election results. + +An [article at *LWN*](https://lwn.net/SubscriberLink/1014603/ac0cfc0a74755501/) +covers many other details about related contention around this election. + +Many of us believe there are OSI board members and perhaps election +candidates who will hear our petition and realize that somehow, these +two elections have gotten off-track and steer OSI to do the right thing. + + +Petition detail +---------------- + +This is an open petition. The signed document is available on Codeberg +[here](https://codeberg.org/OSI-Concerns/election-results-2025), and +supporters are encouraged to submit their signature through a merge +request. If you have a relationship with OSI (e.g., a Member, Former +Director, etc.), pease indicate your relationship to OSI in your signature. -It's suggested that OSI Affiliates have their Affiliate Representative sign -on behalf of the organization; they and other Affiliate employees/volunteers -can of course sign separtely on their own behalf, too. +We ask that OSI Affiliates have their Affiliate Representative sign on behalf +of the organization; they and other Affiliate employees/volunteers can of +course sign separtely on their own behalf, too. Respectfully, -FIXME_SIGNATORIES HERE +SIGNATORIES HERE