The Signpost

Op-ed

As one thousand of us requested, Superprotect has been removed


Related articles
Superprotect

As one thousand of us requested, Superprotect has been removed
11 November 2015

Wikimedia Foundation finances; Superprotect is gone
4 November 2015

Superprotect, one year later; a contentious RfA
12 August 2015

Media Viewer—Wikimedia's emotional roller-coaster
27 August 2014

Media Viewer controversy spreads to German Wikipedia
13 August 2014

One Sunday afternoon in 2014, without fanfare or advance warning, the Wikimedia Foundation deployed "Superprotect" – a new user privilege that enabled the Foundation's paid staff to overrule elected volunteer administrators. Without pausing to describe what circumstances would merit its use, the Foundation immediately used Superprotect to unilaterally enforce the rollout of the Media Viewer software on the German Wikipedia. Several local communities had questioned Media Viewer’s suitability for broad release. The Foundation forcefully asserted its own authority – a climax in several years of disputes with volunteers over various software features. Then-Trustee Samuel Klein later characterized the move, saying it "opposed our wiki values, distracted the projects, and did not solve any pressing problem."

I was moved to urge the Foundation to remove Superprotect, and to disavow its newly asserted authority. I wrote a letter to this effect, invited others to sign it, and delivered it in September 2014 to the organization’s ten Trustees and top two executives.

The letter’s popularity was substantial and diverse. More than a thousand people have signed it; they hail from dozens of language communities and multiple Wikimedia projects. Yet for more than a year, the Foundation declined to publicly acknowledge either the letter or the problems with Superprotect. But what a difference fourteen months can make: on November 5, 2015, Superprotect was removed, and the executive director publicly addressed the issue for the first time, declaring that its "precedent of mistrust" had to be reversed.

What changed in that time? Will the announcement have the desired impact on restoring trust and effective collaboration between the Foundation and Wikimedia volunteers?

Elections have consequences

All that changed, it seems, was Foundation leadership. The executive most strongly associated with Superprotect left his position in April, and in June all three community-elected Trustees lost their bids for reelection to candidates who opposed Superprotect. Several months later, with no further visible changes beyond the announcement of a new software development process, the Foundation regretted having established a precedent of mistrust. Superprotect was removed as suddenly as it was introduced, and with little more fanfare.

I have had only one conversation with a recipient of the letter. A Trustee I spoke with a few months after delivering the letter characterized it as conveying frustration and anger (which, I submit, it did not); acknowledged that she did not remember what specific actions it requested; and opined that those signing the letter probably didn't know what it said, and signed it as a proxy for various complaints.

About a month ago, I heard similar things from a senior staff member working on related issues: he also did not distinctly remember the two requests, and also felt that some of those signing the letter did so without reading it carefully. As I told both the Trustee and the staffer, I believe their words reflect a major failure in their leadership. Anyone seeking to improve Wikimedia's social dynamics should remember two concise requests that generated substantial support. Even if some of the signatories were no better informed about the letter's contents than the two Foundation personnel, many signed it with eyes wide open. Our request deserved timely consideration.

As the Foundation seeks to move forward by finally removing the Superprotect user right – granting at least some of the letter's provisions – what are the implications going forward? Is this yet another unilateral Foundation action, or did the volunteer-driven letter play a role? To assess this question, we should revisit the exact things the letter said – and the things it did not say. Let’s begin with a review of what I intended when I composed the letter.

What I sought to accomplish with the letter

The frustration and anger the Wikimedia personnel identified do, of course, exist. In the weeks before Superprotect was implemented, I felt them deeply. Those feelings were justified: with the Media Viewer, the Foundation had – as with many previous software releases – suddenly activated a new product that was far from ready, and which ran afoul of important Wikimedia values (not to mention, it would seem, intellectual property law) in a number of ways. They did so, in part, to meet a deadline that could and should have been extended, in order to better address the feedback volunteers were generating. The efforts of myself and many other volunteers to get the software rollout reversed were repeatedly dismissed by Foundation staff as self-interested, despite my many years of full-time work (much of it on the Foundation's payroll) to improve Wikimedia's recruitment and retention of contributors. My time and expertise were being treated as a resource without any particular value.

But in my years of Wikipedia editing, I have learned an important lesson: however justified frustration and anger might be, allowing them to guide your action is rarely productive. So I took a step back, and considered my options. I sought the advice of respected friends and colleagues – especially a few who had no particular stake in Wikimedia, who could help me think about the big picture. They reminded me that any effective advocacy had to reach beyond the immediate conflict, and invoke the core principles that united us all – volunteers, Foundation staff, and readers.

What the letter did and did not say

In order to place that common ground front-and-center, I began with the words "Along with you, we envision...", and then quoted from the community- and board-approved Wikimedia vision statement, which states: "Imagine a world in which every single human being can freely share in the sum of all knowledge. That's our commitment."

It's one thing to begin from a place of alignment, but it's quite another to carry that spirit through to a specific requested action. I considered, "we demand that the Foundation offices be dismantled brick by brick before another line of code is written," but cooler heads prevailed. With support from many quarters, I kept pushing back to Wikimedia's core needs around software development, ultimately settling on two simple requests:

… the following are necessary conditions for a healthy and productive path forward. … :
  1. … Foundation should remove the "superprotect" ... and
  2. … Foundation should permit local projects ... to determine the default status of the Media Viewer uninhibited.

I bold the word "necessary" above for a reason: the letter aimed to establish conditions that could permit all parties to work together to solve complex problems. It never aimed or promised to identify conditions that would solve the problems; it was not that ambitious. Sufficient conditions for a healthy and productive path forward were not proposed; that's the work that could commence once a reasonable baseline had been established.

The two requested actions are not arbitrary; neither action would carry much meaning without the other. The first is a technical action, while the second is a social or rhetorical expression of the related principles. The second is what would form the basis for future accountability. Without #2, the Foundation could comply with the #1, but then create "super-duper-protect" – and even if it didn't do so, the very possibility of a different technical obstacle to enforce its will would be enough to demotivate volunteers like me.

My emphasis on specific, achievable goals, clearly tied to the core vision that unites all Wikimedians (volunteer and staff alike), was crucial to the popularity of the letter. A thousand signatures is far more than I hoped for; but as the signatures and comments came in, I heard from many signatories who felt the letter wasn't critical enough of the Foundation, and also from many who liked the Media Viewer software. By carefully choosing specific requests, I established a core position that appealed to a broad range of viewpoints in the Wikimedia community.

Comparing the requests to the Foundation's recent statements

Let's look at two of the Foundation's various statements this month, as they announced the demise of Superprotect:

  1. At the November 5 Metrics and Activities Meeting, Wikimedia's executive director stated:
  2. On November 9, the engineering community manager stated:

I urge you to take a moment and compare numbers one and two in each section. The first request was clearly granted.

But what about the second? The statement quoted – which came not in the initial email message, but in response to pointed questions – addresses how the principles relate to a specific kind of software deployment or configuration – those which are determined by an admin-editable wiki page.

Are all software deployments in which the volunteer community has a stake carried out by wiki pages? Will they be in the future?

I don't know the answers to these questions. We have heard a strong assertion about how principles apply to specific set of pages. Statement #1 above offers assurance that volunteers may participate in a discussion led by the Foundation, but that framing is far out of step with the reality of how Wikipedia and the other projects have achieved their success. So I'm not sure: 14 months later, has the Foundation heard the complaint?

Was the letter finally successful?

Following the Foundation's announcement, I suggested to the first twenty signatories of the letter that we should declare it a success. But nearly everybody who answered said "no," with the following reasons:

(a) The Foundation took too long (about fourteen months rather than, say, fourteen minutes) to fulfill the requests. (b) The Foundation never formally acknowledged the letter's existence. (c) There was initially no statement about local administrators' ability to assess and enforce consensus around software changes. (d) When there was a statement, it was not from top leadership, and it may be at odds with routine practices that don't specifically involve Superprotect. (e) The core problems around software releases and Foundation-volunteer relations remain, and may even be "unfixable."

These are all worthwhile points; I will respond briefly to each:

  1. The Foundation took too long. Any responsible executive or Trustee should have immediately recognized how badly Superprotect would impact trust, in a system that is deeply rooted in trust. It should not have been deployed to begin with, and every minute it survived – before or after the letter – was a minute too long.
    But as bizarre as Superprotect's fourteen-month tenure may be, we cannot change the past. The necessity of its removal is as necessary now as it ever was. In evaluating the letter's success, we should not be too distracted by the many months in which it was unsuccessful.
    Furthermore, the negative impacts of the extended tenure of Superprotect affected the Foundation, as well as Wikimedia as a whole; four good leaders in the organization are now gone, arguably due to Superprotect. The mistrust has hurt many parties, but dwelling on that point accomplishes nothing.
  2. The letter was not acknowledged. This is irritating – perhaps to me more than anybody, given my focus in the letter on clear, limited, reasonable requests. But this shouldn't be about me, or any of our feelings as individuals. It's always been about getting the Wikimedia effort back on track. I devoted my time to this because I believe in an outcome, not in a quest for acknowledgment. If the Foundation's goal is truly to rebuild trust – which goes somewhat beyond the scope of what this letter sought – its leadership might do well to consider the body of research around trust and acknlowledgment, which executive coach Charles Feltman sums up as: "The only known antidote to betrayal...is to acknowledge it and apologize for it."[1] Superprotect was a betrayal of trust. Without acknowledgment of that basic complaint – without Foundation leaders reflecting back what they heard, what resonates, and how it impacts the path forward – I suspect questions about whether the problems have been understood will only continue to fester.
  3. The removal of Superprotect was poorly communicated. The Foundation's statements have been rather random, with bits and pieces being communicated in numerous venues, by various staff and Trustees. All parties would be better served by a single, thorough, carefully written announcement. I would counter that communication in our wiki world is complicated. While I'd like to have a Foundation that is able to rise above the fray with, for instance, a clear press release or blog post outlining principles, in our community, we are pretty good at parsing text and identifying themes. So this is not a big deal in itself.
  4. Senior Foundation leadership should have made the announcement. The Foundation's position would be clearer if it came initially, and comprehensively, in a statement from the executive director and/or the board of trustees. And if indeed there are routine practices that violate the principles stated by the engineering community manager, then yes, there is a legitimate concern here.
  5. Not everything has been resolved. This is true. In particular, Foundation leadership seems to still undervalue the significance of Wikimedia volunteers' efforts. But look closely at the text we agreed to: we never spoke of total resolution in the letter. We asked for a couple of actions that were necessary before we could solve the underlying problems, but we never claimed they would be sufficient in themselves. There is still work to be done to establish a better process, but we always knew there would be.

It's murky – so I would say, "no."

I take the various objections seriously – there are many reasons to be dissatisfied with the Foundation's handling of Superprotect. Some of the objections are incidental, and should be deferred until later. But others are compelling. Why have we not heard a more comprehensive statement about the community's value in guiding software decisions? Is it because the Foundation doesn't perceive that value, or because its leaders think it "goes without saying?" The Foundation created an opportunity to speak with clarity – by announcing a decision on Superprotect – and then declined to address the context. That seems very odd, and makes me think the letter's requests are as relevant as ever.

Was the Wikimedia Foundation's response enough?
Enter your response in this poll

What do you think? Are you satisfied with the Foundation's response to the issues brought up in the letter? Please weigh in on the letter's talk page. And if you feel the conditions are right, take a look at the product development process outlined by the Foundation – your opinion, feedback, reasoned complaint, or endorsement there will help establish whether or not the next step is attainable.

  1. ^ Feltman, Charles (2009). The thin book of trust an essential primer for building trust at work. Bend, Oregon: Thin Book Publishing. ISBN 9780982206836.


Pete Forsyth has been a Wikipedia editor since 2006 and runs a Wikipedia training and consulting business, Wiki Strategies. He worked for the Wikimedia Foundation from 2009 to 2011. The views expressed in this editorial are the author's alone and do not reflect any official opinions of this publication. Responses and critical commentary are invited in the comments section.


+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.
  • It's a good result, if not the best result. Superprotect was symbolic and that symbol has been symbolically removed. A clear statement of good will has been made.
We often (us humans, whether in groups or individually) fail to reap all the benefits of a compromise we agree to, or an action we take because we don't clearly enunciate why what we did was wrong, or the lessons we have learned. This might be considered to be face saving. In fact it often mean that we are not getting the benefits of the sacrifice (in some cases only of "face") we have made.
Let us take this step forward as positive and continue in dialogue with the Foundation - robust dialogue where necessary - to advance the objectives of the community in the furtherance of open knowledge.
All the best: Rich Farmbrough, 21:19, 15 November 2015 (UTC).[reply]
  • A clear statement of good will has been made. That's it for me. It was not a painless move, but a necessary one. I am indeed hopeful about the future of this "dialogue with the Foundation". José Luiz talk 21:55, 15 November 2015 (UTC)[reply]
  • It may sound like a positive act but history has shown that this 'good faith' won't last. With Wikipedia and all its offshoots being Internet based electronic products,its management company of now over 200 employees does not even physically sit in the same building. It's time this huge and cumbersome machine the 'WMF has become stops its navel-gazing in order to come out of its bubble for a moment and realise that there are tens of thousands of dedicated, active editors out here, some even putting in a 20 - 40 hour week for free, who are providing the Foundation with its raison d'être. And then we can really say: A clear statement of good will has been made. Kudpung กุดผึ้ง (talk) 22:26, 15 November 2015 (UTC)[reply]
  • Hi, as point of contact for this announcement, there are a couple of observations that I would like to share. "All parties would be better served by a single, thorough, carefully written announcement." mw:WMF product development process/2015-11-05 is the canonical location of the announcement, and it offers a central location for Discussion. The announcement and its Q&A were echoed in several channels. There was feedback in several places, I have been the single point of contact replying, and I have been updating the Q&A accordingly. The only other WMF speaker making statements about the announcement was Lila, echoing the news on the same day, at an appropriate venue such as the Metrics meeting. "The Foundation (...) declined to address the context." The context of Superprotect is the WMF product development process. This process documented and agreed with the communities is the right tool to prevent and address situations like the one that lead to the use of Superprotect. The draft is under discussion, and there you can find already now several WMF voices discussing openly from their roles as VP of Product, Director of Community Engagement (Product), Lead Software Architect, Agile Coach, Software Engineer... Owners and stakeholders addressing topics like Between "Concept" and "Plan" - Prioritize? or Community input prior to build phase is exactly what needs to happen in order to resolve "the underlying problem of disagreement and consequent delays", from which Superprotect was one manifestation. Peteforsyth and anybody else interested in this discussion, if you have questions or feedback about how the WMF should develop and deploy software, mw:Talk:WMF product development process is the best venue to discuss and agree on the best implementable solutions.--Qgil-WMF (talk) 07:44, 16 November 2015 (UTC)[reply]
To me, the core issue here is that the WMF is treating this as some kind of disagreement over software development practices (just look at the title of the announcement, and the fact it was made on the MediaWiki wiki, which is focused on development of the MediaWiki software), when it isn't. It's a social conflict over the role and powers of the WMF. Few people care about intricacies like how software changes are deployed. What the people who protested over superprotect want is a frank acknowledgement from the WMF (and probably from the Executive Director or someone else high-ranking, who is seen as speaking for the organization as a whole) that they screwed up and won't do it again, and maybe a discussion about when and how the WMF should be able to overrule the community. This whole imbroglio has been a specific demonstration of what I see as the problem with the WMF's mindset. The WMF sees itself primarily as a software development house, which is why its default answer to every problem seems to be trying to cook up some whiz-bang piece of software that will fix an issue. But not all problems are technical problems. The projects are communities, not just bytes sitting on hard drives, and in my view the WMF needs to shift more of its focus towards actively fostering healthy communities instead of just being "the people who run the servers". Of course, in doing so, the WMF needs to be aware of the existing bad blood between it and the projects, and needs to make an effort to patch up relations and work in collaboration with the communities instead of imposing things. I think things like the Community Tech team are a step in the right direction, though I'd like to see something not narrowly confined to the area of software development. --71.119.131.184 (talk) 11:16, 16 November 2015 (UTC)[reply]
"if you have questions or feedback about how the WMF should develop and deploy software the best venue to discuss and agree"... is a FLOW talk board!! (mw:Talk:WMF product development process) Brilliant! Good god. --Atlasowa (talk) 11:51, 16 November 2015 (UTC)[reply]
@Qgil-WMF:, I have a strong positive impression of your role in addressing Superprotect; thank you for your efforts. But there is a fundamental problem with your urging people to use what you call a central place for discussion: the letter in question identifies conditions its signatories consider necessary for having that kind of tactical discussion. The issues, in my view and (I believe) those others who signed the letter, remain at the strategic level, and involve (as stated by the IP immediately above) social dynamics that extend beyond software development. I will not be engaging in depth in evaluating the any proposed development system until I am satisfied that these conditions have been met. I suspect you will find that others who have signed the letter will also decline to engage strongly there, unless they consider the letter's requests to have been granted.
Also, you state that Lila Tretikov was the "only" other WMF person making statements about this; but Trustee Dariusz Jemielniak, and employees Dan Garry and Brion Vibber also responded substantively on the Wikimedia-L list, other staffers have commented on the MediaWiki page you link, and I believe still others have commented in other venues. If you are making some distinction between "official" comments and some other kind of comments, it is opaque to me -- partly because neither your initial announcement nor Lila Tretikov's carried any trappings of officialdom (e.g., press release, blog post, statement that it is a considered action of the WMF, etc.) Could you clarify? -Pete (talk) 13:19, 16 November 2015 (UTC)[reply]
You seem to be forgetting the tens of thousands of Users who did not sign the letter, a letter which appeared either petty in substance or petulant in demand - just like your 'we won't talk to you until all our demands are met' looks petty and petulant. Alanscottwalker (talk) 12:55, 18 November 2015 (UTC)[reply]
@Alanscottwalker: I can't really speak to your perceptions, except to point out a couple things that you seem to have missed:
  • Three of the five poll questions I put forward were addressed to (as you put it) the tens of thousands who didn't sign.
  • I talk to Wikimedia Foundation staff frequently, and have not refused to do so. What you identify as personal petulance is actually straightforward prioritizing. As a general rule, it doesn't make much sense in any context to delve into specifics, when there are unresolved broader issues. -Pete (talk) 20:43, 20 November 2015 (UTC)[reply]





       

The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0