By The Masked Maggot
Just as the Wikimedia Foundation (WMF) was wrapping up its annual show of “wiki-love” and idealism at London’s Wikimania 2014, the latest round of tensions between the foundation and the volunteer community reared its ugly head. The trigger this time was the forced introduction of the Media Viewer to the German Wikipedia, after the community of editors that work on that version voted to disable it. While a slight change to the presentation of images might seem like a minor issue, it points to a deeper and widening rift between the offices in San Francisco and the global community which they serve.
The WMF’s gradual grab, or How the Wikimedia Foundation is harming Wikipedia
While there are all sorts of issues worth complaining about, Wikipedia really is an amazing thing. For 13 years, a vast collection of unpaid and unsung volunteers have written, organized, and advocated for an online encyclopedia-like website that pretty much everyone uses, naively or reluctantly. It’s been a long run where volunteer “editors” have not only created the content, but have also been enfranchised to have a say in how the site is run. Wikipedia’s “consensus” model, for all its problems, has been the real driving force behind getting the product to market, and the “Stone Soup” way that everyone brought something to the table (from content to structure to presentation) worked a bit better in practice than it should have in theory.
Those happy days seem to be coming to an end, and while the seeds for that end were planted at the beginning, the fruit is only becoming apparent now. The Wikimedia Foundation, which accepts all the donations and was (so the Wikipedians thought) dedicated to supporting the volunteers who write Wikipedia’s articles, has over the past year or two managed to alienate its volunteers. The irony is that many of the things causing tensions between the WMF and the volunteers are actually things aimed at improving the interface and attracting more editors.
Making editing easier, in theory
During the first several years of Wikipedia and the WMF, the “coding language” behind the articles was mostly created by volunteers, both at the level of the software, and through creative use of the wiki markup language to create “templates”. This made editing rather complicated over time, as more and more templates became standard and in some ways required.
This wasn’t such a bad thing in the heady days of the mid to late ’00s, because there was a large community of “gnomes” who would follow content contributors along and add the appropriate templates. For example, if someone created an article about a species of fungi or lizards, someone would helpfully follow along and add a “taxobox” and categories to provide a uniform look to all the fungi and lizard pages, while others would contribute more information about the taxon. In many cases someone would even add references to the taxon in movies or comic books, which while not exactly what you’d expect to find in an encyclopedia, did lend Wikipedia its particular charm.
The problem with the stone soup method was that the Wiki Markup Language (WML) was very spare, and mostly only updated when a volunteer coder added a feature they thought would be particularly useful. So, for example, that “taxobox” became a very large and detailed “template” that needed to be added to each page and properly filled in, and even the templates used to add references to articles became complicated and wordy, making it harder and harder to actually read the text while viewing the edit window. This became a higher and higher hurdle for new contributors to get past, and also may have contributed to the jumbled prose since it became difficult to read while writing.
By the end of 2011, many were beginning to complain about the steep learning curve needed to add content. Jimmy Wales even declared his userpage a template-free zone. The calls for an easier editing interface got louder and louder, and the WMF began using its newly acquired millions to create one, eventually introducing the Visual Editor (VE).
A Visual Editor
While VE was introduced to great fanfare, it was an epic flop with the volunteer community. The software was buggy, causing all sorts of problems to pages it was used on. The editor was also unable to add or edit the many templates found on almost every Wikipedia article. Worst of all, it was imposed against the will of the volunteer community.
The flaws of the Visual Editor weren’t just a problem for the people who tried to use it, because the bugs actually caused “damage” to the pages it was used on by “breaking” templates and links, and even inserting random Unicode characters such as chess pawns: ♙♙♙. The gnomes and patrollers suddenly needed to waste even more time looking at every edit that had been tagged a visual editor edit, closely looking over each change to see if any brackets were missing or random characters added.
The Media Viewer, Round Two
As reported here on Wikipediocracy a few weeks ago, the Media Viewer’s introduction on the English Wikipedia was not at all popular with the volunteers, or at least the small subset of volunteers who spoke up about it on the Request for Comments (RfC). The foundation’s representatives, and Deputy Director Erik Möller in particular, dismissed the discussion because not enough people had participated in it. The RfC wasn’t very well advertised, and the community had perhaps been demoralized by the fight over VE, as well as signs from the WMF that “Flow” (a potentially disastrous “improvement” for discussion pages) was going to be imposed in similar fashion following their self-imposed deadlines. Meanwhile on the German Wikipedia, a much larger RfC was underway, and the result there was far more definitive: the German volunteers did not want this feature, preferring the simpler presentation of simply viewing the file pages.
At this point, the volunteers became truly infuriated. The Wikimedia Foundation’s technical department head Erik Möller (a German) was blocked by a volunteer admin. The mailing lists erupted into displays of anger, frustration, and demoralization. Some users even expressed the feeling that the WMF was intentionally alienating the volunteers in favor of its paid staff, and perhaps its Wikipedian in Residence program. One wrote on the “meta” RfC:
They are not worried about chasing the active users away. They don’t want us. They want new users, different users, facebook users, hyperactive people, that make pictures anywhere and edit with their mobile, that add to the article what goes through their mind at the moment, not what is written in reputable books and found out after a long and tiring reasearch. They want masses, a movement, millions of bot-articles. And they hope that the new masses will finally chase the old users away, those who are questioning their godlike decisions and their godlike self-declared superpowers. You see this in every answer: “You, who have argued and voted, are only the old editors, but we care about the silent readers and the fictive new editors. So of course we stand above your arguments and elections.” Our opinions are not in the strategy focus anymore and can be ignored without consequences. In the view of the WMF, the readers are not reading the articles because we editors have written them in a way that is well appreciated, such that it raises multi million dollars of donations a year. The readers visit them because of the great software that WMF develops. They would read anything, if it was just presented by their great software.
Intentional or not, the WMF seems on course to severely weaken its most important asset: the volunteer contributors who have created the product they ostensibly exist to promote.
The crisis on the German Wikipedia remains unsettled at the date of publishing (as it is on the English Wikipedia as well).
The Scramble, and the Fallout
|The official and semi-official line from the WMF has been bizarre to say the least. In response to a comment about the imposition of VE and other “improvements” on co-founder Jimmy Wales’s page, Jimmy had this to say about the “spiritual side” of the issue:
Less interested in the spiritual but rather reframing the issue in an e-commerce light, incoming Executive Director Lila Tretikov wrote:
Our site, technology and content have far reaching impact and scale. Today, we do not fully appreciate ourselves as a top-5 website from an operational standpoint. This was surprising to me and I would like to bring clarity to what this means in practice for the benefit of the staff as well as the community. I would like to outline a number of operational principles to help us navigate towards this goal. These are not something we are inventing: these are widely used high-scale product practices; we will need to modify and adopt for our use. I look forward to discussing them with you all here.
And Erik Möller hints at how they see the volunteers as lesser partners of the WMF, rather than the community they exist to serve:
Lila and I will post more thoughts on the larger issues within the coming days. We deeply regret the disruptive impact this discussion is having on Wikimedia’s mission and our work together. At the same time, working through these questions has long been overdue, and my hope is that we can come out of this with greater clarity regarding how we partner on issues that are often likely to be contentious, which includes user experience changes.
It’s hard to see how these three somewhat contradictory sentiments can help calm things down. The question is: will the volunteers stick around when they’re not given a franchise? Again, the history here is that Wikipedia was built by contributors who gave their time, energy, and expertise to help create their product. The WMF was formed (at least in part) to serve that community and support their efforts. The community had always until recently had a strong say in the policies, content, and appearance of the website. It should surprise no-one that they’re unhappy with the path the WMF is taking now. German Wikipedia administrator Henning Schlottmann, a “Wikipedian” since 2005, succinctly laid out the problem here in a widely endorsed comment, as translated by Andreas Kolbe (emphases added):
Frank [Schulenburg], too, is not looking at the real issues that the MV has ignited. The Foundation has a miserable cost / benefit ratio and for years now has spent millions on software development without producing anything that actually works. This is in large part due to the fact that decisions are made without consultation with the community. On the other hand, it has to do with the fact that people like Erik, Steven and Philippe were recruited from the community, but obviously have no experience in really getting a product “out the door” and completing a project successfully. It was a good idea to employ Rachel, so she can take care of the communication about software development. But unfortunately Erik has severely damaged her chances after less than two months.
My theory: The WMF isn’t up to the job. Nobody who works there really understands and has a handle on software projects. This is evidenced by a horrific track record over many years. That the MV is rolled out even though it doesn’t recognize many licence templates is a symptom. The underlying cause is that the MV is based on a framework that has not been validated. We see the same thing in what is really a very minor issue, the thumbnail display. The layout team wants to abolish the frame and replace it with more white space. That they have not thought of images that need a frame to really show the image (Japanese flag) is one thing. But the guys have deleted the “Zoom” icon in the thumb frame without replacement. Why? Because they have not thought about what function it might have. With image maps that icon is the only way to get to the image information and the licence info! None of them knew that. And none of them asked or tried to find out for themselves what the function of that icon was. It’s the same with the MV. It reads the licence templates according to a microformat. So far so good. But this micro format is not universally distributed. Therefore, it should either have been rolled out only when everything was converted to that micro format and the millions of files had been migrated, or the MV should have used a more flexible model for reading information. But no, the thing has now been in development for XX months and has already cost Y million dollars, so it had to be rolled out now.
But it’s not only software development. What’s it like with user recruitment? How many millions have been invested in this over the past five years? Probably a two-digit number. And how many new authors were gained by it? Correct: practically zero. Why? Like Micha above I don’t see the problem with the editor. Anyone who has the intrinsic motivation to contribute to the greatest free education project in human history will not be deterred by the editor. There are many other barriers that are more important (first and foremost the ability to prepare information appropriately, but I mention this only as an aside). So here, too, the Foundation mucks about, but has achieved exactly no demonstrable results.
Why is it all like that? Because the Foundation (and in a similar way but to a lesser extent also WMDE) has grown much too fast. The unlimited money supply from the fundraising campaigns shows the tremendous enthusiasm of our readers, but it has seduced people into hiring staff without first agreeing on goals and methods. This excessive staff and bureaucracy then very quickly became estranged from its base, the community, and is now fighting for self-preservation. As far as content is concerned, they have nothing to show, so they have to use force. Best, –h-stt? 16:17, 16 August 2014 (CEST)
One thing should be clear: if the WMF loses its volunteers, or even starts to lose them at a faster rate than they already have been, their product will suffer greatly. If the sentiments being expressed by the German community are a sign of things to come, the talented volunteers might begin to take their talents elsewhere.
Image credits: Wikimedia, Flickr/gtall1 ~ licensed under Creative Commons Attribution 2.0 Generic