The Signpost

News and notes

Echoes of the past haunt new conflict over tech initiative

Contribute  —  
Share this
By The ed17
The Media Viewer, seen in November 2013

As with the troubled release of the Wikimedia Foundation's (WMF) flagship VisualEditor project, the release of the new Media Viewer has also been met with opposition from the English Wikipedia community.

A request for comment (RfC) asking whether the new project should be enabled by default for logged-in or anonymous users was closed on 9 July with "clear consensus" that it "should be disabled" for both (the closer added "by default" two days later). A previous RfC on the VisualEditor was closed with similar results. When the proposal's originator asked the WMF if they were going to disable it, WMF staffers quickly shot down the idea. Brion Vibber, the WMF's first employee and currently its lead software architect, wrote that "Perhaps it's time to stop calling self-selected surveys of a tiny subset of our user base 'community consensus'. The vast majority of our user base never logs in, never edits, and never even hears about these RfC pages. Those are the people we're making an encyclopedia for." Dan Garry, the WMF's Associate Product Manager for Platform and Mobile Apps, stated that "This is exactly why there is an opt-out for the feature. We don't expect everyone to like everything we make. That's a reality. So take 10 seconds to go to your preferences and disable it, and you'll never see it again."

Erik Möller, the WMF's deputy director, simply replied "no", and linked to a "detailed explanation" from Fabrice Florin, a product manager: "After carefully reviewing this proposal, we recommend that Media Viewer remain enabled on the English Wikipedia, for a number of reasons ... Overall, we believe that Media Viewer’s benefits far outweigh its downsides. And while the feature still has some limitations, we have collectively identified practical ways to improve it over time." Möller later continued along that line of thought, writing "It's normal and expected that the first reaction to noticeable user experience changes will often be negative. This is why we shouldn't base decision-making solely on early-stage RFCs and first reactions. Just look at the responses to major redesigns by Flickr, [the New York Times], and others—almost universally negative, irrespective of what the data actually says about user and readership growth or decline as a consequence of these changes."

Florin noted that there were several problems with the RfC's implementation, starting with the level of participation (one of their "key concerns"): the most opposes garnered was 64, set against an active editing community of over 124,000. Coupled with the five supports and one neutral, that is a sample of 0.06% of all active editors, and as pointed out in various places by Risker, the opposes could also be set against the 14,681 people who had enabled the beta version of Media Viewer.

Nor did it account for anonymous readers, nearly all of whom do not participate in such discussions. Community members pushed back against the latter point; Isarra Yos wrote that "they know their audience, they interact directly with this audience ..., and indeed they often use the site exactly as this audience would, simply taking things a step further to edit as well." Todd Allen echoed Yos: "I am beyond tired of hearing that those who have volunteered hundreds or thousands of hours per person toward building the greatest educational work in history do not have at heart the interests of those who would use it."

The plot thickened after MZMcBride posted JavaScript code that, if used, would leave the enabling or disabling of MediaViewer to English Wikipedia administrators. Yet when this was applied less than an hour later by Pete Forsyth, he was reverted by Möller, who threatened that the WMF would "temporarily revoke [Forsyth's] admin privileges" if he edited the site's JavaScript again. This edit was made under Möller's Eloquence account; unlike most WMF staff, he uses one account for his personal and professional edits.

Erik Möller, the deputy director of the WMF, reverted an English Wikipedia administrator's modifications to the site's JavaScript that would have disabled the Media Viewer

That the WMF would have the legal authority to do this is clear, but it is less clear if they were planning to justify it under any English Wikipedia policy. Their exclusive software mandate comes from Wikipedia:Consensus, under the section header "Decisions not subject to consensus of editors". It lays out the differences between various communities, such as MediaWiki software developers, and continues "These independent, co-equal communities operate however they deem necessary or appropriate, such as adding, removing, or changing software features ..., or accepting or rejecting images, even if their actions are not endorsed by editors here."

But for removing an administrator's toolkit, the closest applicable policy is Wikipedia:Office actions, a rarely invoked process that lays out how the WMF can remove "questionable or illegal Wikimedia content" when after a "formal complaint [is] made off-wiki." Only four articles are currently affected by it. The page does not say anything about blocking editors or removing userrights, except in two cases. When the WMF takes an official office action, administrators making "unauthorized modifications" will see their actions reversed and, "possibly", their administrator tools removed. There is also an explicit caveat for when an editor is repeatedly adding content subject to (a) DMCA takedown notice(s).

Where do the WMF and community go from here? There is an ongoing request for arbitration that as of publishing time has a majority in support of accepting the case, but the arbitration committee's own jurisdiction prohibits them from reviewing "official actions of the Wikimedia Foundation or its staff". A section has been opened on the RfC's discussion page proposing that the community reassess the tool in six months. "In the meantime," states Tom Morris, "the Foundation can go back [and] seriously rethink how they engage with the community—threatening desysops is not the way to do it. If the issue is the Foundation believes that the average reader would be more keen on MediaViewer than involved editors, then the Foundation can do user testing and surveys to show that."

In brief

+ 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.

Disambiguation link

I've made this edit to correct a disambiguation link in the article. Zhaofeng Li [talk... contribs...] 06:27, 13 July 2014 (UTC)[reply]

Correction to quote from closure of RFC

The original closing statement was "There is a clear consensus that the Media Viewer should be disabled for both logged-in and logged-out users." This was not reflective of the actual discussion, which was about disabling default, not disabling Media Viewer entirely; I note that Armbrust has subsequently revised his closing comment, but in the interim, this is at least in part related to the next steps, which resulted in the application of a script that did exactly what Armbrust's initial close said it should do. Risker (talk) 07:30, 13 July 2014 (UTC)[reply]

Thanks, Risker. I've added text to clarify this. Ed [talk] [majestic titan] 14:23, 13 July 2014 (UTC)[reply]


  • I don't argue for or against this bit of software, but the attitude attributed to WMF is reminds me of words I've heard before. Vibber said "Perhaps it's time to stop calling self-selected surveys of a tiny subset of our user base 'community consensus'. The vast majority of our user base never logs in, never edits, and never even hears about these RfC pages. Those are the people we're making an encyclopedia for." Richard Nixon's Vice President Spiro T. Agnew in 1969 said "America's silent majority is bewildered by irrational protest." Richard Nixon said those who vocally protested his policies were vastly outnumbered by the silent majority who supported him. Perhaps surveys of the Vast Silent Majority should be taken before someone asserts what they think. Edison (talk) 13:38, 13 July 2014 (UTC)[reply]
  • I am quite offended, as I'm sure many others are, at the capricious attitude displayed by WMF staff. If we, the editors, are to be disallowed from decision-making the least WMF could do is re-write "Office Actions" to demarcate the limits of their authority. Threatening a punitive block to an admin acting on consensus is beyond the pale. Chris Troutman (talk) 15:57, 13 July 2014 (UTC)[reply]
  • If the decision making capacity of RfCs isn't questioned for other matters, why is it questioned for this one? Are all RfCs now suspect? It seems to me that the people most interested in what happens to the English Wikipedia would take part in these discussions, so even if that turns out to be a small number, its decisions should be given a lot of weight. If there's a problem, it's with the English Wikipedia's diminished overall participation and possibly a lack of way of getting Wikipedians to take part in the most site-critical discussions. What happened since the heady days when we got virtually everyone to comment about SOPA? Stevie is the man! TalkWork 19:54, 13 July 2014 (UTC)[reply]
    • Stevietheman, please see above. The code that was added to the mediawiki.common.js didn't just disable the software as the default, it disabled it entirely. That wasn't even the intention of the person who added the code, and it wasn't the correct interpretation of the RFC - the initial close of the RFC didn't match the actual discussion that occurred. It's entirely right that the script was removed; it wasn't even the consensus on the RFC. Risker (talk) 20:03, 13 July 2014 (UTC)[reply]
      • Thanks for the technical points, but I'm referring to the WMF's attitude about RfCs, which stands no matter what the particulars are. Stevie is the man! TalkWork 20:06, 13 July 2014 (UTC)[reply]
        • If a project develops an internal consensus that its users must have accounts separate from their SUL account, the WMF isn't going to agree to implement it - and they shouldn't. Thisi s exactly what the WP:CONEXCEPT purpose is. Core features remain enabled in some form on all projects. In this case, the user inserting the code disabled a core feature - and not even with community consensus at that. To be honest, I don't think the WMF's attitude to RFCs is all that different from the community's. Risker (talk) 20:13, 13 July 2014 (UTC)[reply]
          • I wouldn't call the Media Viewer a "core" feature. We can edit and keep working with or without it, and the fact that the Foundation is trying to shove it down to us is concerning. I understand that the code inserted by Pete was not the best solution, but the response from Vibber and Moeller left us with little to no room to explore alternate paths. → Call me Hahc21 23:13, 13 July 2014 (UTC)[reply]
            • Core features are those that are attached to MediaWiki, as opposed to local features or gadgets or scripts. Risker (talk) 00:56, 16 July 2014 (UTC)[reply]
              • No, not really. As noted at the top of mw:core features, the term "core" is used in a variety of contexts, but I don't know of any in which a MediaWiki extension (i.e., mw:Extension:MediaViewer), which adds an entirely supplementary feature, is considered a core feature. Extensions are inherently not part of MediaWiki core, of course. There's also currently a wmf:Template:Staff and contractors#Core Features team, but they were not involved in MediaViewer. I don't see how it's appropriate for you or Edokter to suggest that this feature is somehow vital without first clearly establishing why you believe this to be the case. --MZMcBride (talk) 03:15, 16 July 2014 (UTC)[reply]
                • MZMcBride, MediaViewer may not be in the core of MediaWiki, but it part of the Wikipedia installation, so it is regarded as a core feature of the website. Consider Extension:Cite (which was moved from core some time ago). Do you not consider that to be a core feature? -- [[User:Edokter]] {{talk}} 09:35, 16 July 2014 (UTC)[reply]
                  • Cite is vastly more important than MediaViewer as it's actively used by articles and editors, but neither feature represents core functionality. Both extensions are not part of MediaWiki core and it's trivial to do without either of them. MediaViewer is not considered a core feature of Wikipedia the Web site. It's insane to suggest that MediaViewer is. MediaViewer is a bit of JavaScript that hijacks clicks on files. Editing is a core feature of Wikipedia. Logging in is a core feature of Wikipedia. Serving content to readers is a core feature of Wikipedia. File upload support is a core feature of Wikipedia. MediaViewer is not a core feature. --MZMcBride (talk) 13:11, 16 July 2014 (UTC)[reply]
  • I find Brion Vibber's comment highly offending to the Wikipedia community as a whole, as well as clearly disingenuous. → Call me Hahc21 19:57, 13 July 2014 (UTC)[reply]
  • The leading vote recipient in the Wikimedia Foundation 2013 elections got 787 votes, from our "active editing community of over 124,000". Any decisions based on the authority given to persons appointed by this board which was elected by a "tiny subset of our user base" would seem to be based on a pile of sand. I didn't bother to vote in this RfC, per WP:SNOW, because I didn't want to unduly pile on to what was looking to be a lopsided result. At best, this was looking like an uphill battle. I've watched the videos of our new executive director's first two monthly talks to employees, which are posted online. In each, she emphasized that the employees work for "the users". It's time to specify how "what the users want" is determined. If it's not a WP:RFC, then the whole set consensus-based guidelines on which this project operates needs to be reconsidered. Maybe we need to hold a constitutional convention. Wbm1058 (talk) 20:33, 13 July 2014 (UTC)[reply]
There will always be some questions on who represents the community better, the WMF Board or 60 people on an RfC. Personally, I think the self-selected RfC participants would have to show why they consider themselves the representative of the whole movement. The WMF was started by the founder, does hold some election, and has gone out of its way to get feedback.
There really shouldn't be much doubt about where the community stands on many of the controversial issues that seem to be debated forever on pages like this on on RfCs. Just take a random sample of about 400 editors, say 4 times a year, ask them about, say, 6 current issues. Delve deeper on the next survey if opinion on an issue is closely divided. You'll get a series of "where the community stands" over time and on many issues, with a margin of error of about 5%. What more info? Do same thing with readers (non-logged in accounts). Cost? Maybe $100k/yr if the WMF wants to farm it out, maybe cheaper if done in-house. This could solve those seemingly unsolvable questions - do editors want better media capabilities? Do editors want paid editors regulated? (The ultimate RfC puts this at about 80% yes - why did it take us this long to figure this out?) Should the system of admins (or Arbcom, or ...) be overhauled? It would be pretty easy to actually find out rather than argue about it forever. Smallbones(smalltalk) 00:49, 15 July 2014 (UTC)[reply]
  • I actually prefer MediaViewer to the previous mechanism for viewing image thumbnails. I am, however, very unhappy with the bullying engaged in by Erik Moeller and hope that ArbCom weighs in on this matter with action and a statement of community expectations of appropriate behavior by WMF employees. Carrite (talk) 15:14, 15 July 2014 (UTC) Last edit: Carrite (talk) 15:16, 15 July 2014 (UTC)[reply]
  • Out of curiosity, has anyone pointed out to Brion Vibber & others who agree with him that their argument that RfCs don't represent the majority of users is little more than a paraphrase of the old Usenet claim, "The lurkers support us in email?" -- llywrch (talk) 16:11, 15 July 2014 (UTC)[reply]
That's an intereseting point. Which users are you talking about, though? Because the majority of people who click on images do not edit, they are readers. We can't actually say what they think, except for some results from a WMF-based survey whose legitimacy was questioned. It would be interesting to set two laptops side by side on the same page, get someone who only reads WP and never edits to click on the same image on the same page, and ask them which of the "standard" view or Media Viewer they prefer when looking at images. I have no idea what the result would be. But disabling MV as default for "unregistered" users (i.e., all the millions of readers) would remove the option entirely for them. A reader survey that allowed readers to compare the two "side by side" so to speak would be much preferable to everyone insisting they know what would be best. Risker (talk) 00:50, 16 July 2014 (UTC)[reply]
@Risker:, the preference of readers is certainly an important consideration, but it doesn't trump other legitimate considerations. We have the legal rights, and the fair treatment, of copyright holders and the subjects of photos to consider; and we have the strategic priorities of the Wikimedia movement to consider. The image to the right comes from the strategic planning process; the red arrow on the right-hand side indicates the important link between readers and contributors. Very few readers become contributors, but it is of vital importance to the movement that some do. When a software change makes the very concept that it's possible to improve the page virtually invisible to readers, that damages that red arrow. How much damage? We don't know, because the WMF didn't test that. -Pete (talk) 01:29, 16 July 2014 (UTC)[reply]
Umm, no, I don't buy that, Peteforsyth. I'll lay odds most readers don't even realise that they can click on an image and wind up somewhere else (and that's pretty shocking all by itself); heck, I didn't figure it out myself for months after I started editing. You remember, back in the day, when we used to have those little image placeholders and encouraged people to upload their images? Do you also remember that 90% of the images uploaded in response were deleted because they were copyvios, and that the "new editors" thought that since they'd personally scanned the image or done the screengrab, they owned the rights? We don't have that problem anymore - we got rid of the placeholders, people stopped contributing copyvios, and the net effect was...well, we weren't gaining new accounts as quickly but we also weren't overloading admins or having people leave in a huff because their one contribution had been declined. And we had fewer copyvios. A lot of editors never upload images. IP editors can't upload images - it requires an account. (This is probably a net positive, because we've already got enough, um, self-portraits to fill an encyclopedia. Okay, bad example.) The "Edit" tab is the entry point for almost everyone, not images, at least not from Wikipedia. Ask the question, you'll get an answer; if you want, I'll ask the question: what percentage of new editors to English Wikipdia made their first edit a media upload? Data from May 2013 and May 2014 for comparison, both dates from before MediaViewer was accessible to new users. How does this sound? While we're at it: percentage of active editors from both months who have never uploaded an image, just to answer that niggling question. Thoughts? Risker (talk) 02:32, 16 July 2014 (UTC)[reply]
@Risker:, I'm touched that you remember the placeholder thing -- starting the centralized discussion to get them removed was my first substantial effort to push for a broad change on Wikipedia. Your points are all good ones, and it seems like you are saying basically the same thing as me: there are significant questions around the reader-to-contributor transition that should be carefully and transparently considered. -Pete (talk) 15:41, 16 July 2014 (UTC)[reply]
And now we have mobile frontend doing something very similar to the placeholder image thing, but pushing everyone to commons. Life's a circle I suppose. Bawolff (talk) 02:16, 17 July 2014 (UTC)[reply]

Annual plan

I found the risks section of the annual plan to be quite interesting. In particular, "A competitor could provide an interface (reading, editing, or both) that is significantly better than ours" [1]. Well I could certainly understand how that could be a potential risk to the foundation, I do wonder if such an event could actually be a good thing for the Wikimedia movement as a whole (competition breeds innovation and all). Bawolff (talk) 22:29, 14 July 2014 (UTC)[reply]

Certainly competition could be good for the foundation and for the movement as well. One possibility could be for the WMF to fund some type of internal competition, e.g. a "Wikipedia" for academics, or perhaps a version, call it "Edgepedia" that is willing to experiment with new media applications, moving closer to the cutting edge of technology (or maybe call it Mediapedia!) Smallbones(smalltalk) 00:49, 15 July 2014 (UTC)[reply]
Wikimedia's mediapedia for a more media-immersive experiance by using a non-Mediawiki interface! Bawolff (talk) 01:30, 15 July 2014 (UTC)[reply]


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