English Wikipedia anti-SOPA blackout/Post-mortem

From Meta, a Wikimedia project coordination wiki
Jump to: navigation, search
Editing levels appear to be unaffected.

Things that went well[edit]

Technical Issues[edit]

  • Google advising us on how to prevent harm to rankings (comment from Jorm) Risker 03:21, 19 January 2012 (UTC)
    • Please clarify - does this mean the decision to replace the full blackout by a javascript-only blackout? Boud 16:09, 19 January 2012 (UTC)
      • My understanding is that the decision to do a JS/CSS blackout, as opposed to changing the HTML, was made prior to the communication with Google. -- Tim Starling 18:21, 19 January 2012 (UTC)
        • NOT making anything 503 was done on the advice of Google LeslieCarr 18:50, 19 January 2012 (UTC)
          • The javascript-only issue turned out not to be a biggie, but it would have been helpful for the discussion so that people could have made more of an informed decision about how to !vote. First Light 20:09, 19 January 2012 (UTC)


  • Media mentions
  • Requests for interviews responded to chapter execs and members, WMF board, WMF staff Risker
  • Social media channels (Twitter, Facebook, G+, Blog) gained more followers and helped spread message: stats

Community and Cross-Community[edit]

  • See separate page for activities by other WMF communities
  • WP Staff identified the need for a final focused conversation and implemented it (in a flawed, but effective version). Philippe (WMF) 04:16, 19 January 2012 (UTC)

WMF Handling[edit]

(WMF Handling to include things like the responsiveness of Legal team, etc., for example.) Gyoung 05:27, 19 January 2012 (UTC)

Things that did not go well[edit]

Technical Issues[edit]

  • Extensive single-threading through a few staff left a lot of volunteers flailing trying to help because they were unable to self-serve. Important issues had to be repeated many times in order to catch the right person's attention. (partially helped by the etherpad but it was underutilized for gathering clearly important trouble reports)
    • It sounds silly, but i wonder how many people are really using a full featured local mail user agent, with many folders and exact filtering. It makes processing a flood of mails much more possible. What i see today is a shift towards a mix of Chat and comparably underdeveloped web applicatons like Webmail or Facebook. Many young people don't bother no more with configuring local mail applications at al (and won't know what i'm talking about here.) I'm not saying this does not work, but say it's inefficient. If there were a specific 'task force prep camp' it could include education in efficient information processing, and usage introduction into several things, including etherpad and video conferencing.
  • Mediawiki needs to develop some kind of etherpad extension, or have one running on the site. While we might not use it for general editing, we badly need an in-house collaborative live text editor, with a panel at the side for live collaborative dialog about the live text, and the ability to have multiple docs grouped together being worked on by the same people.
  • Good :) However I didn't know, nor did most others. At best I heard such things were available on some unknown external site, not on our own systems. It would have been fantastically useful. Suggest in future when it becomes obvious users will be working on a major or time-pressured document collaboration where this would help, staff/seasoned users should consider suggesting this immediately or much more quickly, and raising awareness among those actively involved. FT2 (Talk | email) 20:32, 19 January 2012 (UTC)
  • Longer lead-up times with better initial planning. Pulling this off in the time frame available was monumental, but it could have gone better with only an extra week of time.--Jorm (WMF) 06:14, 19 January 2012 (UTC)
  • Simple english got hit with the blackout (why?)
    • The technical reason is bug 25591. We actually knew about this going into it[1] and I thought we had this covered.[2] I visited the Simple English Wikipedia during the blackout and didn't see the blackout overlay. How long were they affected by the blackout? Kaldari 08:37, 19 January 2012 (UTC)
      • Around an hour? It seemed from IRC that jamesofur fixed it. SJ talk | translate   16:56, 19 January 2012 (UTC)
  • The wordpress blog was overloaded due to the high traffic the blog post caused. 3 part fixes. #1 Moving the server to a beefier server. #2 Cutting down comment display from all to 20 per page. #3 Implementing a cache on the box. LeslieCarr 18:54, 19 January 2012 (UTC)
  • On Camino, the screen was very light gray with white text. ~~EBE123~~ talkContribs 17:18, 4 February 2012 (UTC)


Community and Cross-Community[edit]

  • Could probably have been much clearer about what was and what wasn't open for long-term discussion (esp. re: blackout screen designs). Things that were not intended to be "call for entries" were assumed as such.--Jorm (WMF) 06:13, 19 January 2012 (UTC)
  • Early discussions which started December 10th began on Jimmy's talk, which despite garnering lots of responses, could not be used to declare consensus of any kind. Starting a proper RFC earlier would have meant less WMF involvement would have been needed to generate an unequivocal consensus for a blackout. Steven Walling • talk 06:24, 19 January 2012 (UTC)
    • Having a pool of community facilitators who work on this sort of task regularly might have helped here -- it's a slightly different set of considerations than mediation or arbitration. Processes like a standard summary notice in the next week's Signpost (or the equivalent on the sister-project/language-project in question) would be useful. With pratice we could define a better cross-wiki announcement system for such things as well. SJ talk | translate   16:56, 19 January 2012 (UTC)
      • Yes, to more and earlier communication of the discussion. Some long-time contributors seem to have not heard about it. First Light 20:11, 19 January 2012 (UTC)
    • Agree re RfC vs. Jimmy's talk page, but it's worth recalling that there was in fact an earlier village pump discussion in mid November, about the proposal of a new user to black out the logo for "American Censorship Day" in protest of SOPA/PIPA. Regards, Tbayer (WMF) 20:36, 19 January 2012 (UTC)

WMF Handling[edit]

  • List of important messages needed, so people can get working on them, and drafts need lining up earlier so they are believed ready to go 24 hrs before, allows a last review "the next day"
  • In focusing on on-wiki consensus, we forgot to post to the English Wikipedia mailing list which some were unhappy about. -- Steven Walling • talk 06:22, 19 January 2012 (UTC)
  • en:Wikipedia_talk:SOPA_initiative#javascript-only_blackout – !voters weren't told about the use of JavaScript during the vote. --Michaeldsuarez 19:22, 19 January 2012 (UTC)
  • [3] – !voters weren't allowed to decide the length of the blackout. --Michaeldsuarez 19:22, 19 January 2012 (UTC)
  • I was told by one WMF admin that there was some technical reason why the blackout had to be 24 hours.[4] Discussion elsewhere leads me to believe that the actual reason was fairness to all timezones. I was told by another WMF a WC admin that there was a technical reason why we could not have use a 503 redirect[5][6][7][8][9][10][11][12][13](Last paragraph.) Discussion elsewhere[14] leads me to believe that the actual reason was a Google caching issue.
There are two possible ways to interpret these events. If the technical issues were real, I have two WMF admins making claims about the servers and then going silent when I questioned the claims. If the technical issues were bogus, I have two WMF admins using the old "we can't do that because of technical issues" excuse rather than explaining the actual reasons for a decision. (In this case "too busy, get back to your question later" would also have been a perfectly acceptable answer.) Either way, WMF needs to do a better job when engineers not connected with WMF ask technical questions, and WMF admins who make claims about the servers need to make it clear whether they are speaking from personal technical knowledge or are passing on their understanding based upon conversations with the server administrators. --Guy Macon 13:23, 24 January 2012 (UTC)
I'm going to have to point out a third possible way to interpret these events, which is the true way to do so: You weren't speaking to WMF employees. With the exception of your last posted edit (by Kaldari, who is an engineer and not a sysadmin), none of the people you were speaking with were employees of the Foundation or are generally associated with managing the servers, code, or anything along those lines. Thus, your accusation about "WMF admins not responding" isn't really valid: there were no WMF sysadmins involved in your conversation.
For what it's worth, Kaldari is correct. Google does not index our site in the same way as the rest of the internet: they poll recent changes and then re-index pages along those lines. 503 errors would do nothing for them. --Jorm (WMF) 17:39, 24 January 2012 (UTC)

Actually, I was speaking to one WMF employee, but I wrote "WMF Admin", not "WMF Employee". I chose that wording because https://en.wikipedia.org/wiki/User:Philippe_%28WMF%29 says

"I am the Head of Reader Relations for the Wikimedia Foundation, and my work on this username is typically in my capacity as an employee... If it's an official action it will be clearly labeled... Unless otherwise stated, any edit to this wiki by me is an act of a regular member of the community and administrator, not a legal or official action of the Wikimedia Foundation."

The answer he gave me was not clearly labeled, so I referred to him as a "regular member of the community and administrator", or "WMF Admin" for short.

https://en.wikipedia.org/wiki/User:Gmaxwell and https://commons.wikimedia.org/wiki/User:Gmaxwell have at the bottom of the page, "Category: Wikimedia Commons administrators". I should have used the more accurate "One WMF Admin and one WC Admin" rather than incorrectly stating "WMF admins", and for that I apologize, but the fact remains that two admins who have nothing to do with managing the servers gave me what appears to be dubious information about the technical details of what can and cannot be done on those servers, and then went silent when I asked for details. (The other user who kept telling me I was wrong wasn't an admin of any kind, and thus I don't consider his comments to have anything to do with this post-mortem.) --Guy Macon 20:07, 24 January 2012 (UTC)

Things that could have been better[edit]

Technical Issues[edit]


  • Inadequate supply of prefab press-release grade material left reporters mostly riffing off a single somewhat unfortunate AP article. (We had old ones which were mostly still applicable, but they weren't easily found and were not updated)
  • Insufficient access to clearly spoken interview subjects left many articles without quotes from 'our side' (there are some exceptions, but where people were actually interviewed things went well)
  • Yes on both, for my part. The media needs and relations need doing in advance as some will want to comment beforehand too. This time it was very tight though and just getting it as we did was very good by all concerned. FT2 (Talk | email) 19:36, 24 January 2012 (UTC)

Community and Cross-Community[edit]

Decision to black out, and pre-blackout preparations[edit]

  • More effort to involve the larger En.wp community in the discussions at an earlier point. Risker 03:25, 19 January 2012 (UTC)
  • Banners pointing to RFC should have been global from the time the RFC started Risker 03:25, 19 January 2012 (UTC)
  • RFC Close should have made mention of the extensive discussion that occurred prior to the !vote (and I say that as the person who wrote much of the closing statement) Risker 03:25, 19 January 2012 (UTC)
  • OTRS - messages, queues, etc needed sooner, before announcements. We had them pretty fast, but emails will start arriving in advance of the day, or even before any announcement, so they are needed sooner than might be expected. FT2 (Talk | email) 18:05, 20 January 2012 (UTC)
  • OTRS again - email responses must often vary in tone as time passes, so planning is needed. For example before any announcement ("unconfirmed but you might want to..."), between announcement and day ("will take place on"), on the day ("is currently taking place"), and once the day ends ("took place on"). In particular we should aim not to send "blackout will only last 24 hours" emails once it's already over, and when there's an hour to go it might be more appropriate to say "It ends at 00:01 19th" rather than "It will last 24 hours". FT2 (Talk | email) 18:05, 20 January 2012 (UTC)
I don't know about this, FT2. Until the decision was actually posted, nobody knew exactly what the decision was going to be, and there were enough variables in the possibilities that investing a large amount of effort into writing messages that would not be used (instead of focusing on copy-editing and improving ones that *would* be used, once the decision was made) would probably not have been a good use of volunteer time. Perhaps translating the messages used by the Italian Wikipedia OTRS volunteers during their blackout might have been worthwhile, but that's about it. From my perspective, Volunteer Response Team actions were an example of "what went well". Risker 19:18, 20 January 2012 (UTC)
For the month it was being discussed, any few queries could be handled ad-hoc. For 3 days it was voted. It was predictable by 15- 16 Jan that some kind of emails would probably be needed (90% is hard to miss); the team could have begun thinking about drafts. In fact the first email saying "we probably need templates" was sent late on the 16th, on the 17th we were still discussing our tone regarding donations, and they were finally added to OTRS about 8 hours (I think?) before the blackout. Discussions related to incoming mail processing and tagging, what to tag for analysis, etc, took place shortly before the blackout. The discussions were brief when they happened. They could have easily been contemplated or drafts suggested by the 16th. Then, nobody (self included) thought ahead to note that language needed changing as events progressed. On Friday afternoon, 48 hours after the end of the blackout, an otrs volunteer finally and belatedly asked whether a template should be changed from present to past tense.
OTRS volunteers did a phenomenal and awesome job. But this section is "what might we do more or catch better in future events". Anticipating queues and templates beyond doubt are two of those. As with all drafts they can be tweaked up to the moment, but the thought to consider drafts earlier, and thoughts of the kinds of drafts probably needed, could be anticipated more in advance and firmed up upon events. FT2 (Talk | email) 02:11, 21 January 2012 (UTC)

During the black out[edit]

  • Super-user editing of wikipedia was generally good, see this table on my talk page for the five edits I believed were problematic due to happening during an edit lock. (All editors were highly collegial regarding this, and all edits were otherwise good edits). Fifelfoo 06:36, 19 January 2012 (UTC)
  • With respect to this, I think we would need some advance decisions and published guidance on how to deal with potentially injurious vandalism discovered during a blackout, particularly if content is still fully visible to mobile users. --Maggie Dennis (WMF) 11:58, 19 January 2012 (UTC)
  • I recollect seeing a number of BLP and copyright related edits made during the blackout, and didn't make any note of them as they appeared to be activities connected with Oversighters, Stewards or Staff's responsibility due to their role. Responding to edits with legal consequences seemed sensible, even when otherwise locked against editing. Fifelfoo 12:24, 19 January 2012 (UTC)
  • What constitutes legally actionable vandalism is obviously open to interpretation. --Maggie Dennis (WMF) 13:07, 19 January 2012 (UTC)
  • Maybe consider removing all the permissions, not just the editing ones, with the possible exception of suppression from Oversighters. I know that some editors, simply for the sake of exploring and with no ill will at all, changed permissions around; while they had no ill effects, there was a bit of hard feelings about it. Risker 06:49, 19 January 2012 (UTC)
  • There seems to be some low level generic ill-will regarding edits or changes conducted during the black-out. I didn't address permissions fiddling or css templates in user-space as these didn't relate to the consensus as closed. Fifelfoo 12:24, 19 January 2012 (UTC)
  • I strongly agree with Risker here, and made the suggestion before the blackout that certain user groups (especially the oversight group) should be whitelisted. It complicates matters when editing is left open to a user group which is generally unconnected to the project. Most stewards didn't mind helping out (although some understandably didn't want anything to do with it), but it would have been more appropriate for a local user group to have access for these emergency purposes. It turned out not to be much of an issue but for the future, it would be great if the stewards were not the only volunteer users allowed to edit. Thanks, PeterSymonds (talk) 13:07, 19 January 2012 (UTC)

WMF Handling[edit]

  • Better zip code data for the CongressLookup extension. Sadly, most of the publicly available data is spotty or outdated, but if we had had a few extra days we could have spent more time vetting the data and given people more accurate information on their representatives. Kaldari 08:44, 19 January 2012 (UTC)
    A request to pass upstream? Here, perhaps the Sunlight Foundation. SJ talk | translate   16:56, 19 January 2012 (UTC)
    I'm in contact with their data guy already :) Kaldari 00:55, 20 January 2012 (UTC)
    ZIP code validation. ZIP codes like 1_0 and 999999999999 should result in a "invalid ZIP code" error, instead of listing some random (and almost certainly incorrect) congressman. --Guy Macon 16:35, 21 January 2012 (UTC)
  • Conferencing can be enhanced to great effect. At times emails and irc were being shared by multiple people in a location (eg the projection display in the "war room"). Other times they were only showing on individual laptops or not at all. The result was that it wasn't clear to whom one was talking - whether one was addressing the entirety of a room or whether one was languishing in one person's email inbox. It wasn't clear whether to ask or "ping" again, or ask someone else, or if they were in the same room in which case "no benefit".
There may be times a user feels an urgent concern and wants to conference chat with multiple WMF staff, or to know to whom they are talking. The fact that dialog was sometimes conferenced at the far end and sometimes not was unknown to many participants. Can WMF give thought to workflow and "heads-up" in this area, 1/ to assure users where their dialog is at, 2/ to make most efficient use of resources, and 3/ to ensure users who do seem to need group dialog or collaborative consideration of a point can obtain it eventually in an ordered manner (ie queue/pingback) if appropriate.
My own experience was that the one time I was told "You are on the projector" was incredibly helpful and reassuring - it said "Don't waste time but you have everyone's ear briefly".
  1. During many irc chats (eg office hours) there is often a side channel where one person receives, clarifies, routes, and prioritizes all incoming requests. During a major events, can a specific irc/email be allocated as standard in advance, with a dedicated person to staff them, so users can engage appropriately and without issues, and everyone's time be better used.
  2. If conferencing projectors are in use, can the channel topic be set and a bot be used to notify the channel "you are live on projector" or otherwise?
  3. Option for "read receipts" on email?
FT2 (Talk | email) 01:53, 21 January 2012 (UTC)
  • Voice is a lot denser than text for some issues. Skype can take incoming skype or phone calls local to many countries, the "pro" version allows good quality voice conferencing, and many users use it as "de facto" voice software. Can WMF consider adding skype or a voice conference system, so if needed and subject to agreement, collaborating community members can have that facility to collaborate directly in small targeted groups or in dialog with WMF? FT2 (Talk | email) 01:59, 21 January 2012 (UTC)

Responding to concerns[edit]

Not sure where to put this, but one thought I had was that it might be an idea to not only improve how to handle the nuts and bolts of a situation like this in the future, but on how to handle the concerns raised. There were a number of strongly argued stances against the blackout. One is en:User:Scott MacDonald, the other is at User:Martijn Hoekstra/Open letter. There are probably others as well. While it is clearly a minority opinion, it is one that I'm very sympathetic towards (I personally had and still have deep misgivings about the Rubicon that was crossed here). So what I'm asking is whether there is any intention to collate links to internal concerns/protests and to respond to such concerns, or whether the intention is solely to fine-tune things for 'future' protests? My suspicion is that if events like this are repeated or become common, then support will fall drastically. Tactics like this only work if used sparingly. Carcharoth 03:09, 21 January 2012 (UTC)

1/ It's pretty clear from discussions that "extremely sparingly" and ideally not at all is expected. 2/ Agree, part of the community's post mortem should be to look at these and similar, and while noting they are a small minority, try to draw what lessons we can for how to help, what can be done, and anything else. If relevant for WMF etc, bring the key points here too. Good point. FT2 (Talk | email) 03:28, 21 January 2012 (UTC)
I think part of the "small minority" was because the opposition was never given comparable "exposure" to make their case. At every step of the way, the Wikimedia Foundation proponents basically pushed the message that Wikipedia was in (my phrase) mortal peril, and what does "the community" want to do about it? Framed in that manner, it's not hard to see the panicked reaction it generated, especially given the immense media scaremongering on the topic in general. By the way, I cynically think that we're going to see this more and more, as the incentives for the politicization here remain. I keep saying the following, and I think it's going over people's heads, but this was classic "manufacture of consent" of the masses by the policy-makers. And that's not going away. -- Seth Finkelstein 12:09, 21 January 2012 (UTC)
I saw complaints it wasn't advertised in (say) the wikien-l mailing list. Frankly though I'm not attributing blame games on it; anyone on any viewpoint could have and enough proponents and objectors saw it and none on either side did. The discussion didn't seem "panicked" or "scaremongered". It looked pretty thoughtful. Learn for future though, again. FT2 (Talk | email) 14:20, 21 January 2012 (UTC)
To me, it's not about the mailing-list notices. To put it in political-science terms, it's the overall issue of how the matter was framed for public consumption by policy-makers. Where's the archived central notice? There were of course thoughtful comments in the huge mass of material that Wikipedia generates for anything. But I saw too many people who thought that if there was a single link to an infringing site anywhere on Wikipedia, then all of Wikipedia could be taken-down. Misinformation like that is the panic and scaremongering. -- Seth Finkelstein 14:39, 21 January 2012 (UTC)
That is the plain reading of the law. Are you saying that no judge would ever agree that linking to the pirate bay's website or providing information or links related to copyright-breaking software that has gained widespread attention (which I hold should be covered, you may disagree), was "facilitating" someone? If a law exists, assume it will be used. If a law exists and money or an example depends on it, assume it will be pushed to the limit.
What I saw confirmed that a risk existed. I am not desirous of providing a law to test whether the doomsayers or naysayers are right, any more than I would give Iran the H-bomb as a test to see if it would genuinely use it or not. The risk isn't worth it and precedent says envelopes get pushed. if such matters were not intended to happen, it's their job to draft a law under which it cannot happen, not to expect others to rely on assurances how it is expected to be used.
I think this area's been discussed (though perhaps not resolved, do you have specific proposals or just a complaint?). I'm content to leave the final word, but in summary, enough legal views said "there's a serious concern" and so does the plain reading. A large site could be an "example to all" to change the net at a stroke, or smaller cases could be used to build up precedent. We may have to differ. FT2 (Talk | email) 23:32, 21 January 2012 (UTC)

(outdent) My specific proposal would be for there to be some sort of PRO/CON statements that people have to read before voting. But that's kind of self-contradictory with the whole contention that this was a "war of choice" by the WMF - i.e. obviously, then they're not going to allow strong opposition to put forth refutations, that's sort of the point of controlling the media debate. We have indeed gone around the specifics - it's way too late now, but I really should have FAQ'ed the points. Take this as a sketch of part of an against-the-motion statement (obviously, this is colloquial, not what would be in a formal statement).

Q: But Wikipedia was in mortal peril! A: Even the Wikimedia General Counsel conceded regarding site take-down, that "The new version now exempts U.S. sites like ours." (see also e.g. Techdirt - Pirate Bay immune). Why do you believe the mortal peril misinformation?

Q: But laws can be misused! A: Haven't you just created a grotesque Wikipedian version of the terrorist-scaremongering "One Percent Doctrine" where "If there's a 1% chance that (a proposed law can be used against Wikipedia), we have to treat it as a certainty in terms of our response.". Obviously, that way lies madness. So what would determine when there's a protest? Practically, when those in power start beating the war-drums.

-- Seth Finkelstein 11:46, 22 January 2012 (UTC)

I think you actually just undermined your own argument (not the first time). Any user of tens of thousands is free to set up a page or proposal like that and test public opinion on it, in a page where WMF intervention or staff comments would be clear and not covert, and at best a tiny number could be persuaded to any given view before someone would say something (crossref how very quickly canvassing even 6 - 8 people in community elections comes out). Nobody has done so; if someone wishes to then the outcome will speak for itself; either way your contention is testable by omission or outcome. For that reason a simple assertion it's so isn't very weighty and to me knowing how outspoken people here are when they disagree, the fact seems suggestive. (I'll collapse this subthread or at least some replies to it soon, or move much of it to the talk page, it's too off-topic).
As for my own stance, see above, something this important, if someone wants to convince me a law is right, I expect them to word it appropriately and to reduce so far as practical major areas of collateral damage - because once passed it is to be expected it will be used. If you ever follow my work on wiki you'll find that noticing flaws or careless thinking that could bite badly later is something I've done here for years too, and I'm no less forgiving when I find them on WMF or the community's part. I set high standards when competent people propose far-reaching actions. I didn't see that at all in SOPA. FT2 (Talk | email) 15:34, 22 January 2012 (UTC)
I think Seth has a point here. My feeling after seeing the rapid collapse of the bill is that we've had something of a Castle Bravo moment, finding our political impact to be much stronger than we imagined. Unfortunately, I think we're going to face fallout from that, because so many different groups are going to want to own that power. While the action was appropriate considering user sentiment and the importance to Wikipedia, in hindsight I think it might be best to go to smaller actions sooner - simple site banners without blackouts or professional PR - so we can see how much effect those have on their own. And the occasions on which we have to do this need to be kept few and far between, and closely tied to Wikipedia's role in building free culture. Wnt 21:41, 23 January 2012 (UTC)
A different and valid point. Not much to add. FT2 (Talk | email) 19:37, 24 January 2012 (UTC)