Talk:Image filter referendum/Results/en/Archive

From Meta, a Wikimedia project coordination wiki

Releasing results[edit]

Interim message[edit]

(The message below was posted on the Image filter referendum/Results/en page on September 2, to advise the community of a delay in releasing results. It is placed here for future reference now that the preliminary report has been posted. Risker 01:59, 4 September 2011 (UTC))[reply]

Due to the very high vote count, and the need to review comments manually, the results of this voting will be delayed. The committee regrets this, but wants to be certain that the results are error-free and the comments are accurately portrayed.

We expect to release totals on 3 September, 2011.

There were more than 24,000 votes and almost 7,000 comments submitted. While the committee will not review each comment before publication of the results, they will all be reviewed before the committee disbands and a representative sample (nearly 2,000) will have been reviewed by publication of the results.

We thank the community for its patience, and for the outpouring of opinion.

For the Image filter referendum Committee,

Risker 04:05, 2 September 2011 (UTC)[reply]


When are results coming out?[edit]

It's September 1. And? — The preceding unsigned comment was added by Patio (talk) 2:23, 1 September 2011 (UTC)

... Rich Farmbrough 16:13 1 September 2011 (GMT).

Are the answers that bad that the foundation is shaking?--91.10.49.253 21:22, 1 September 2011 (UTC)[reply]

No, but when the timeline was developed nobody expected there would be over 24,000 votes and almost 7,000 comments that would have to be manually reviewed; the last Board election had fewer votes than there were comments in this one, and this plebiscite has had a larger response than any other in Wikimedia history. We're working as quickly as we can, but it's very important to be able to provide a genuinely representative sample of the comments. Risker 04:10, 2 September 2011 (UTC)[reply]
I assure you, the "Foundation" isn't shaking. :) I am the only Foundation employee with access to the full results so far, and I'm not shaking. I'm a little overwhelmed by the sheer volume, that's all. But I've read and collated nearly 1000 comments in the last 24 hours so I suppose that's natural. Philippe (WMF) 05:25, 2 September 2011 (UTC)[reply]
This is most disconcerting: if only one person has access, there are no checks and balances. Combine this with the biased setup of this referendum, the canvassing spam mails, and the hiding of negative comments.... I'd like to Assume Goof Faith, but it is made exeedingly difficult. Zanaq 11:16, 3 September 2011 (UTC)[reply]
He said the only WMF employee, not the only person :) The full committee is here. Results very soon, we hope! -- phoebe | talk 01:33, 4 September 2011 (UTC)[reply]
We really want to see the raw data and also the PGP to verify that all votes are counted. Otherwise you could tell us a fairytale, insert "random" votes, and so on. I usually assume good faith. But in this case I'm very concerned, since so many facts make the whole thing very biased.
  1. A report from a single person is the source (No other organization e.g. the ALA was involved, Sue worked for the same company, all coincidental?)
  2. The questioning itself is very biased (No "no" option. "I don't care" is no "No!")
  3. Spam mails that where filtered before sending them (For example: No user with "bot" in his name was informed)
  4. The whole proposal is obviously ignoring the five pillars.
  5. ...
How can you still expect us to trust you? If i assume good faith, than i must admit that you made one stupid mistake after the other. If i don't assume good faith, then most parts are fitting together. --Niabot 13:57, 3 September 2011 (UTC)[reply]
We sent out 700,000 emails and put a banner notice on every project page for a couple of weeks.... How many responses were expected? Rich Farmbrough 22:01 3 September 2011 (GMT).
Good question, Rich. I think it's just too rare for us to have these sorts of actively-promoted community discussions, so we have to think back to the relicensing proposal to find a similar one. We should be doing this sort of outreach and mutual communication all the time - say at the 1% level of banner-promotion if not the 100% level. Then we would have a finger on the pulse of the community, what sorts of topics interest readers and editors, and would not be surprised. [we would also get many more good ideas or offers of help that way!]
To Niabot: there is no conspiracy afoot. (not even in the filtering of email recipients, though that is embarrassing!) As with all community votes, raw data will be made available. I'm not sure what you mean by "no other organization was involved" -- the vote was administered by Software in the Public Interest precisely to avoid such concerns. Many good things develop 'one mistake after another', and the history of the wiki and of Wikimedia is no exception. If we are afraid to make mistakes we can't learn. Please don't shame people into not making mistakes: help them learn from them. Let's make sure we have more (and more meaningful) community polls, including real referenda (open referendum tools are becoming popular). Let's make sure we have rules of thumb for releasing data. Someone has been unhappy about data-release being too slow after every public vote I can recall. Assume raw data includes dupes, socks, and identifying information. What subset of that can and should be availalbe immediately? How soon should the rest be available, and in what formats? SJ talk | translate   22:44, 3 September 2011 (UTC)[reply]

We want the immediate release of the numerical results. We are not patient anymore. If you have nothing to hide, you would have published this part in the first minutes of September 1st. --Eingangskontrolle 19:17, 2 September 2011 (UTC)[reply]

Release raw data first?[edit]

Is it so difficult to represent the raw data of the result, or will we only see the interpretation by the WMF? I'm curious what takes so long to represent a plain table with the plain results. --Niabot 06:10, 2 September 2011 (UTC)[reply]

With the comments, we must ensure that there is no publicly identifiable information present (and there is). The committee has chosen to release both pieces (comments and the vote info) at the same time, rather than release an incomplete picture. Philippe (WMF) 06:12, 2 September 2011 (UTC)[reply]
Then release the poll results first and the comments later on. Can't be so hard. --Niabot 13:59, 3 September 2011 (UTC)[reply]

Please release the incomplete picture now: only the raw and anonymous vote data. Release the interpretation separately, including (filtered) comments. That would be much more transparent. Zanaq 11:18, 2 September 2011 (UTC)[reply]

Why has no one responsible for organizing this 'referendum' not responded to this very sensible suggestion to explain why it hasn't been implemented? If you're busy reading comments, you can do that later. Just publish the damn results and we can all pick over their implications together! Trilobite 08:51, 3 September 2011 (UTC)[reply]
Didn't the mail go out? I thought there was notice that we wanted to delay a day to release graphs of the data as well? At any rate, it will be posted soon. -- phoebe | talk 01:36, 4 September 2011 (UTC)[reply]

This 'referendum', which was nothing of the sort given that the decision to introduce a filter has already been made, was a train wreck already. It's now descending into farce. The results of an electronically conducted ballot can be released immediately in raw form at the close of polls. There is no excuse for any delay in doing this. If you're concerned that people's comments will compromise their anonymity you can just filter comments out and publish every part of each submission except the comment for now, stating where applicable that a comment was left and will appear later when it has been checked for personally identifying information. In future, it would be strongly advisable to have a warning on the voting page stating that anything people write in the comments section will be made public, making it their responsibility to stay anonymous if they so desire. I cannot understand what the hold-up is with these results. Transparency in Wikimedia governance is already shot to pieces. Don't make the situation even worse! Trilobite 08:49, 3 September 2011 (UTC)[reply]

Agree with all the people who say you could publish the results as we're supposed to be a free (as in open) community. So many people took the time to vote and have the right to know what we all want A.S.A.P. Please, WMF. There is no reason to wait. You can apparently, but we have more things to do... S.I. 'Patio' Oliantigna 10:32, 3 September 2011 (UTC)[reply]
I have to agree that the delay is becoming suspicious, especially when the reason given is to read comments before providing a "random sample". The combination of having someone take time to read over what the comments say and decide which ones to release gives the impression that someone might be taking the time to put a spin on the results. Also, I think I saw someone make a comment about translating comments, which seems extraordinarily dubious, considering that people were still trying to get translations of the poll ready a few days before it closed. Translating comments from every country on Earth is a process I'd expect to take months, not weeks. As far as personal information is concerned, I simply don't care - I hope we had the proper disclaimers on Meta - you submit something, it gets published. If people really want to redact such information they can do so post facto, as I've often seen at the Wikipedia Refdesk. Wnt 01:03, 4 September 2011 (UTC)[reply]
WMF committed to not publicly releasing identifiable information about voters. Do you really expect them to say, "Oops, Wnt's impatient, and you didn't write your comment in English, so we couldn't tell what it said and we released it anyway. Sorry about that: our confidentiality guarantee only applies to English speakers."
Or did you think that when they said "random sample of comments" they meant "random sample of English-language comments"? WhatamIdoing 02:58, 4 September 2011 (UTC)[reply]
In the future, we should make it clear that submitting a comment in a poll is just like submitting an edit to Wikipedia - that it becomes public and immediately available (and make that happen). Wnt 08:10, 4 September 2011 (UTC)[reply]
Why? Please explain your reasoning in detail, because I really don't get this suggestion. Risker 16:04, 4 September 2011 (UTC)[reply]
I agree with Wnt. Comments (but not the votes that go with them) should be public and immediately available at the end of the vote, and remain anonymous unless the commenter explicitly identifies themselves. (Although there probably should be some smarts warning people of identifying information being included, and perhaps something confirming the identity if they do want to include some.) Mark Hurd 05:12, 5 September 2011 (UTC)[reply]
Is your goal to reduce the utility of the text field for communicating between the poll-giver and the poll-taker?
In a similar poll a while back, I once provided a text comment that I very much did not want to be made public (to protect the guilty). I thought it important to provide the information, however, because my "yes" answer, while definitely accurate, involved a situation so unusual that the poll-giver could not reasonably have expected it, and I was seriously concerned that, without such feedback, the poll-giver would interpret the responses to conform with the common stereotype, rather than the diversity of the real world.
If I had not been reasonably certain of the poll-giver refusing to release comments, and thus protecting the privacy of the other person, I would never have provided that comment (which I heard later was useful to them).
Also, there's nothing to guarantee the accuracy of such comments. Would you be happy if libelous comments were instantly released? Would you be happy if I adopted a policy of writing "Mark Hurd is a convicted criminal and should be banned from WMF projects" in free-form text fields for all future WMF polls? Should that sort of comment be publicly released without any review? WhatamIdoing 17:54, 5 September 2011 (UTC)[reply]

┌─────────────────────────────────┘
I do agree the comments need some review (which I sort of implied by saying "something confirming the identity [of] identifying information" (slight paraphrase of the above)). I have also added {{NOINDEX}} here and would expect a similar thing on the release of comments. I do concede actual "immediate availability" is wrong; I suspect I'd prefer each submission having the option of it becoming public, once reviewed. Mark Hurd 10:27, 7 September 2011 (UTC)[reply]

That's basically what we have now: submissions are released after review. WhatamIdoing 17:24, 7 September 2011 (UTC)[reply]

Congratulation[edit]

Regardless of whether the release is slow or is it not, I want to congratulate to the committee, the board and to the community in general, for such an big response. I remember the sorrow of the organizers for small participation in the board election. I took extra here some care to participate this time.

I hope, that the results will move the Wikipedia (and other projects) bit forward too :). Have a good time and thank you for your hard work. Reo On|+|+ 15:13, 2 September 2011 (UTC)[reply]

Thank you for sharing, Reo On. I hope to see proper referendums that originate with the community in the future as well. SJ talk | translate   22:46, 3 September 2011 (UTC)[reply]

Results now released[edit]

Please see the page for the preliminary report. Risker 02:01, 4 September 2011 (UTC)[reply]

Disappointing result. From the results it seems support for this ridiculous thing is pretty clear. I wonder what kind of influence the biased questions and spammail have had: we will never know, so the Foundation has been legitimized to push this waste of resources forward. Congratulations. I hope the :nl:wikipedia community will decide to disable this. Zanaq 06:55, 4 September 2011 (UTC)[reply]
Interesting - I actually read the result as 'no consensus', which on Wikipedia would typically mean it doesn't go forward. The key number being the "Average of those with a preference" for "It is important for the Wikimedia projects to offer this feature to readers.", which is 57% - i.e. it's slightly favoured, but not strongly favoured (if you turn it into a black-and-white yes-no, then I'd say it's the analogy of less than 2/3 saying yes). Mike Peel
Perhaps another key question ought to have been worded, "Should image filtering be added (yes or no)?" Voters' assessment of the importance of the feature is significant, of course, but so, too, is other voters' assessment of the importance of not doing it. Focusing only on the former, and then on the various characteristics the feature would need when added, suggests a bit of bias in favor of image filtering. But in any case I think you are correct, Mike, that the numerical results of even this apparently slanted poll suggest that there is no consensus that image filtering is a good idea. It will be interesting to see if the Foundation will implement the feature anyway. If so, perhaps other considerations are in play. Jonathan Lane Studeman 09:47, 4 September 2011 (UTC)[reply]
Yes, it is important to realise that the result actually is no concensus. No consensus usually means that things are kept as they are, i.e. there will be no filter. / Achird 10:38, 4 September 2011 (UTC)[reply]
This wasn't a yes/no vote, but a sort of market analysis. The result is that there is in fact, alas, some "market" for such a function. We can't make this fact go away, but we do have to draw limits as to what it means and what it doesn't mean. This vote, for example, is not a vote that "editors who add images to articles without rating them, or without rating them in a way that is popular, shall be blocked". It is not a vote that image hiding code needs to appear in the Wiki source of articles. This market can be reached in many ways, including by allowing offended users to circulate their own personalized lists of files they don't like, which are "tagged" in the articles only by having a filename. Wnt 15:45, 4 September 2011 (UTC)[reply]
Both true -- there is clearly no consensus, and there is significant interest in such preferences, both negative and positive. [If it weren't for the latter, we wouldn't be having this discussion at all -- it is much simpler and more comfortable to do ignore that class of suggestions, on a number of counts.] To your last point, there seems to be strong agreement that no 'tagging' system will work. SJ talk | translate   19:41, 5 September 2011 (UTC)[reply]
Disappointing for who? Given the premise was that this was a "how do we" survey, I'd say the result shows a very strong message that the community does not want to have this system. The survey materials were not encouraging people who did not want the system at all to make their feelings clearly known , or even to vote. And yet all that could be achieved was a luke-warm response to the "importance" problem. And this is thrown into high contrast by the definitive answer to the "culturally neutral" question - which is a key focus for the hard questions the proposal generates. Rich Farmbrough 17:35 4 September 2011 (GMT).
I agree neutrality is a prerequisite for any core features; I wa surprised that answer was not even more definitive. How did the survey materials not encourage those opposed to such a system to vote? SJ talk | translate  
Is neutrality important for surveys? IMO, this survey was not neutral. IMO, this survey is also not sound; it should have used a five-point w:Likert scale (or similar) rather than 1-10. It should have been developed onwiki. It should have been reviewed by RCom. As it is, I have no idea who devised or approved this survey.
People who oppose an image filter had no obvious way to clearly register this. It did not respect that there are people who believe Wikipedia should be completely uncensored, and the editorial control should be in the hands of the editors. I am a supporter of an image filter, if done right. However I would prefer it is not done rather than done badly, as it would undermine the encyclopedia and, if it is not supported by the community, it will have been a waste of WMF money. This survey is not giving me much hope. Remember Pending Revisions? If you dont have the community behind you, it wont work. I am shaking my head at the way the WMF is continuing the centralise and use top-down approaches, and the result is often that the community rejects the WMF solution.
There should have been a question like "Do you believe an image filter is appropriate for an encyclopedia?" and/or "Do you believe an image filter is appropriate for Wikipedia?" Why? because we *know* there is a segment of the community who believe that Wikipedia specifically should be uncensored, and it is only right that a survey measures that segment of the community, if only to estimate how many people will leave if this is implemented in a way that fundamentally undermines their vision of Wikimedia. A fundamental in our community is that we respect the viewpoints of people we disagree with. This survey does not respect a large segment of our community who put free-speech and the 'wiki' ahead of the readers. Their philosophy is not without merit - free-speech and the 'wiki' is what made the readers come in the first place. It would have been useful to include a question like "Have you been confronted with an image on Wikipedia that you wish had been hidden by default?" followed by "Which topic was this image on?"
This could have been a useful survey; it could have been a survey which helps the readers communicate their preferences to the editing community and the WMF, which could have brought about real change. As it is, I doubt that any meaning can be obtained from the results. John Vandenberg 08:49, 10 September 2011 (UTC)[reply]

Requests for further results or visualizations[edit]

Table formatting[edit]

Done. Philippe (WMF) 18:00, 4 September 2011 (UTC)[reply]
 ? 0 1 2 3 4 5 6 7 8 9 10 Votes Votes with a preference Average of those with a preference Median
Total 269 3763 790 1163 978 715 2819 1800 2670 2957 1308 4791 24023 23754 5.7 6
% 1.12% 15.66% 3.29% 4.84% 4.07% 2.98% 11.73% 7.49% 11.11% 12.31% 5.44% 19.94%
Total # %
? (Abstain) 269 1.12%
0 3763 15.66%
1 790 3.29%
2 1163 4.84%
3 978 4.07%
4 715 2.98%
5 2819 11.73%
6 1800 7.49%
7 2670 11.11%
8 2957 12.31%
9 1308 5.44%
10 4791 19.94%
All votes 24023
Votes with a preference 23754
Average (Mean) 5.7
Median 6

Would anyone mind if we moved to a different format for the table, like one of my examples above? I personally prefer the first, because it takes less work to make this change. The existing one is hard to scan for trends and runs off the end of the display on all but the widest computer screens. WhatamIdoing 03:16, 4 September 2011 (UTC)[reply]

No objection from me.  :) Philippe (WMF) 03:31, 4 September 2011 (UTC)[reply]
Nor me, making the tables more user-friendly and readable is a *good* thing. Risker 03:53, 4 September 2011 (UTC)[reply]

Results by project?[edit]

Since the votes were cast separately according to the user's principal project, it would be valuable to see a tabulation by project, to see if there Is particularly strong or weak support by any one project's users? DGG 04:35, 4 September 2011 (UTC)[reply]

I was actually interested in this information, too. Was it stored when users took the poll? Killiondude 05:21, 4 September 2011 (UTC)[reply]
Project-voted-on information is part of the non-public information linked to the vote, so this will need to be worked on by someone who is cleared for private information. For some smaller projects (which make up the majority of projects, although they do not account for the majority of wikimedians), the numbers of active users are so small that a public release of project by project data for these groups may be sufficient to identify editors to their votes. However, for larger projects that have a significant number of editors who participated in this plebiscite (say, 350 or more voters or roughly 1.5% of total votes), it might well be reasonable to provide this information publicly, with a collective figure for projects with less than that number of votes. From my own perspective as an administrator of this election, I agree that this could bring some useful insight to the results, and I have personally recommended that such a study be done. However, something else to keep in mind is that editors were able to vote from just about any project they wanted, so there is no guarantee that the data will be attributable to the user's principal project. Since this is the first plebiscite of its kind within the Wikimedia family, we're still working out some of these sorts of details for analysing the data while protecting user privacy, and resources weren't allocated for anything more than the data that's already been provided. Risker 05:52, 4 September 2011 (UTC)[reply]
Please also add the overall percentage (number of votes) per project to see the overall distribution of participation. One interesting question is how language centric the poll was. This can also be done for smaller projects without risking a leak of personal data.
It would also be interesting to know the number of invitation mails that where send per project/language. Don't know if you can still do this, but would be a very nice addition to the statistic. --Niabot 08:19, 4 September 2011 (UTC)[reply]
Hmm, no reaction so far? Are there any plans to release per project data? That would be the most interesting data. Do you don't have the needed data, not time or do you plan to hide it? --Niabot 09:53, 5 September 2011 (UTC)[reply]
My apologies, I thought we had made it clear that we are gathering requests for further analysis and release. That's what this section was. Therefore, it's on the list. We're in the process of requesting the resources to execute as many of these as are possible. There was no attempt here to ignore the question, I honestly thought we had communicated it. --Philippe 18:43, 5 September 2011 (UTC)[reply]

Result by date?[edit]

Can you release date and viting-project data for the full set of results? I'd also like to see a comparison of results by date as the vote progressed. SJ talk | translate   05:23, 4 September 2011 (UTC)[reply]

If you mean number of votes on each date during the election, this is public information that's available in the voter list. I suspect that an enterprising coder could probably come up with a script to gather that data fairly quickly without needing the committee to do this. They would just have to ensure that their script doesn't include the "greyed-out" votes that represent votes cancelled when the user revisited their vote. Having said that, there was an enormous increase in voting in the 72 hours after the emails went out. I've commented above with respect to the voting-project data. Risker 05:57, 4 September 2011 (UTC)[reply]
There should be a way to compare the results before the e-mail to results after: I speculate regular users vs. incidental users and the biased nature of the spam mail will have had some influence. Zanaq 06:58, 4 September 2011 (UTC)[reply]
Well,you can look at it yourself, the voters list with time stamps on the votes is publicly available. The emails notifying wikimedians of the plebiscite went out starting late on August 18th. It's pretty clear that people will speculate all kinds of things. Risker 14:21, 4 September 2011 (UTC)[reply]
Would there be a hyperlinklink to these data? Zanaq 15:22, 4 September 2011 (UTC)[reply]
From any project, go to Special:SecurePoll. The Image filter referendum is the bottom vote listed; click on the link to the list, which lists all of the voters with a timestamp to their votes. Risker 15:27, 4 September 2011 (UTC)[reply]
The list is empty for me. Can't see any result. --Niabot 15:30, 4 September 2011 (UTC)[reply]
http://wikimedia.amellus.net/index.php/Special:SecurePoll/list/230 -- Marco Aurelio 15:37, 4 September 2011 (UTC)[reply]
Ah, thanks MarcoAurelio. Risker 15:39, 4 September 2011 (UTC)[reply]
Not very informative. At least we can see that the vote count increased drastically on August 19th. --Niabot 15:46, 4 September 2011 (UTC)[reply]
As I've noted above, I have no doubt that we have a huge supply of script-writing editors who could produce a nice graphic representation of the daily vote counts based on this public information. Risker 15:49, 4 September 2011 (UTC)[reply]
I must say that this is not the only data I'm interested in. Thats why i said that it isn't very informative. Results/Attendance per language/wiki would be much more interesting. --Niabot 15:52, 4 September 2011 (UTC)[reply]
You can see which voters voted but not what. I'd rather have the reverse data: see when/what but not who. This is useless. Zanaq 16:34, 4 September 2011 (UTC)[reply]
I'm going to try to get you vote count by day today. Philippe (WMF) 17:06, 4 September 2011 (UTC)[reply]
Day by day
Rough vote distribution
' Voted Total Votes
16-Aug 1509 1509
17-Aug 1137 2646
18-Aug 1126 3772
19-Aug 8198 11970
20-Aug 4650 16620
21-Aug 2359 18979
22-Aug 1738 20717
23-Aug 955 21672
24-Aug 597 22269
25-Aug 482 22751
26-Aug 363 23114
27-Aug 284 23398
28-Aug 316 23714
29-Aug 415 24129
30-Aug 326 24455

Great, thank you! So we could reasonably have stopped the banners after 5 days and gotten almost the same quantity of votes. (I'd like us to always be thinking carefully about how we use our banner space, which is an extaordinarily powerful messaging tool; it tells the world what to think about our work, our priorities, our processes, and our passion.) I'd also love to see the vote totals for those voters from the first 5 days v. those voters from the last 10 days, to check for any significant differences in perspective. SJ talk | translate   23:37, 5 September 2011 (UTC)[reply]

Actually, I believe (and am confirming) that the fifth day dropoff reflects when we DID stop the logged out /anonymous banners. I think we probably could have gotten another 1000 votes by leaving them on one more day, but would love to experiment with something like only running them on weekends or something... I worry about just getting the people who happen to log in during one five day period or something. There are a ton of combinations we could play with. Philippe (WMF) 00:10, 6 September 2011 (UTC)[reply]

Request for more analysis[edit]

From Tim Starling's post to Foundation-l: He requests a correlation analysis of some kind. For example, it would be interesting to know whether those who support the filter have differing views on cultural neutrality to those who oppose it. Philippe (WMF) 01:25, 5 September 2011 (UTC)[reply]

word frequency[edit]

First lets eliminate words like "the", "do", "a", ... "some", "there", that will be found in any text in english language. And then single letters with no meaning at all like "n" and "s".

Lets concentrate on the keywords:

Word Frequency
images 4274
wikipedia 2523
image 2190
feature 2155
users 1678
filter 1476
content 1462
 
Word Frequency
important 1157
controversial 1143
censorship 1127
user 987
categories 874
neutral 865
wikimedia 817
pictures 773
 
Word Frequency
offensive 724
view 720
culturally 648
cultural 647
default 603
children 593
sexual 577
filtering 533
wiki 524
question 519
option 502

Some of these words were part of the query so its not a surprise to find them here. But there is one word, that was not intended to be used by the users, but is relatively high on the list. Think about that single word. --Eingangskontrolle 10:56, 4 September 2011 (UTC)[reply]

I don't see how. If I remember correctly, I commented something like "This is not about censorship". Single words don't mean anything. --Elitre 12:05, 4 September 2011 (UTC)[reply]
Quoting from the document itself: Note that differences using modifiers such as “not” were analyzed as well. For instance, the word “controversial” appears 1143 times in the comments, with 19 instances of “non-controversial”, 11 instances of “uncontroversial”, 2 of “noncontroversial”, 1 of “not-controversial”, 1 of “not-so-controversial”, and 7 of “not controversial”. These numbers are representative of other modifiers sampled. These modifiers make up a small but significant percent of the words and should be evaluated more fully for the committee’s final report.. Philippe (WMF) 20:59, 4 September 2011 (UTC)[reply]
What's the word that was used 8,908 times? Nyttend 12:08, 4 September 2011 (UTC)[reply]
Great question. I just checked the raw log I was sent, and it's blank. I'm going to ask the guy who ran the query. Philippe (WMF) 12:56, 4 September 2011 (UTC)[reply]
Torchwood. Rich Farmbrough 14:10 4 September 2011 (GMT).
Interesting one, actually - I just heard back from the guy who did the graph... the missing word is " ' ". Just a single apostrophe. Philippe (WMF) 18:54, 4 September 2011 (UTC)[reply]

@Elitre - you are right in saying that single words mean nothing. They just document that certain questions around single words were discussed. So this list is just another way to throw sand in the eyes of the users.
More important are the 41.2% "negative" versus the 29.8% "positive" - which means that no mayority to introduce filters was reached. --Eingangskontrolle 19:03, 4 September 2011 (UTC)[reply]

I'm fairly certain that if I'd removed single words (I thought about it) you'd have come back at me with allegations that we were witholding information. I chose to over-disclose, and - of course - I'm throwing sand in your eyes. This is getting really sad.  :( Philippe (WMF) 19:25, 4 September 2011 (UTC)[reply]
It was sad from the beginning, that the WMF did not ask the real question. But this list of words alone is just that, a list. Language is the the art of connection words to sentences. The only interesting information to be found in this list is that "censorship" was used by a number of users. We have to know the context to get the whole picture. --Eingangskontrolle 07:36, 5 September 2011 (UTC)[reply]
Please, please, never do a frequency analysis or word cloud again. This is an embarrassment. Just read the comments and categorize them. I know it's seductive for a coder to go and write a frequency analyzer but this is much worse than ignoring all the text field comments. Comet Tuttle 01:22, 5 September 2011 (UTC)[reply]
+1 --Hagman 18:20, 5 September 2011 (UTC)[reply]
Did you read the report? It makes it clear that we did exactly as you suggested, Comet Tuttle... we read the comments and categorized them. They were also sent through word frequency. However, I'm confused that you seem to think we did the second and not the first part. I assure you, as the report says, every comment will be read and categorized. Philippe (WMF) 18:45, 5 September 2011 (UTC)[reply]

Vote analysis by project and degree of passion[edit]

The results for some questions are polarized, with higher levels of 0 and 10's rather than middle values. This happens quite a lot (think "pending changes") so we need to anticipate this and try to design feature surveys to be more useful in the face of polarized replies.

In this case, there is no data so far how replies vary by community/language. For example, if hypothetically we found (for illustration and ignoring the cultural stereotype) that most US respondents rejected filtering and most middle east respondents supported it, that suggests a possible solution of developing it, then enabling on some wikis and not on others, would leave more people feeling fairly satisfied. While nothing will be that stereotyped, perhaps a breakdown by region or language will be revealing as to attitude/preference/readers' principles.

There is no data so far showing why supporters support it, or why rejecters reject it. It might be that some reasons are not incompatible, or can be worked around. Similarly, the principles drawn on in comments could be important and revealing. A user who feels "Wikipedia should respect cultural views" or "Wikipedia should be uncensored" is expressing a different kind of view (one related to global use) than one whose objection is based on "waste of time" or "no value to me" but has no other strong objection.

Future surveys of this kind should aim to allow this kind of data, if it's practical.

What I'd like to see in further analysis:

  • How did responses relate to project/culture/geographical region?
  • For those replying 0-1 (or 0-2) out of 10, and those replying 9-10 (or 8-10) out of 10, and left comments - what were the principles/reasons they gave that appeared to be their main reason for their strongly held positive or negative views?

FT2 (Talk | email) 13:43, 4 September 2011 (UTC)[reply]

1 - 10 is too fine for people to judge. This probably contributes to the polarisation. 1 - 5 is plenty enough choice. 86.181.174.161 14:10, 4 September 2011 (UTC)[reply]
It will also will be hard to distinguish people who said "It is important" meaning "It is important and it would be terrible" as opposed to "It is important and it would be great"Rich Farmbrough 14:15 4 September 2011 (GMT).
And it is hard to tell what people thought voting 0 or 1. Did they mean "totally unimportant but harmless" or "I want to express my opposition"? So FT2s Question ist quite interesting. What did those people with strong oppinions comment? Adornix 14:45, 4 September 2011 (UTC)[reply]
I think it comes down to the fact that we didn't ask whether people support the filter or oppose it (or present arguments on each side of that question), so we simply don't have that data. To try and derive it from what was asked does a disservice to the referendum process, which was about how, not whether, to implement a filter.--Trystan 15:29, 4 September 2011 (UTC)[reply]
What irks me is that the vast majority of victims of this "feature" – the readers – were disenfranchised entirely. That's more than a little odd for a fundamental shift in the project's philosophy. — Coren (talk) / (en-wiki) 18:57, 4 September 2011 (UTC)[reply]
Given that 99.5% of readers are not editors, I would like to see one other analysis: whether the few users who have many edits responded differently from the many who have few or no edits. The problem being that discussion of the issue is by those who are engaged in Wikimedia which is a tiny minority. If it turns out that of the 24,000 responses, 12,000 are by people with few or no edits, then those are more likely to be representative of responses of the 99.5% of non-editors. In other words, the debate here may be self-defeating because it's a debate between people in the tiny minority of engaged users. What I'd like to know is whether responses by people with very few or no edits, differ qualitatively from those by people with many edits. If so, which set of responses probably matters more...? FT2 (Talk | email) 21:44, 4 September 2011 (UTC)[reply]
Or might have been were the methodology in this survey not hopelessly broken. A first year stat student could have told you that you don't give such a large scale, or provide a middle value; and the questions were framed by begging the implementation as a forgone conclusion. The first could be salvaged, the latter not so much. — Coren (talk) / (en-wiki) 22:53, 4 September 2011 (UTC)[reply]
What I'm looking for is what we can get from the data we have. For example examination of strong views v. reasons in comments is possible, and if IP or account data was kept to detect or prevent multiple votes (and is still available), then analysis by geography/language/region and qualitative differences between respondents with very few or no edits v. the rest might be possible. They might give useful information even after the limitations in the survey. FT2 (Talk | email) 23:09, 4 September 2011 (UTC)[reply]
It may have been better to ask the "end users" too. But nobody will be a "victim" of this new feature because it's opt-in and nobody who doesn't want to will get images filtered. Adornix 19:13, 4 September 2011 (UTC)[reply]
You're making the presumption that nobody will be "opted in" against their will, or by social pressure. Or by whoever provides their sole 'net access; much of the non-occidental world can only access our projects through public terminals (cafe, libraries, etc). Do you really believe they won't put rules mandating such filters in place?). You're also presuming that the oh-so-convenient infrastructure we'll gleefully provide will not be used by anyone else. If there is in-band signalling "classifying" images (as must be if there is code to hide them), then it becomes trivial for anyone holding a mandatory proxy to filter them – user preference be damned.

Neither of the presumptions are valid in the real world. — Coren (talk) / (en-wiki) 19:42, 4 September 2011 (UTC)[reply]

In addition to the potential for third-party censorship that Coren describes, the victims of a filter, and the warning labels which it makes use of, are those a) prejudiced against certain content because we make the choice to label it "controversial", and b) those groups who are most likely to have their own depictions labeled controversial. For example, if depictions of certain people with disabilities are deemed "controversial" because they disturb some users, we become an instrument of reinforcing societal prejudices against those groups. I am not entirely opposed to the implementation of the filter, but implementing it while turning a blind eye to its negative effects means we can't do anything to mitigate them.--Trystan 04:34, 5 September 2011 (UTC)[reply]
Just because labeling the image may upset the subject or creator of the image, doesn't mean the reader's feeling shouldn't be take care of. There is no right or wrong of this subjectivity and selectivity of individuals. What you're trying to do now is like this: a man on a street which is full of passersby shows his disfigured face openly. When the passersby are turning their face away or covering their eyes with hands, that man fiercely force them to look at his unfortunate face by turning their face on him and move their hands from their eyes. That man then eloquently stats that the law does not forbid him from showing off his disfigured face openly. -- Sameboat (talk) 04:56, 5 September 2011 (UTC)[reply]
I have some difficulties understanding your line of argumentation here. Are you saying that disabled people shouldn’t have a place in public life because some people might find them ugly? -- Carbidfischer 08:02, 5 September 2011 (UTC)[reply]
No, but they shouldn't stop the others from avoiding them by using the filter. -- Sameboat (talk) 08:06, 5 September 2011 (UTC)[reply]
So you value the right of not seeing disabled people higher than the right of disabled people to a place in society? You gotta be kidding. -- Carbidfischer 09:15, 5 September 2011 (UTC)[reply]
Both things do not conflict each other. That vulva image was eventually restored on the de.wp main page to Jimbo's demise. (If you prefer, you can vote for disable/nude image to be displayed on the main page too.) Now we just ask for the option to hide it before our own screen, not yours, not even arguing its right to be displayed openly. You guys are so insatiable to strip that option from us too. -- Sameboat (talk) 09:29, 5 September 2011 (UTC)[reply]
There is nothing wrong with a vulva on the main page: it is a featured article. Featured articles will sometimes be displayed on the front page including the relevant picture. The only neutral conclusion is to show the article including the relevant picture on the front page. Zanaq 11:18, 5 September 2011 (UTC)[reply]
So wrong in so many ways, surely not from your POV. You have not a single apprehension to those who actually in the wrong place to have that vulva image loaded unexpectedly, do you? All you can say now is "it's only your fault for not able to accept that image", right? -- Sameboat (talk) 11:50, 5 September 2011 (UTC)[reply]
That is called neutrality: the only values we should care about are encyclopedic values, like verifiability. If featured articles will appear on the main page, then all featured articles should be treated the same, regardless of content. Not make some exceptions based on some very personal objections. There is no neutral standard for objectionability, so we should not use objectionability as criterion for any encyclopedic content. Zanaq 18:26, 5 September 2011 (UTC)[reply]
Great, whenever talking about the practical problem of loading sensitive image by some user, you just use the encyclopedic values as your escape route. We are NOT denying its encyclopedic value. Now you're that disfigured man (objectionable image) who forces everyone (who want to apply the filter) to look at him because the law grants him the right (encyclopedic value). You're purposely making someone uncomfortable in the name of encyclopedic value, even though they have already respected your right to display objectionable image as you please. -- Sameboat (talk) 22:31, 5 September 2011 (UTC)[reply]
So the purpose of making vulva a featured article is to offend people, yeah right. If the disfigured man crosses the street, other traffic is supposed to make eye contact. There is no purpose here: the man is traffic so will have to be treated as traffic. You cannot deny his right to walk the street so people won't have to make eye contact. I see no-one here making other people uncomfortable on purpose, that is just a side-effect of neutrality. If only non-neutral ways are available to fix this and draining the resources, it shouldn't be fixed. If you don't want to see pictures of a penis, don't visit the article. If you don't want to see random featured articles, don't visit the main page. Or maybe we shouldnt publish featured articles on the main page at all. But once you do: any content should be treated exactly the same: that is the only culturally neutral (and let's not forget feasible) way to handle things. Zanaq 11:10, 6 September 2011 (UTC)[reply]
Thanks for restoring my belief in humanity. ;-) -- Carbidfischer 12:13, 6 September 2011 (UTC)[reply]
So you're telling people don't visit the main page either. The disfigured man I'm speaking about is using force to remove the eye band from people who don't want to look at him. If this is the humanity you're talking about, I call it selfishness.
The eye band supplied by WMF, this is what the debates stemmed from. IMO when the platform to display images which may turn off some visitor deeply is provided by WMF, it is absolutely appropriate for WMF to also provide the countermeasure for such dilemma. -- Sameboat (talk) 01:16, 9 September 2011 (UTC)[reply]
Actually, the bigger problem with Zanaq's "encyclopedic" argument is that Wikipedia is not the sum total of all WMF projects. Commons doesn't care if the material or its viewing is "encyclopedic". Wikibooks's basic goal is specifically to host material that is not "encyclopedic". The encyclopedia project is the biggest, but it is only one of dozens of WMF projects. WhatamIdoing 15:16, 6 September 2011 (UTC)[reply]
Other projects may have their own goals and editorial judgement which also probably do not include objectionability, since any two people will never fully agree on what's objectionable and what is not. Zanaq 07:58, 7 September 2011 (UTC)[reply]
You’re evading my question. Plus: You can’t strip anyone of anything that this person doesn’t already have. -- Carbidfischer 09:48, 5 September 2011 (UTC)[reply]
I'm not. It's your problem for correlating 2 unrelated issues. Disable people have their right in the society, but that must be upon respect to other people who don't find their appearance swallowable. Filtering those image is only the user's own affair, not the subject or creator of the filtered image. -- Sameboat (talk) 10:43, 5 September 2011 (UTC)[reply]
Apart from your interesting position on disabled people, you make a good point. Filtering is indeed only the user’s own affair and not an affair that Wikimedia or the communities have to address. -- Carbidfischer 11:00, 5 September 2011 (UTC)[reply]
No. Before this proposal, we do not have a direct mean to filter the sensitive images from Wikimedia projects without 3rd party script which may pose a threat to our computer security. WM gives us a reliable way to achieve this goal without worrying about security issues, just as it offers an open platform to share knowledge. We just want to hide the images selectively, apart from that request, we demand nothing more. The filter will not cause further problem to the usage of the labeled image. You can proudly display any images passed featured picture/article, vulva, disability, nudity, Mohammed, any image you want without considering users who dislike them. -- Sameboat (talk) 12:11, 5 September 2011 (UTC)[reply]
There are no more or less security concerns if this filter is available or not. It is not a replacement for childcare software nor it is a filter for the whole web. It serves only to discriminate our own carefully chosen content. --Niabot 12:22, 5 September 2011 (UTC)[reply]
It's not just childcare software, it's a filter for every user of all ages. It's only you would discriminate the labeled image to justify your opposition against the WMF proposal. We, the filter users, will never do such stupid thing to challenge the labeled images to be unsuitable for WM projects. -- Sameboat (talk) 14:27, 5 September 2011 (UTC)[reply]
Now you made my day. What a great joke. ;-)
An image is already discriminated by choosing to hide it under certain conditions (possible misuse by third parties aside). Especially if the reader would not choose to hide this particular image, when some other group of people decide to hide it. It wasn't his choice, it was a choice made by others for him. That's already enough to be called censorship, while the filter is the censor's tool. That isn't only my conclusion. [1] [2] [3] [4] [5] [6] ... --Niabot 00:02, 6 September 2011 (UTC)[reply]
Say it again when the WM filter categorization is actually abused by 3rd party as a censorship tool. I only believe such hypothesis when the result comes out. -- Sameboat (talk) 01:18, 6 September 2011 (UTC)[reply]
Say it again when Mubarak/Gaddafi/... actually abused their power to threaten their own people. I only believe such hypothesis when the result are dead people laying on the streets. Quite a statement you made here. If this is already the case and until we know it, the damage is already done. --Niabot 11:29, 6 September 2011 (UTC)[reply]
The ability to hide individual images on a case-by-case basis if you find them upsetting is the equivalent of your right to look away from persons with disabilities. A label-based filter is the equivalent of demanding that all persons with disabilities be pre-labeled as objectionable and then properly announced so that you can avoid seeing them at all. That process of labelling and warning is deeply incompatible with NPOV projects, intellectual freedom, and basic human dignity.--Trystan 22:56, 5 September 2011 (UTC)[reply]
Whether the proposal compatible with NPOV or not is only the matter of philosophical view. AFAIK, normal user would never be able to tell if the image is labeled if the filter is never enabled. I doubt there's any intellectual freedom as WM server belongs to private domain, any law related to freedom of speech of the US is not gonna applicable here. But I admit the filter would discourage some user from contributing images. -- Sameboat (talk) 23:40, 5 September 2011 (UTC)[reply]
The Harris Report describes intellectual openness as "the overriding principle that animates all Wikimedia efforts – the unrelenting, unremitting and rigorous commitment to non-censored openness and “intellectual freedom” that Wikimedia attempts to provide for the world." The report discusses "modifying" this principle to balance it with other considerations. I would use stronger language (i.e. infringement) to get the point across, but intellectual freedom is still the core principle for the projects.--Trystan 23:43, 5 September 2011 (UTC)[reply]
That freedom is the privilege granted to us by WMF. Opt-in filter is not censorship, no matter how hard you prove the possibility of the filter being abused by 3rd party, it's still essentially not a censorship tool. Any "censor-" terms in the report have nothing to do with this proposal. You still have your intellectual freedom even the filter is implemented. -- Sameboat (talk) 01:18, 6 September 2011 (UTC)[reply]
Re comments: I'm disappointed that the first response to the comments people left was to put them thru a pattern-crunching program & create a pretty picture of which words were used the most. After looking at this image, I felt that I would have had more impact in this plebiscite had I simply wrote "censorship" repeatedly until I used all of the space allotted, than what I did write.

What this "plebiscite" is asking is an ideological question, which means the opposing schools of thought are not going to find a common ground to agree on. (By "opposing schools of thought", I mean each side believe that the other is a bunch of ill-informed scumbags who are subverting the ideals of the projects. We're arguing over religious beliefs here, people. When the Christians, Jews, Moslems & Atheists all agree on who Jesus Christ was, then maybe we'll have a consensus on this issue.) Unless someone can produce a solid reason why this would benefit any Wikimedia project, then the Foundation is simply pissing away finite resources that would be better spent on something else -- such as US$50/€50 gift cards to established editors to help them buy research materials to create content.

As for the issue of knowing what the users of our content actually wanted, the best way to have obtained that knowledge would have been to contract with a polling firm, who would find a large enough population of Wikipedia/Wikimedia users for an accurate sample, & ask them. -- Llywrch 22:33, 4 September 2011 (UTC)[reply]

Agreed. Especially considering the delay in the results, why spend time making such a picture? That kind of communication where all that matters is the word count - that's a modern-day politician's kind of discourse. Not a bridge for the communication of ideas but a wall meant to convey only the murmurs of human voices while holding back their content. Wnt 08:03, 5 September 2011 (UTC)[reply]
A third-party poll (of a much smaller # of people; you don't need thousands) to supplement any projects-wide poll sounds like a fine idea. [consider the value of the banner space we're using; it's ok to spend 0.01% of that on a sanity check that improves our techniques.] We can compare survey methods with the pro's in the process. There are advantages to both mechanisms -- we certainly need a smoother method for surveys, community-driven proposals, and actual referenda, and I'm glad we're putting energy into making that happen. SJ talk | translate   19:51, 5 September 2011 (UTC)[reply]
But a real poll might yield real results and I’ve come to think the Foundation doesn’t want to get facts in their way. -- Carbidfischer 08:02, 5 September 2011 (UTC)[reply]
I can't speak for the Foundation, but the only point to running such votes is to get useful data and 'real' results. Clearly we have much to learn about survey design, but I see that as a learning experience. (we should definitely involve the Research Committee in such efforts and build that sort of expertise within our community, and we run many surveys every year in different contexts.) SJ talk | translate   19:51, 5 September 2011 (UTC)[reply]
A lot of people responding to my comment above seem to have missed an important point, one that I have been repeating over & over. Creating this filter is a waste of finite resources. The Foundation is wasting money on this Maybe we volunteers ought to start our own grass-roots counter-campaign this year to encourage people not to donate funds, & hit the WMF in the pocketbook where it hurts. Maybe then you'll understand why so many people are upset at this stupid proposal. -- Llywrch 06:30, 6 September 2011 (UTC)[reply]
You need not frame this as 'we v. the WMF'. That suggests a lack of direct influence and participation in Foundation and community priorities that isn't apt. Public funds raised are mainly given to the projects as a whole, with the Foundation as mechanism. "Hurting donations" hurts the movement as a whole, yet hurts "the WMF" less than saying, as you do here, "this is a poor use of time and money" and pointing out better alternatives. [the WMF's bottom line, unlike many foundations, isn't "How much money did we raise this year?", it is "How well did we serve the projects and mission this year?" And that is defined in large part by community input and participation.]
Unfortunately, this and all previous rounds of discussion about controversial content started with community groups advocating for change. We wouldn't even be discussing this if that were not the case. At some point that perennial community debate was picked up and amplified by a Board and Foundation discussion about whether anything should be done to address the concerns of those community groups, and all of the related discussions about the different audiences we serve and who does and does not have a voice in wiki-driven planning -- and here we are 18 months later.
So by all means start a campaign. But start it to support your view of what we should be doing instead. SJ talk | translate   21:17, 6 September 2011 (UTC)[reply]
As I said further down on this page, I find it hard to believe that Wikimedia is so blatantly incompetent. It’s not like you had to invent polling, there are people who know how to poll and have polled before. You could, like, have asked one or two of them.
If your really see this referendum as a learning experience, you’re not just incompetent, you’re cruel. Thousands of people had to work through this poll just so that you and your friends could learn some things that are common knowledge for everybody who knows the least bit about polling? -- Carbidfischer 21:52, 6 September 2011 (UTC)[reply]

When taking the survey, I was annoyed by the leading questions - all of them after the first one (importance) are irrelevant if the respondent is opposed to adding the filter in the first place. I would even go as far as saying the data from questions 2-6 is significantly skewered by this.

...suggests the presence of a sizable number of voters who wished to express their opposition to the basic idea of the image hiding feature.

To me, this sounds like a complaint about respondents misusing the survey form. But the large mass of '0' votes on all questions really means there was frustration over the implicit assumption that such a filter would be acceptable. In future surveys or polls, I'd recommend more general questions about whether such a filter is "right" or "wrong", and why. Specific details about implementation of such a filter don't seem very important until the bigger question, "should we?", is answered by the community.

Further, I think the free-response portion of the survey carries significantly more weight than the multiple-choice answers. This is a very deep issue morally, and free discussion and expression of thoughts works much better than "rate 1 to 10" questions. For this reason I'd love to see a massive open straw poll on whether the filter is a good or bad idea. Technician Fry 15:53, 5 September 2011 (UTC)[reply]

Results by edit count[edit]

May you give the results by user edit count? Let's say, logarithmically organized: with <10 edits, <100, <1000, <10,000, <100,000? --Millosh 07:45, 5 September 2011 (UTC)[reply]

Even 10 is a fairly dedicated editor, by global standards. FT2 (Talk | email) 07:53, 5 September 2011 (UTC)[reply]

That would most likely give an unfair result. I spend more time on Wikisource than on Wikipedia. But the number of edits is larger on Wikipedia. -- Lavallen 11:30, 5 September 2011 (UTC)[reply]
Lavallen points out exactly my hesitation to that type of analysis. Philippe (WMF) 18:52, 5 September 2011 (UTC)[reply]
I would also like to see this breakdown, even understanding the bias that may be involved. Within a given project, edit count is a reasonable-if-flawed estimate of activity. [visualizers: if you have a better one, use that instead :) ] SJ talk | translate   19:51, 5 September 2011 (UTC)[reply]

Statistical data give quantitative, not qualitative results. And quantitatively, it is easy to suppose that the most of editors with more than 10k edits are more involved than the most of editors with less than 1k edits. There are exception, of course, but if we want to have better insight into those numbers, I think that it would be useful to have results presented in that way. --Millosh 06:02, 6 September 2011 (UTC)[reply]

My premise is that editors with more edits tend to be negative toward image filter and I would like to see that proved or disapproved. If everything stays the same by edit count, then it could be said that just significant minority is against the filter. If I am right, those votes are not just significant minority, but the votes of the core of our community. If something else, it should be analyzed. --Millosh 06:02, 6 September 2011 (UTC)[reply]

I agree with Milosh, it would be very useful to present us that data, and I agree with his premise that editors with more edits tend to be negative toward image filter. So, is it technically possible to show us that data? (Note: It can be edit count on the project from which user voted or global edit count, because some users are active on several different projects)--В и к и в и н др е ц и 09:01, 6 September 2011 (UTC)[reply]
I think it would be more interesting to see the opinion off the readers of our pages. The opinions of the editors look less interesting. As an editor I can create a filter of my own. (And yes, I already have one.) -- Lavallen 09:26, 6 September 2011 (UTC)[reply]
Well, the editors have to do the work, so, if they don’t want to do it, it’s kinda pointless to look at the readers’ opinion ... -- Carbidfischer 10:50, 6 September 2011 (UTC)[reply]
Mere readers were not permitted to participate in the referendum. The eligibility requirements specified registered editors only, with a minimum of ten edits before 01 August 2011. WhatamIdoing 15:22, 6 September 2011 (UTC)[reply]

Results by country[edit]

I would like to see some analysis of the results broken down by country, or by language. John Vandenberg 07:00, 10 September 2011 (UTC)[reply]

German Wikipedia poll[edit]

84% opposed in ongoing poll on de:wp[edit]

At the moment vote 84% of the germans wikipedians in a real poll against this filter. Other wikis have to love filter for these result. Maybe there are other reasons (kind of questions,...)? An-d 07:17, 4 September 2011 (UTC)[reply]

Correction: 84% of the active parts of the German community who voted in an open poll. You're likely not attracting a lot of casual editors or readers with a poll like that.  :) I'm not saying it's bad, just that we have to state clearly what it means. Philippe (WMF) 08:02, 4 September 2011 (UTC)[reply]
Thank you for your specification. I believe in the Law of large numbers An-d 08:16, 4 September 2011 (UTC)[reply]
You can believe in UFOs if you want. Law of large numbers needs random samples of the whole population and multiple experiments. Who can vote in a German Wikipedia poll? I guess only registered users with > X edits. That is not a random sample. Enjoy your numerology knowledge. Emijrp 09:09, 16 September 2011 (UTC)[reply]
Well, WMF polls are not likely to be that attractive for casual editors either. -- O.Koslowski 08:18, 4 September 2011 (UTC)[reply]
Actually, I hope that we CAN do the analysis on these numbers to see how effective this one was in reaching casual editors. Our experience in vote checking showed that it might have actually done okay at that, but that's anecdotal and highly skewed by the fact that we were looking for particular trends associated with multiple voting. I hope we can get some analysis done to tell us if that's correct or not. Philippe (WMF) 08:19, 4 September 2011 (UTC)[reply]
An-d, I believe in the law of large numbers, too. And I believe that 250 editors, out of the thousands of editors at de.wiki, is what we call "a small number".
Furthermore, I believe in en:Groupthink, and I know that the first few comments in an open poll tend to bias all subsequent responses. Consequently, the law of large numbers operates best when the later participants have no idea how the previous people responded. In this respect, the referendum (with its far larger numbers and correct use of a secret ballot) produces far more reliable results that one page at de.wiki. WhatamIdoing 00:05, 5 September 2011 (UTC)[reply]
From the tone of the comments on the poll page, I'd hoped for a very strong rejection of the filter. But what we need to remember that a poll about importance of one issue needs to be compared to feelings about the importance of other issues. This outcome is a collective lukewarm "huh maybe" to one idea about how the devs can spend their time. Maybe other features, if put to such a poll, would get 90% of voters saying that they're important. I'd expect any proposed feature with a weak consensus behind it to do at least as well as this, probably better. So I'm still suggesting the WMF should look at the commentary on the talk page, understand how appalled editors are about the idea of people trying to form consensus about which images to annotate for special hiding, and rethink whether there is a way to allow independent groups of editors to make their own ratings without needing to patrol and enforce their "consensus" categories against all other opinions. Wnt 08:19, 4 September 2011 (UTC)[reply]
I know that an open poll like this (asking for importance, without questioning deficits) and a project internal poll (asking for acceptance, stating pros and cons) is quite different. Overall this is not so surprising to me. I could guess that this kind of questioning would provoke such a result. Asking for importance is something very relative. Asking the people if "nice weather" is important and asking them if "being able to breath" would be important, would both easily receive a majority. But if you would ask them if "nice weather" is more important than "being able to breath" would give a complete different picture. Like Wnt already said. This must be accounted for. --Niabot 08:36, 4 September 2011 (UTC)[reply]
You both raise good points. There are a few different questions to be asked about how to proceed: Should any work at all be done to give readers more control over how images are displayed to them? And if 'yes' or 'mmmaybe', how much effort should go into that? The latter question is important to any consideration of such ideas. And having some point of comparison [for priority] would be a good idea for future 'feature' discussions. SJ talk | translate   20:12, 5 September 2011 (UTC)[reply]
You're likely not attracting a lot of casual editors or readers with a poll like that. This is my biggest concern - on first blush the image filter is a great idea (barring, if I may be so rude, knee jerk reactions), it is only on reasonably close analysis, and with some technical knowledge that the significant problems begin to show. Reading between the lines, or explicitly, some of these had been apparent to Harris, the board or the designers, it is unfortunate that they weren't collated and either presented as part of the information on the poll, or indeed used to inform the construction of the poll itself. Rich Farmbrough 14:10 4 September 2011 (GMT).

The committee was interested in a particular subset of those who used a positive tone in their comments We have known that from the start of this farce. --Eingangskontrolle 10:33, 4 September 2011 (UTC)[reply]

Interested in *why* they used a positive tone, and what they expected. Risker 14:23, 4 September 2011 (UTC)[reply]

FYI: German-language Wikipedia users reject image filter[edit]

I would like to draw your attention to a voting in German-language Wikipedia that at this point of time strongly rejects the introduction of image filters. Above all, the community deeply deplores that the Foundation did not ask whether to introduce a filter, only what kind of filter would be preferred. Voting runs through September 15, 2011. The toll now stands at 230 against/ 43 in favour of the filter/ 14 abstentions.--Aschmidt 12:21, 4 September 2011 (UTC)[reply]

Updated: 288 / 51 / 14. --Tutenstein 19:13, 6 September 2011 (UTC)[reply]
I doubt the legitimacy of the poll from a single local wiki project would have the power to revoke the decision of Foundation. But this would be a shame if DE.WP bureaucrats have the power to prevent its local users from using the filter once it's implemented. -- Sameboat (talk) 14:34, 4 September 2011 (UTC)[reply]
This is a precedent. The Foundation will not be able to neglect the opinion of a vast majority of users on German-language Wikipedia. We just say no, it seems.--Aschmidt 14:46, 4 September 2011 (UTC)[reply]
If the DE Wikipedians vote on that poll, that means they're already well aware of this referendum. That means they could express their opinion through the referendum rather than in their home Wiki, so that local poll is totally unnecessary. And please don't give me that "this referendum is a fraud" crap, it's no more than a cheap excuse and insult to reject something you don't like. -- Sameboat (talk) 14:54, 4 September 2011 (UTC)[reply]
I gather your strong language indicates we have got it right.--Aschmidt 15:04, 4 September 2011 (UTC)[reply]
That poll is not democracy, the "majority" just want to prevent the local proponents from using the filter in DE.WP that WMF intends to offer. Stripping the freedom they should have. -- Sameboat (talk) 15:56, 4 September 2011 (UTC)[reply]
Quite the opposite. They have the freedom already. ;-) --Niabot 15:59, 4 September 2011 (UTC)[reply]
That poll asks one question. A question that was absent inside this referendumm (sic!). --Niabot 15:26, 4 September 2011 (UTC)[reply]
Sameboat, what restricts freedom is the foundation proposal. The freedom deWP should have is the freedom to accept of reject the filter for their WP. If there is anyone having the priviledge to control the way the deWP chooses to have material and the way it chooses to display it, should it not be the consensus of the people who work on that WP? This is just the same as deWP having the right to use Proposed changes,, which they have done, rather than making it a WMF decision that every project must have it (or must not have it) . DGG 16:05, 4 September 2011 (UTC)[reply]
I don't see the point of WMF proposal restricting freedom. And the opinion of having the filter should be based on readers themselves, not the constant contributor group of 1 particular local Wiki. I'm trying not to use that ideology to describe that poll, but it's a very close one in this case. And what if user from other Wiki who have already set the filter switches to DE.WP and suddenly find that the image should be hidden actually appears right before them? I directly refer to that problematic image once on the DE.WP main page. -- Sameboat (talk) 17:04, 4 September 2011 (UTC)[reply]
The reason why the editors get to choose what goes on their wiki is this: it is the editors that are charges with the maintenance, not the readers. The editors are the ones who will need to keep the POV-pushing ideologues at bay. The readers are important, don't get me wrong; but if they are allowed to vote, they might not have a deep concern for the logistics of the proposal. This is one reason why wikis have voting eligibility policies. Case in point: unified Serbo-Croatian on en.wikt, where the votes were 40-50 but it was closed as 30-21 (this was before the voting policy). — Internoob (Wikt. | Talk | Cont.) 17:16, 4 September 2011 (UTC)[reply]
Yes, editors can decide thing that is related to editing. But the filter proposal is directly about READERS' preference, not just editors. That's why the opposition disgusts me, they attempt to use their own power to control the readers' preference for satisfying their own philosophy like a tyrant. I do not advocate non-contributors' right to vote, but we contributors should be more comprehensive on the readers' side, not just ourselves. -- Sameboat (talk) 22:03, 4 September 2011 (UTC)[reply]
You forget that the editors will have to correct the false positives and false negatives, deal with the editors who will try to push their POV/culture into the categories, listen to the complaining offended readers, etc. It's not a thing that will support itself without the editors. — Internoob (Wikt. | Talk | Cont.) 04:44, 5 September 2011 (UTC)[reply]

Just to make it clear: I for one reject introducing the filter because we were not asked if we wanted it in the first place. Period. This is no way to run a wiki or indeed any community these days. And then, if the problem is in the content you find on Commons, please go and solve in on Commons, i.e., delete the problematic content there and do not hold users responsible for their filter settings if they encounter something they don't want to see. No education project relies on explicit material in any way. Never. I repeat, the Foundation will not be able to deny the German community's voting. If we say no to this filter we mean it this way. --Aschmidt 16:22, 4 September 2011 (UTC)[reply]

This is clearly not the point, you're denying other readers' preference to hide the image which is essential to the topic and is subjectively accepted by another group of users. The filter is here to solve this dilemma. -- Sameboat (talk) 17:04, 4 September 2011 (UTC)[reply]
On the contrary, the filter is to be introduced to hide the dilemma. If hides the content rather than removing it altogether. I don't want anyone to filter anything in Wikipedia. Every user should be delivered the same content. Apart from that, I wonder why you are against democracy in our projects? It is a precedent to announce far-reaching decisions like this without asking for the community's consent in the first place. This was a big mistake, as we now can see.--Aschmidt 16:37, 4 September 2011 (UTC)[reply]
I just repeat, it's not democracy. And the DE sysops just act like a supremacy to control its visitors' option to hide the image. Obviously you've no right to force someone to accept the image that may disgust them. And that Vulva image once on the de.wp main page is the perfect explanation why this filter is proposed. The issues about the process of the proposal, ask WMF, not me. -- Sameboat (talk) 16:51, 4 September 2011 (UTC)[reply]
Wait a second, please, you are responsible for your opinion, not the Foundation, right? And then, of course, it's not democracy, that's just what we are claiming.--Aschmidt 16:54, 4 September 2011 (UTC)[reply]
I am not representing the WMF, so the problem for not asking de.wp community's consent about the proposal is not of my concern. The referendum is already sufficient for me to express my opinion on that matter. -- Sameboat (talk) 17:00, 4 September 2011 (UTC)[reply]
There is no "forcing" people to see images. Any more than we are "forcing" them to read stuff they may find unacceptable. Bad editorial decisions should not be used as an excuse for more bad decisions. Rich Farmbrough 17:18 4 September 2011 (GMT).
This is enforcement. When WMF offers the tool to hide image, the de.wp sysops just use their power to ban it, preventing anyone from using it. Who did ask to see the Vulva image on the de.wp main page? Are you that desperate to let everyone see the Vulva image to satisfy your educational purpose? This is far over the line. -- Sameboat (talk) 23:05, 4 September 2011 (UTC)[reply]
de. sysops don't have the "power to ban" the tool. Bad editorial decisions should not be used as an excuse for more bad decisions.Rich Farmbrough 15:24 6 September 2011 (GMT).
  • The significance here is that the de: project feels strongly enough about this to suggest forking. Rich Farmbrough 17:22 4 September 2011 (GMT).
This is not about forking, but rather about pointing out to the foundation it has made a mistake that has to be corrected now.--Aschmidt 17:35, 4 September 2011 (UTC)[reply]
But it may be about forking if the foundation doesn’t change their mind. -- Carbidfischer
Wikimedia is a global projekt - let find us together a way out of this trouble. Forking is no serious option. An-d 17:57, 4 September 2011 (UTC)[reply]
@Carbidfischer: Yes, if they don't change their mind. It would be the next stept, and it would be a logic one. I expect the foundation to move toward the German-language community.--Aschmidt 17:59, 4 September 2011 (UTC)[reply]
In theory, the foundation could fork from the community, just like Oracle did it with the Open Office community. --Liberaler Humanist 19:48, 4 September 2011 (UTC)[reply]

I doubt very much that anyone would seriously consider forking because of an opt-in filter. What might happen is that dewiki ignores the filter by creating local copies of the filtered files, thus removing them from the influence of commons categorization. But that's about it. --Tinz 00:04, 5 September 2011 (UTC)[reply]

This is far too cumbersome considering the amount of labeled images to be transferred. That's why they start that poll to justify the ban on the filter. -- Sameboat (talk) 01:09, 5 September 2011 (UTC)[reply]
Only images that are included in articles would have to be considered. But ok, even if you are right, the important thing would be to set an example. This could easily be achieved by transferring just a few sexually explicit images and images of Mohammed that are included in the high-traffic articles. --Tinz 01:39, 5 September 2011 (UTC)[reply]
Yes my point wasn't "ZOMG de will forkz!" but that de: took it seriously (and negatively) enough to suggest forking. Rich Farmbrough 15:26 6 September 2011 (GMT).

CentralNotice translations[edit]

Is it translatable somewhere? Liangent 12:46, 4 September 2011 (UTC)[reply]

Yes, see Image filter referendum/CentralNotice. Mathonius 14:20, 4 September 2011 (UTC)[reply]
My language (zh-cn in preferences) has translation and it's listed as "published" but I'm still seeing the English one on zh.wikipedia. Liangent 10:33, 5 September 2011 (UTC)[reply]


Expressing opposition[edit]

Asking if there should be a filter at all[edit]

As previously mentioned in a reply, this survey had a fairly significant problem: the questions were about how to implement the filter, without asking a more basic question of if there should be a filter at all. I ask that, at minimum, an informal vote be held to determine community consensus on that issue. Additional useful information would be for support of an option to disable all article images, rather than a category-based filter. If the Foundation has unilaterally decided that there will be a category-based image filter and damn what the community thinks, I believe the community has a right to a transcript of the discussion and vote at which that decision was made. — The preceding unsigned comment was added by 98.230.51.109 (talk)

(Edit: gah, that was me! MaxHarmony 15:02, 5 September 2011 (UTC))[reply]

The first question was already asking: "it is important for the Wikimedia projects to offer this feature to readers." -- Sameboat (talk) 15:03, 4 September 2011 (UTC)[reply]
I'm quite sure Wikipedians are intelligent enough to vote 0 (totally unimportant) to express their opposition to the filter. A "I don't want a filter"-Option would have been better though. Adornix 15:14, 4 September 2011 (UTC)[reply]
But will it counted as opposition? I don't think so. The questions and answeroptions were designed to show some importance anyhow regardless of the answers given and thats enough to go forward. --Eingangskontrolle 20:07, 5 September 2011 (UTC)[reply]
In the absence of more explicit input, yes, people voting that it seems unimportant are being seen as an approximation of opposition to the proposed feature. Looking at this data [with some 'high importance' voters leaving negative comments, and some 'low importance' voters leaving positive ones] I too wish that the separate question had been asked. SJ talk | translate   20:54, 5 September 2011 (UTC)[reply]
When I first encountered the survey, my sentiments were very similar to those of 98.230.51.109; I was disappointed that such a major decision had been unilaterally reached without seeking community consensus or even input. Moreover, I personally recall the thought process of "if I respond with 0/10 in a concrete question of importance to voice fundamental objection, will my other responses and text input be methodologically invalidated as partisan or irrelevant?" going through my mind, which indeed resulted in a considerable artificial deviation in my case. CMBJ 16:49, 4 September 2011 (UTC)[reply]
A helpful anecdote, thank you.
The intoduction of the filter has already been decided by the Board of Trustees (Image filter referendum#Background). Purpose of this referendum was explicitly not to decide wheter we want the filter or not but only how it should look like.
@Adornix: It's not the same to ask "how important is such a filter" or "do we want such a filter" (even though both questions sound similar they aren't equal)... Chaddy 16:02, 4 September 2011 (UTC)[reply]
The assumption that the board is incapable of changing it's mind, is one that I find staggering. Rich Farmbrough 17:14 4 September 2011 (GMT).
I didn't assume this. I only said what the referendum was about. Chaddy 17:52, 4 September 2011 (UTC)[reply]
Chaddy is right: I suspect this is the primary reason there wasn't an explicit "do you want this?" question -- a concern about what it would mean if the community definitively rejected a Board resolution. However, as has been pointed out, that's precisely what referenda should be for. We should welcome this kind of direct feedback: If something is important enough for a projects-wide poll, and the community rejects it, clearly we shouldn't carry it out. (As we do not have clear acceptance or rejection here, the next step will require more discussion and consideration.) SJ talk | translate   20:54, 5 September 2011 (UTC)[reply]
as can be seen from the voters list, I waited until the last day to vote, because i had some difficulty deciding whether to vote at all--whether to participate in such a referendum was acceptance that the WMF had the moral right to pass such a resolution, and whether even voting on a referendum presented as how to implement as issue was legitimate. Should I vote on questions of what way to facilitate censorship?

There is a considerable difference indeed between saying, "it is not important that we have such a filter" and "it is very important that we not have such a filter". The skewed distribution of votes on the first question (the very few 1,2, and 3 votes as compared to the number of 7, 8 and 9 votes), indicates that those who opposed it felt very passionately about it. Risker said above that the examination of the comments indicated that also. DGG 18:11, 4 September 2011 (UTC)[reply]

Now I see what Rich meant about the materials not encouraging those who oppose the whole concept to vote. Thank you. Yes, opponents feel passionately and (by their comments) on ideological grounds. proponents may feel strongly but more on usability grounds, along the lines of 'that would be a useful feature', or 'I know someone who would appreciate that'. Only a small proportion of commenters said that they were eager to use it themselves. SJ talk | translate   20:54, 5 September 2011 (UTC)[reply]
I am actually not strongly opposed to this feature, but IMO the process left a great deal to be desired. First it was misleadingly called a "referendum", and then there was this confused question about how "important" it is to implement the feature, which seems to me like a weasly way of giving people the impression that they could oppose the feature, while allowing a get-out to implement the fait accompli anyway, even if most people said it wasn't "important". 86.160.217.239 21:59, 4 September 2011 (UTC)[reply]
Oh, I see. The Board unilaterally made this decision. In that case, I believe that the arguments made by each board member, and the final vote of each member, should be available to the community. MaxHarmony 15:02, 5 September 2011 (UTC)[reply]
I refer you to the resolution on controversial content, already public on the Foundation's website. Philippe (WMF) 18:51, 5 September 2011 (UTC)[reply]
Max, to your original question: the Board did not suggest a "category-based image filter", it asked that readers be given more choice over when/whether they are shown an image. The use of "new categories" as currently proposed has been a source of controversy whenever it was discussed, on-wiki and elsewhere. Most of the discussion and arguments derived from the Harris report, which was developed and argued publicly here on Meta. SJ talk | translate   20:54, 5 September 2011 (UTC)[reply]
There is also another issue that has been overlooked here: every dollar, euro, &c. spent on this proposal is a unit of currency that could have been spent on something else. Something that the communities who support the Wikimedia projects could agree needed to be done. Even the act to consider whether to filter images was an unbelievably stupid decision; the entire Board ought to be ashamed of themselves for allowing it to be put on their agenda. I am glad to say I did not vote for any of the successful candidates for the Board in the last election, & I hope we get some fresh blood on the Board immediately. -- Llywrch 22:44, 4 September 2011 (UTC)[reply]
I tend to agree; in particular I find it really frustrating that they called this a referendum, but it was really more of a survey. At the very least we should have a non-binding referendum asking whether or not we want an image filter in the first place- and that should be the only question on the referendum.--142.68.162.67 03:21, 5 September 2011 (UTC)[reply]
It wasn't a referendum. A reader clicks to the main page of the referendum. They see a presentation of the filter (described in a positive light) then votes on questions written in a way which gives the impression that the filter WILL be used in the future and it is simply a questionaire of HOW to implement it ... with no encouragement for debate, no encouragement to read the talk page (of which the casual reader has no clue about ... a page filled with criticism) a FAQ that was totally biased during the first week ... and yet still the committee expects to get a balanced view of how the community feels about the filter? The way you ask a question will dictate the meaning of the final answer. The way you present a "referendum" will heavily influence the users' response. How the users click their way to the "vote" through positive descriptions of the filter and a total lack of encouragement for debate or seeing other peoples view...gives you a totally scewed result compared to how it would be if there was a more balanced approach. Almost every referendum ... in the history of forever ... poses one question and provides two options: yes or no. This was poorly executed and it wasn't a referendum. --Shabidoo 05:12, 5 September 2011 (UTC)[reply]
Llywrch and others keep assuming that this filter will be terribly unpopular with the donors. Donor reactions are hard to predict, but I don't believe that's going to prove true. In fact, I recommend that the Fundraising committee test an ad that describes the personal, opt-in filter as a project they need funding for. It might attract money from people who currently don't donate. WhatamIdoing 15:29, 6 September 2011 (UTC)[reply]

Lacking important information[edit]

There doesn't seem to be a single sentence about how the respondent would state agreement/disagreement. Does 1 signify complete agreement or complete disagreement? There should also be a link to a deactivated survey or screenshot of the survey so that the reader can see how the questions have been posed visually. Dittaeva 14:53, 5 September 2011 (UTC)[reply]

Excellent point, thank you. I'll make sure the committee discusses this. Philippe (WMF) 18:56, 5 September 2011 (UTC)[reply]

For me the most important missing information is a statistic per language/wiki, and for overall attendance. One question was if the filter should be cultural neutral. Many found this important. So i think it is important know the opinions per project and the the overall distribution of the votes as well. For example: If 80% of the votes came from EN than it wouldn't be representative for all the other projects. Somehow this should be the case if i look at the current poll at the German Wikipedia with an average 80-85% against this filter. --Niabot 15:13, 5 September 2011 (UTC)[reply]

In response to the cultural point, I would hypothesise that there are larger number of Muslims on en.wiki than just about anywhere else, even though the proportion is relatively low. In my view the only reason a filter is necessary on a sitewide basis is to deal with the Muhammad problem; individual Wikis should have complete autonomy on filtering beyond that. 82.13.161.114 14:20, 6 September 2011 (UTC)[reply]

Next steps and conclusions[edit]

For a list of specific problems to address, and suggestions for moving forward, see the next steps and analysis page.

Conclusion? Talk about "burying the lead"![edit]

What's the gist of it all? In puzzlement, I am your friend, GeorgeLouis 11:14, 4 September 2011 (UTC)[reply]

Agreed, this page is awful. When a normal person clicks "View the results," he wants to see what will result from this vote, not "the median was 7 and the committee was pleased with the use of the text field." Townlake 06:25, 5 September 2011 (UTC)[reply]
I also agree. I barely passed statistics in college so the page is pretty meaningless to me. I'm also rather surprised at the use of fancy quotation marks since they're a violation of the Manual of Style. --Kitsunegami 07:16, 5 September 2011 (UTC)[reply]
This is Meta, not the English Wikipedia. The rules are completely different here. Meta's Manual of Style doesn't say one word about quotation marks. WhatamIdoing 17:39, 5 September 2011 (UTC)[reply]
What use of fancy quotation marks? Philippe (WMF) 19:01, 5 September 2011 (UTC) (struck out, found it. Appears to be a relic of the program used to write this. I'll make some changes later, but that's stylistic and I'd rather focus on getting more comments rated first).[reply]
Amen! Conclusion section? Aaadddaaammm 08:37, 5 September 2011 (UTC)[reply]
The committee is not charged with making the conclusions: that's the Board of Trustees' job. The committee is charged with execution of the voting, which it did. Philippe (WMF) 18:47, 5 September 2011 (UTC)[reply]

A filter is not a no-brainer[edit]

Let's give the Foundation and the committee who made this survey every benefit of the doubt. Let's presume they truly and sincerely thought that the idea of this image filter was entirely uncontroversial, and that it would have near-universal support once the details are worked out. That the survey was so rife with methodological problems that – unavoidably – biased any result strongly towards something that looks superficially supportive of the feature yet still renders a highly mitigated result should signal an immediate "all stop" to forging ahead blindly with this idea.

What I'd expect now from the committee/WMF is an acknowledgement that the image filter is nowhere near the no-brainer they imagined it to be, and a commitment to not do any further work towards implementation until a real community discussion has taken place. Further, at least some signal that they even allow for the possibility that this may not be a workable idea in the first place. — Coren (talk) / (en-wiki) 13:29, 5 September 2011 (UTC)[reply]

    • What do you mean by a "real" community decision, Coren? A poll with a turnout of over 24000 members voting is hardly a trial run, yeesh. The signal is quite clear, don't move ahead with the idea. People already have the option to block images in their browsers - if they don't want to see them, they don't have to. Why should we be adding another layer to stuff which already exists? BarkingFish 18:35, 7 September 2011 (UTC)[reply]
The feature's implementation as proposed is clearly itself controversial, which I don't think was the intent. Better discussion is needed. (Obviously the topic is emotional, but much concern expressed in the poll was on implementation details of such a feature, not the fact of it.) It is possible, as you suggest, that the very idea of offering visitors extra control over how and when images are displayed is not workable -- or not workable by us at present -- or workable but too complicated to be worth the time investment, given the current affordances of MediaWiki.
I haven't seen any alternative proposals to the early example drafted on mediawiki.org, so it is hard to say. I think it likely that there are ways to empower readers in this manner that are not too controversial or complex. SJ talk | translate   22:53, 5 September 2011 (UTC)[reply]
Thanks for the "blue response", SJ. I'm heartened to see that our comments are being read. I do think that the very idea of image filtering as propose is fraught with insurmountable problems, but I'm willing to be convinced otherwise.

I can identify four specific problems with this idea, and I would like being able to express them in a constructive setting rather than what felt as flailing in a crowd; what are your (collective) plans for examining this in detail? I made the error of not participating earlier, and I intend to not repeat the same mistake. — Coren (talk) / (en-wiki) 00:47, 6 September 2011 (UTC)[reply]

There's now a page for consolidating issues and suggestions here: Image filter referendum/Next steps/en. Please add your four problems there, and discuss related issues noted by others and/or ways you could see them being resolved. SJ talk | translate   04:28, 6 September 2011 (UTC)[reply]
That would be a reasonable reaction if the Foundation really was not aware of the implications of their decision. -- Carbidfischer 13:50, 5 September 2011 (UTC)[reply]
We might be getting a trifle harsh. If you're going to give the committee "every benefit of the doubt", you must assume that neither they nor whoever came up with this filtering idea in the first place considered the possibility that libraries, internet cafes, churches, and other community portals could use the filter to censor WP to their own arbitrary standards. Or perhaps they did, which would explain the question about allowing anyone to filter content vs. only logged-in users.
The next question is whether restricting filtering to logged-in users would solve that issue, or whether a library or cafe or church could simply log in and hack away, with their users none the wiser -- i.e. whether anyone wishing to view WP unfiltered would be able to do so, even at a community portal; and if not, whether there would be some clear indication that what they were viewing was not the unexpurgated WP.
I too would like some indication that the committee understands that there are previously-unpondered problems that need to be worked out before (or even if) this plan goes forward. Cheers, DoctorJoeE 14:51, 5 September 2011 (UTC)[reply]
You seem to believe the committee will be executing the filter. That's not the case. The Board of Trustees and the WMF are the parties charged with that. The committee's work is done when the final report is produced (which will happen after the evaluation of the comments and analysis has completed). Philippe (WMF) 18:54, 5 September 2011 (UTC)[reply]
Thank you for clarifying this, Philippe. The specific role of your committee was, indeed, not entirely clear. — Coren (talk) / (en-wiki) 20:08, 5 September 2011 (UTC)[reply]
Maybe I can add a little additional information here. Coren, it's not the case that the Wikimedia Foundation felt the image hiding feature would be non-controversial: quite the opposite. The Board of Trustees wrestled over this issue for months on its mailing list and at in-person meetings: it exchanged literally hundreds of e-mails on this topic, as well as having phone calls, face-to-face meetings and so forth. It's an extremely difficult topic: it's controversial and serious and emotionally hot.
Ultimately, the Board of Trustees passed the resolution that Philippe and I think also Chaddy have posted elsewhere on this page, requesting that I develop and implement the feature. I then decided to stage the referendum in order to find out more about editors' opinions. There are pros and cons to the idea of having a referendum, but on balance I felt like it was a reasonably good way to do further consultation with the community in addition to the extensive discussions that had already been held on-wiki. I was interested in what numerically-quantifiable results might look like, and I was also interested to see to what extent the results of a simple poll-type set of questions might or might not be different from the tenor and tone of the on-wiki discussions. (I didn't have preconceived ideas about that: I was just really extremely curious.) So we created the referendum.
How we did that: Philippe, who works for the Wikimedia Foundation and has experience facilitating the Board elections process, organized a committee to help run the process and analysis. The people on the committee in my understanding are disinterested and process-focused, and are essentially doing facilitation work: I don't know what their personal opinions are about the feature, and their personal opinions are intended not to factor into this. (And by the way, I am grateful to them for helping do this: it's hard work, and I'm sure sometimes can feel a bit thankless.) The actual substance of the referendum --the scope, questions, etc.-- was hashed through mainly by me, Phoebe and SJ in a couple of long talking sessions in Boston at the Wikipedia in Higher Ed summit: it was not developed by the committee. I expect that if we do something like this again we will approach the questionnaire aspect quite differently, based on our own views of how it's played out as well as the views expressed here, and the views that I expect will be expressed by the committee in its post-mortem -- there are some obvious improvements that should get made. But I don't feel too terrible about that: I think despite the flaws, the information that's resulted from the process will be useful. And it's not our only mechanism for consultation. And now we know more about how to run a referendum well, than we did before.
So where are we at right now? The committee is continuing with its analysis. People are kicking around the results. I think it's too early to say at this point what influence the results will have. I have my own preliminary thoughts, but I will hang onto them while more information comes in: I'm interested to see what additional analysis will tell us, and I'm interested to hear more of the conversation here. My purpose in speaking here right now was just to aim to clear up what looked like a little confusion about the process, and particularly about the role of the referendum committee. Thanks. Sue Gardner 01:56, 6 September 2011 (UTC)[reply]
If you want input from the community then you should not ask biased questions. There was no way to really express a simple "No, we don't want/need this". Now you tell us, that the feature is "not ... non-controversial". But why was no word to be found inside the opening page that it IS controversial? I only see a rough description and what it is intended to do. But not a single comment on possible side effects and that it runs against the primary rule of NPOV.
It is also obvious that the questions will not lead to any real conclusion as stated under A view from someone who does this for a living. Overall it looks more like a farce then a real intent to improve the project. That is my opinion on this topic.
At the end one personal question: Did you ever read the Library Bill of Rights and especially the subsection Labeling and Rating Systems? If you did, what is your personal opinion on this rules in context of Wikimedia and it's projects? --Niabot 17:04, 6 September 2011 (UTC)[reply]
Hi Niabot. And sure, I can take a crack at answering that.
I think I read Phoebe say somewhere the other day (foundation-l?) that there's a significant difference between a library and Wikipedia --- which is a statement I agree with. A library (as the ALA's Labelling and Rating Systems article explicitly states) is not endorsing the material in its collection. We, though, are. As I believe Phoebe said, Wikimedians are editors, not curators. Every day Wikimedia editors exercise editorial judgment over the materials in the projects ---is it neutral, is it notable, is it understandable, is it informative--- and in exercising that judgement throughout the editorial process, editors are shaping the material, not merely collecting or cataloguing it. So, we are endorsing the material we present. It's our material, made by us. That's an important distinction in my view.
So. I've been really interested in the positions taken by the ALA and other library associations, and we researched their views and spoke with librarians as we worked through the controversial content project. (I want to specify too: we did not just speak with Americans. I don't have the data at hand, but the project had a significant international component, with the research component including studies from around the world. And I know that Robert Harris spoke at length with several Canadian librarians.) We have a lot to learn from libraries, because our missions and goals are similar, and they have centuries of experience that we don't yet have. But having said that, I do not think that the mission and goals and practices of libraries and the Wikimedia projects are identical, and so I do not think that we should necessarily follow exactly their path.
Specifically WRT labelling and rating systems: The ALA distinguishes between viewpoint-neutral directional aids, and prejudicial labels that are used to restrict or forbid access. IMO the system that is being imagined for the Wikimedia projects is viewpoint-neutral and informational in intent. The Wikimedia movement is not intending to express a judgement about whether the material it's providing is offensive or should not be viewed by particular individuals or groups: we are aiming to provide information so that people can make their own individual decisions about what they choose to see.
The ALA also says this: “Directional aids can have the effect of prejudicial labels when their implementation becomes proscriptive rather than descriptive. When directional aids are used to forbid access or to suggest moral or doctrinal endorsement, the effect is the same as prejudicial labeling.” And I can easily imagine reasonable people interpreting that to mean that a NSFW or similar label on Wikimedia project material, is prejudicial. But I would disagree with that interpretation.
First off, it's important to note that the image hiding feature isn't intended to forbid or restrict access to anything. This is a feature which is entirely opt-in, and the effect of which can be undone at any time with a single mouse click: one of its guiding principles is that no content would be denied to any user. So, it would not forbid access to any material.
Therefore, I think the key question is this: does enabling content to be screened from view (“shuttered,” as I think it is described somewhere) constitute “moral or doctrinal endorsement.” If we were to enable all users the ability to screen from view all images, then any question of moral or doctrinal endorsement disappears. However, if we restrict ourselves to a small number of categories because of technical limitations or (more likely in my view) usability considerations, then the question of moral or doctrinal endorsement becomes a real one. In that event, IMO, the way for us to avoid doctrinal endorsement is to ask our readers to decide for themselves what types of images they would like to have the ability to shutter. This would be pretty elegant, in my view, because this feature is intended to be for readers' benefit. And therefore: the better we enable readers to make their own decisions, including contributing to the actual architecture of the feature, the better able we'll be to meet their needs and avoid clunky guesses and missteps.
This kind of process would also have the benefit of being broadly-based rather than limited to a small group, which is important because it will work to mitigate against undue influence by people or groups with specific agendas. Imagine this as being somewhat like our system for determining notability. To determine notability, we go outside our own borders and survey the world to see what judgements it's placed on certain topics. We would, in effect, be doing something similar with this. Do we make “doctrinal endorsements” when we decide that some topics are worthy of a Wikipedia article and others aren't? I don’t think so – we make practical decisions, based on data and information created by large numbers of people outside of our projects (a form of crowdsourcing, although I dislike that word). I would argue that we could categorize images for purposes of this feature in the same way, and with equivalent elegance and neutrality.
It's a long answer: I'm sorry – but it's an important question and I'm glad you asked it. I should note as well that this is my personal opinion of the issue. Board members and other staff members will likely have somewhat –and perhaps significantly-- different opinions from mine as expressed here. Sue Gardner 03:28, 7 September 2011 (UTC)[reply]
You said that in your opinion "the system that is being imagined for the Wikimedia projects is viewpoint-neutral and informational in intent." I believe that this is your intention. But as the mockups already have shown, there is very less room for viewpoint neutrality in such an approach. It always leaves simple questions: At which point does an image depict violence? What is to be considered sexual content? At which level of clothing is someone depicted nude? ... The answers would be guidelines. Essentially every guideline is a question for itself. This are "yes/no" questions without any room for compromise (hidden or not) or a balance between positions.
All this questions will need to be answered. But can they be answered in a neutral way? We had many examples for images inside the discussion where it wasn't clear and opinions strongly divided. (If you request, I could present you some images and you could let some people play the judge/censor/whatever) Now we may ask how we can achieve something like this without being prejudicial. In my opinion this is not possible and even worse: We have no sources for images. In articles we cite the opinions and choose them by relevance. Deciding about pictures is something very different, since this is the POV of the man or woman that puts them into one of the categories.
In a library the user decides what to look at, ideally without any possibility to sort out controversial content beforehand. I for myself compare Wikipedia, especially Commons, to a library. A library gathers books with images. We gather articles. We may be editorial for the articles itself, but from the point of the reader we are essentially a library that hosts millions of articles. Thats why I think that we should go along the same rules and should not label our content as controversial or not.
Inside the mail conversation there was one question that got my attention:
6. Do libraries have a policy about demarcating volumes within the library itself -- in other words, some sort of system that identifies certain volumes as potentially objectionable, or safe?
No – this practice is definitely frowned on and should not be happening. This is counter to our principles, philosophies and practices. The idea of “rating” is abhorrent to many librarians – and really who is to decide? Rating and filtering only create a sense of false security. We work from the understanding and position that people should be able to make their own decisions; parents should monitor children’s reading. If this practice does take place in a library, it would be because the staff are misinformed, misguided and sadly unaware of IF principles; or in the case of school libraries, perhaps forced to do so by administrators who have very little understanding of IF principles. On the point of “safe”, see the Bernier article from the reference list below. [7]
But thats exactly what our filter is supposed to be doing. We would identify images as potentially objectionable. Otherwise we could use our current categorization system and would not need additional categories (no matter how they are implemented) that serve this purpose. The labeling alone would discriminate some of our content. That we provide the option to filter objectionable content (we choose the categories) already indicates that we don't consider our content as equal in terms of neutrality - going strictly against our main goals of NPOV. That we aren't always neutral is a fact, but we should aim for that goal. The filter project itself works in the opposite direction.
You stated:
"First off, it's important to note that the image hiding feature isn't intended to forbid or restrict access to anything. This is a feature which is entirely opt-in, and the effect of which can be undone at any time with a single mouse click: one of its guiding principles is that no content would be denied to any user. So, it would not forbid access to any material."
I think you miss the point that labeling the content as controversial is already the problem (see above). If we give the user the choice to hide images with controversial content he actually doesn't decide for himself, because it wasn't him who decided to hide which image or why not to hide it. The decision for a particular image was made by others. For example: If you checked "sexual" as an filter option. Would you open an image hidden from your view in an article about a topic that you just found? Would you have more doubts then before? Would this not discriminate a valuable addition to the article? Would not some users (especially children) learn that this must be something bad? Is sexuality and the knowledge about it something bad?
Now i have offered some concerns about the general problem with neutrality. All comes down to some very basic problems. 1. Who decides? 2. Is the decision neutral or would it be based only on personal preference? 3. Do we decide for the readers or do they decide? (Under the premise that we label the content) 4. Who will do all the new work? (Time which could be invested in improvement of the content itself)
Two ideas that i would find workable:
  • A simple feature to easily hide all images as default with the option to make them visible at any point. This has the following advantages: No one would question the neutrality of the approach. It's very easy to implement. I doesn't cause any extra work for the authors (or admins because of editwars about categorization). A user can feel absolutely safe and always makes his own decision what to view and what not.
  • If children are the concern (read a lot about them in the reports), why not create 1 or 2 extra Wikis with content aimed at children? They would have it easier to understand the topics. Content not suitable for children could be left out, and so on.
I would bet, if you would give the voters the choice between this three options (filter, hide all, extra wiki), the current solution wouldn't get the majority. --Niabot 09:22, 7 September 2011 (UTC)[reply]
That all can't be meant serious anymore. "It's our material, made by us. That's an important distinction in my view." says Sue Gardner. I work since 2005 for Wikipedia as a contributer. In all this time it was perfectly clear that it is not our material, that we collect the knowledge of the world (like a library collects books) and that we don't produce own knowledge. And i know that this position the big majority of conributers shares with me. The question is: Are we contributers wrong or and that would really piss me off are those few people at the head of Wikimedia not knowing anymore what Wikipedia is, what is going on here, what is done here? In the last few weeks i believe this more and more. And really it pissed me off too that this poll not was made "if the filter should come at all", but like one of the Wikimedia workers could tell a german newspaper called Die Zeit that it was made specifically this way taht we could have a feeling "yes, we got asked", but it was made that we had not to decide if it comes or not, but just to make us not this angry when it comes, because we got asked. Hell, my ass!!! So dear Sue, please make a reality check what Wikipedia is and wants. And have the greatness to admit taht you are wrong on those points and that you have confused Wikipedia with whatever, with soemthing that creates a product out of its own (what we don't) and not seen it anymore just as the collection of knowledge from somewhere else (what we are).
Sincerly, a really pissed off Julius1990 11:09, 7 September 2011 (UTC)[reply]
I've logged out of my contractor account (User:Mdennis (WMF), mentioned because I don't want anybody to think I'm trying to hide who I am) and in again under my volunteer account because I just want there to be no doubt that I am speaking as a long-time, prolific volunteer contributor to Wikipedia, not as the liaison. This is entirely my own opinion, based on my volunteer work which dwarfs by a rather large magnitude the months I've held this temporary contract.
Whether or not "It's our material, made by us" may depend on how you're looking at it. The articles we make are "our material, made by us". We make countless editorial decisions daily whether something is appropriate for our readers. We decide which facts deserve emphasis and which images are appropriate for display in what articles. The article Budapest, for example, is full of images that have been hand-picked for display there by involved editors. In other words, they made that display. The only way it would not be true that our articles are "our material, made by us" is if we had some kind of image randomizer that selected from among available images to display...in which case, of course, we would not be choosing the images and they would be statistically unlikely at any given moment to have anything to do with Budapest. Whether we approve of the filter or not, it is intended for readers of those articles and does not impact the collection and storage of materials at Commons.
In terms of the general tone here, the Wikimedia Foundation was specifically asked by the Board of Trustees here "to develop and implement a personal image hiding feature that will enable readers to easily hide images hosted on the projects that they do not wish to view, either when first viewing the image or ahead of time through preference settings" that would "visible, clear and usable on all Wikimedia projects for both logged-in and logged-out readers." They were asked to consult with the community about developing and implementing this. They weren't asked to decide if they wanted to develop and implement it, or to find out if the community wanted it developed and implemented. Sue is responding to the request of the Board that we, as volunteers, elected. It seems a bit unfair of us to elect a board to manage the Foundation and then to blame her for being managed. :/ --Moonriddengirl 14:28, 7 September 2011 (UTC)[reply]
Your words pass the responsibility back to Board of Trustees. Thats why you should state them the current situation. I don't know if the board ever discussed about possible side effects. But as we can see inside the German community a big majority is against the implementation of this filter (average 85%).[8] So it is a decision by the board without any influence by the community itself. Since we are not in real life, we don't have any means to protest against this not trustful proposal by the trustees. But i must wonder why it was even considered, since not long time ago another poll made clear, that this should not happen, not even on the main page (average 94,7%, 233:13). [9] That many authors are angry with the result (whether or not the WMF or BoT are responsible), should be clear at this point.
Back to the wording of "our material, made by us": Yes we make the material (by collecting facts from sources) and yes we decide what kind of material is suitable (best illustrative) for the article. But this decision is (or should) never be affected by means that an illustration is controversial. All that matters is, if the image is the best illustration for the topic. Thats how images are picked. It's not about personal taste ("Hey, that image looks nice. Ok, it doesn't show how bloody it was. But it looks so pretty with that sunset. ..."), its about the illustration of the subject ("People starving to death, because of shortage"). The world is no pink easter bunny.
The filter servers to hide content by personal taste ("this disgusts me, i don't care about the truth"). This has nothing to do with the encyclopedic goal of gathering and distributing of knowledge. It only serves to discriminate content, which is against any scientific/educational/didactic principles. You should keep this in mind. --Niabot 15:21, 7 September 2011 (UTC)[reply]
Maybe Sue was afraid to say what the thinks or maybe she didn’t know any better. The more interesting point is: How did the board get to this 10:0 decision and why did they think the community would accept their decision, without being asked? -- Carbidfischer 14:54, 7 September 2011 (UTC)[reply]
It's nice taht you use the word "storage", what for me insists that you also see that wikpedia, commons and the other projects are likely to compare to libraries and that their view on their mission statements could be easily an example for us. We provide knowledge and images, yes, but the wikipedia does not anything different than a library. We just have putted together more close the infrastructure and the people providing it.
The foundation in the last month was more and more focusing on the problem of loosing contributers. Why does the foundation not respect them. That is a fundamental question about the direction of the wikipedia projects. We contributers make the wikipedia big and like it is, we have the right to get asked if a form of censorship at all should be allowed. Philippe Baudette says ""Die Frage ist nicht, ob der Filter eingeführt werden soll oder nicht. Der Stiftungsrat hat dies ja bereits beschlossen". ("The question is not, if teh filter comes or not. The Board already decided.") (http://www.zeit.de/digital/internet/2011-08/wikipedia-filter-referendum/komplettansicht) So what was the intention of this poll? It was to give us a feeling of deciding, without having really to decide ... That let me vomit.
When the board thinks it can decide such fundamental question without asking us, there is something wrong in the system. I know why i didn't vote in the last election, because i trusted noone of the candidates. And i got proofed right ... And i'm happy that noone of the members can think he/she got the mandate by me to decide this without the community. The tower of ivory problem maybe. That the really main contributers of the German Wikipedia are voting in - the here already mentioned - Meinungsbild the way they do, should let the foundation and the board think. Without the communities, wikipedia is nothing. It wouldn't even exist. It is a matter of respect ... and that's why not just me is this angry. Julius1990 14:59, 7 September 2011 (UTC)[reply]

(to the left) Julius 1990 wrote: "Wikipedia does not anything different than a library."
What? Writing encylopedic articles is a complex editorial process quite different from collecting and cataloging books. Every article is or should be result of a meticulous reflection on what is important for the reader, what has to be stressed and what has to be omitted. You may want to describe that as a form of censorship - since this term became deliberatly used especially in the german wikipedia and it's "Meinungsbild". We, the authors and photographers of this project, decide every day, what readers should read and what they do not have to read. This we do with every single edit in every single article. So we don't simply collect "the knwoledge" (TM) but process and reprocess this knowledge in a kind I would readily call creative. It is our material made by us" as Sue called it and not a mere compost heap of rotting facts as you seem to see it. Adornix 15:24, 7 September 2011 (UTC)[reply]

We decide by one primary principle: What is the best way to describe a topic/term as it is?
An illustration is not chosen by "how nice it looks". It's chosen, because it is the best available illustration to illustrate the facts. We don't make the facts, we gather and represent them. All we argue about is the presentation of the facts. Thats our job. If something is controversial or not has no meaning (at least ideally). --Niabot 15:35, 7 September 2011 (UTC)[reply]
Hm, I don't really see the contradiction :-) Oh stop: your description of what we do is far too simple. Of course we care if the article text is well written and of course we care if the image chosen is well photographed or drawn. And if we have two or more images with the same amount of information we usually chose the more appealing one, don't we? And of course is your simple looking "we gather and represent the facts" much too simplifying since this representing is much more a question of good and responsible choice as you and julius like to describe it. Adornix 16:19, 7 September 2011 (UTC)[reply]
If you do more than that, than you do it wrong. If you have 3 different pictures of the same subject then you choose the image with the better quality or which depicts the subject better. Some personal taste may be involved, but this isn't a problem in this case. All three images are equally informative. Making the preference to show an image, that is less illustrative, just because the more illustrative alternative could be controversial, is an issue and against the principles. You seam not to understand the difference between choosing for the goal of knowledge/education and to chose by pleasure for the eyes. A perfect image about a non controversial topic can easily meet both criteria. A picture about a controversial topic, which has high educational value, usually does not please the eyes of all viewers. That is a fact. But never the less, from an encyclopedic view, both images have the same value and both should be represented under the same light.
I conclude: We (should) decide about illustrative quality. This filter would decide about personal taste. But personal preference or taste has no room inside an encyclopedia. --Niabot 16:37, 7 September 2011 (UTC)[reply]
What a conclusion :-) And it is more or less right - if partially unrealistic. But the filter is not about changing the content of Wikipedia at all. It is about personal choice of the reader. I think this point is constantly confused in this debate. Adornix 16:43, 7 September 2011 (UTC)[reply]
No, it's not. If we on this are right, and if nothing for us as contributers would change, then it is the question what the point of all this is. It is: what do we show the reader, to what give we him access? And on this we can easily compare to the guidelines of the library associations and there is no need for this filter to see ...
And then there is the point, why on such afundamental point the people who are making teh wikipedia (and that is neither the Board nor teh Foundation) don't get asked "if", but there is made such a poll just for giving teh feeling we got asked if it's in teh end not true. It's awefully fake to give with a poll the feeling to decide, while the decision already is set up ... It is more than insensitive by the Board and Foundation to do it this way. And it is teh wrong way to go about a fundamental question. Julius1990 16:50, 7 September 2011 (UTC)[reply]
Niabot, it might be "wrong" to consider the aesthetics of an image, but that is how it is actually done in practice. For example, there's yet another discussion at en:Talk:Pregnancy over which image to put at the top of the article. Many editors want to follow the example of the German Wikipedia's article at de:Schwangerschaft, but they are loudly opposed by some people who want to have a "beautiful", "precious", "warm" photograph at the start of the article, and who demand the complete exclusion of any fully dressed pregnant woman as insufficiently pretty. (Nobody wants to remove any image. The question is only adding one image of a fully clothed woman and moving one image to the same, highly relevant place that the German Wikipedia puts it.) If everyone agreed that prettiness was unimportant, then we wouldn't be hearing these arguments about beauty. WhatamIdoing 17:01, 7 September 2011 (UTC)[reply]
That fact, that there is such a difference between projects, is already a good reason to not implement it the way it is. While EN usually is very critical about such content (making it controversial), DE accepts it as a usual depiction. At first we have less problems with the choice of images then EN. The audience does not complain as much. When was the last time in DE someone wondered about the illustration of hentai? 1 year ago? How frequently this happens at EN? There is a huge gap. But considering the daily count of readers there is only a very little group that complains.
Now we can go on to the real problem. The filter is designed to be global and it's goal is to meet the conventions of all nationalities. An impossible goal right from the start, as your little biology example already points out. --Niabot 17:18, 7 September 2011 (UTC)[reply]
Oh yes, it is :-) This filter as Sue explained it will change nothing for the contributors. It is a feature for readers. And the filter is not meant to and will not work as prevention of access. Sue made it totally clear, that noone will be prevented of access to any image. If the reader wants to see a filtered image they will have the simple option to click a link or button. Of course this is a technical question too and I understand concerns that this feature may provide means to misuse or even censorship!
Your second point is more serious of course. I think the WMF made a terrible mistake. Adornix 17:05, 7 September 2011 (UTC)[reply]
It will actually change a lot. Every image has to be reviewed, it has to be discussed and unavoidably will result in edit wars. At the end it will bind the time of many admins, that have less time for the needs of the authors. No change comes without side effects.
Yes this feature is very prone to be misused. It attracts vandals and censors, it's internal data can be used to limit access through filters of third parties (The button will not work anymore) and it opens the door to censorship. A step we never wanted to go in the first place. --Niabot 17:26, 7 September 2011 (UTC)[reply]

The next step... already decided?[edit]

...has been already decided: "and design/implementation suggestions will be anonymized and passed on to the design team." They could have spared us the voting to this pseudo-referendum and all the rest. They have already decided and what a big part of the community thinks counts nothing. --Dia^ 21:03, 5 September 2011 (UTC)[reply]

That is complementary to, not in place of, an overall review of the proposed feature. Further high-level discussion is already taking place about what this feature hopes to accomplish and whether/how that could be reasonably implemented. SJ talk | translate   23:32, 5 September 2011 (UTC)[reply]

Lack of any request for community input[edit]

I'm going to make a very specific assertion/accusation here. There was zero interest in involving the online community in the process surrounding the decision about which questions were going to be asked.

In particular, we've seen evidence about the lack of involvement on this wiki page, as well as others, where suggestions for modifications to the questions were clearly made, with rationales given for why the changes should be made, but these suggestions were simply not acted on or responded to in a timely manner. Similar suggestions have been made via the mailing lists, and I'm sure they have also been made by personal email. In addition, there have been no requests for community input - be it via the mailing lists, on-wiki, or by any other means - for suggested modifications to the questions. From the date that the questions were posed to the date when the referendum started there were no modifications made to the questions, despite the questions asked during that period and those raised while the poll was running.

To clarify my position: I don't give a fig about whether this filter is enabled or not. I can perceive the positives and negatives on both sides of the issue, and right now I don't think either side has the winning argument. I do care a huge amount about whether the actions are supported by, and have involved, both the editing community and the readers of the projects. My current viewpoint is that this so-called 'referendum' has been pushed through without community input, without the input of either community- or academic-recognised experts on the topic, has been asked to the wrong audience (i.e. the editors rather than the readers) in a biased fashion (i.e. it's focused on current editors, rather than the world-wide population) and lacks publicly available metrics that would be useful in counter-acting the systemic bias that is inherently present within its results.

As much as I hate to say this (seriously, as I would love to see a solid, community-supported way forward here), I think this has been a wasted opportunity that has taught us nothing more than we really need to figure out a better way to organise and run a Wikimedia referendum that inherently involves as much community input as possible. I have the utmost respect for those involved in organising this referendum, which makes it even harder for me to post this, but I really think these issues need to be tackled before either this filter or future referendums take place. Mike Peel 23:39, 10 September 2011 (UTC)[reply]

Basic background info[edit]

It would be helpful if a short paragraph on the background was added. For example, how would the proposed image filter system operate? Some text from the main article on the referendum might be helpful (from: http://meta.wikimedia.org/wiki/Image_filter_referendum/en), such as the text below:

Background description[edit]

The Board of Trustees has directed the Wikimedia Foundation to develop and implement a personal image hiding feature.

Its purpose is to enable readers to easily hide images on the Wikimedia projects that they do not wish to view, either when first viewing the image or ahead of time through individual preference settings. The feature is intended to benefit readers by offering them more choice, and to that end it will be made as user-friendly and simple as possible. We will also make it as easy as possible for editors to support.

The feature will be developed for, and implemented on, all projects. It will not permanently remove any images: it will only hide them from view on request. For its development, we have created a number of guiding principles, but trade-offs will need to be made throughout the development process. To aid the developers in making those trade-offs, we are asking you to help us assess the importance of each by taking part in this referendum. [edit] Why is this important?

In the 2010 Harris report, two of the recommendations (7 and 9) were to create a way for readers to hide images they did not want to see; and that there be an option for readers to hide all potentially controversial content.

There are several rationales for a feature like this. Images of sexuality and violence are necessary components of Wikimedia projects for them to fulfill their mandates to be open, free and educational. However, these images – of genital areas and sexual practices on the one hand, or mass graves and mutilated corpses on the other – will inevitably still have the power to disturb some viewers, especially if they are children, or if they are happened upon unintentionally. The point of the opt-in personal image hiding feature is to help alleviate that surprise and dismay, by making the images unavailable for viewing without a second command. Often, within the Wikimedia world, this is referred to as the principle of least astonishment, or least surprise.

On the other hand, we believe that this command should only delay, not prevent, the presentation of these images. Access to information on Wikimedia Foundation sites should be compromised only to the extent needed to satisfy our responsibilities to respect and serve all of our audiences. A shuttered rather than a completely hidden image satisfies those responsibilities.

Excellent point, thank you. I'll make sure the committee sees this. Philippe (WMF) 18:56, 5 September 2011 (UTC)[reply]


Other comments[edit]

Extend the filter to texts[edit]

Then we won't need to burn books in the future, we can filter them away and thus hide what we don't like, e.g. slavery, religious wars, etc. John.St 11:36, 5 September 2011 (UTC)[reply]

Think of the children! — Internoob (Wikt. | Talk | Cont.) 20:08, 5 September 2011 (UTC)[reply]
It's even worse than what you appear to believe. If only the filter were planned to give each and anyone the ability to filter exactly the images they want filtered (that would be a truly "personal" image filter btw). -- But that is not what is going to happen. This thing is basically a decency filter, with some additional filtering categories (violence, and then "8-13" unspecified others) thrown in for greater mass appeal. It was never planned as anything else than a nudity filter, everything beyond that is just window-dressing. The amount of intellectual dishonesty displayed by the WMF is staggering. --87.78.237.242 21:04, 5 September 2011 (UTC)[reply]
This is false. Three specific types of images, representing three persistent complaints from users, have always been planned, not just nudity. Sexual images is one; the other two are violence and images of religious significance.
I will assume that you have made this false statement out of ignorance rather than telling a malicious lie, so let me suggest that you go read the Harris report, where the three types are described, so that you will not make the same mistake again. WhatamIdoing 15:34, 6 September 2011 (UTC)[reply]
No default option to hide war and weapons? OMFG... must be a idea of american people ^^ --Niabot 15:37, 6 September 2011 (UTC)[reply]
Like Sameboat, you're leaving me no choice but to regard all your contributions as the aggressive trolling they really are. Your strongly worded response is quite revealing though. --87.78.50.127 23:17, 6 September 2011 (UTC)[reply]

A view from someone who does this for a living[edit]

Hi. I'm a statistician. I worked in political polling for a few years and am now back in academia. I was asked to look at this by one of my students, who knows that I'm familiar with Wikipedia. I actually did vote while the referendum was out, and have a few comments.

  • Purpose
    It looks to me like the people who organized this referendum had an entirely different purpose for doing so than perhaps 90% of the voters - certainly than most of those who commented here and in other fora about it. The community was confused because they've never really been asked before if they thought this was a good idea, so some assumed that this was asking their opinion of the feature as a whole, some realized it wasn't, and some realized it wasn't and answered as though it was anyway. This came back to hurt you with:
  • Poor question choice
    The first question on this referendum can mean one of a few different things:
    It is important for the Wikimedia projects to offer this feature to readers.
    1. Do you support or do you oppose this feature?
    2. This feature will be implemented. Do you think implementing it is important or unimportant?
    3. Do you personally want this feature?
    4. Do you think readers want this feature?
    And while these are all good questions to ask, your results mean less than nothing if you ask all four questions at the same time. Further, your results would mean very little even if you asked all four of those questions - you need what are called 'cross-tabs' - the ability to ask, for any combination of questions A and B, "Of the people who selected choice 1 or 2 for question A, what were the results for question B". Since you can't do that, because you didn't ask whether people support this feature, you can't tell which people think "It is not important that the feature be usable by both logged-in and logged-out readers" because they don't think anonymous editors need it and which people think "It is not important that the feature be usable by both logged-in and logged-out readers" because they think the feature is silly to begin with.
    Here is an example of someone who had a problem with this:
    When I first encountered the survey, my sentiments were very similar to those of 98.230.51.109; I was disappointed that such a major decision had been unilaterally reached without seeking community consensus or even input. Moreover, I personally recall the thought process of "if I respond with 0/10 in a concrete question of importance to voice fundamental objection, will my other responses and text input be methodologically invalidated as partisan or irrelevant?" going through my mind, which indeed resulted in a considerable artificial deviation in my case. CMBJ 16:49, 4 September 2011 (UTC)
    How the questions relate together is also unclear. When you're polling you need to make sure that everyone's answering the same question, which means that "It is important that the feature be usable by both logged-in and logged-out readers" should be worded as "If this feature is implemented, it is important that the feature be usable by both logged-in and logged-out readers", so that those who oppose the feature in the first place know that their vote in favor of this question won't be construed as overall support for the feature. I suspect that is why you have significant votes of zero for "It is important that hiding be reversible" and "It is important that the feature allow readers to quickly and easily choose which types of images they want to hide (e.g., 5-10 categories), so that people could choose for example to hide sexual imagery but not violent imagery". For a real-world example of this type of question choice, see [10].
  • Poor answer choice
    And continuing on from the first part, you can't tell which people think "It is not important that the feature be usable by both logged-in and logged-out readers" and which think "It is important that the feature not be usable by both logged-in and logged-out readers". And it's because you ignored a very common and simple set of answer choices (Strongly oppose, strongly support, weakly oppose, weakly support, neutral, no answer). Asking people to rate their support on a scale of zero to ten might help you to analyze each individual user's relative preferences (how many people think it is more important that it be easy than that it be culturally neutral) and relative strength of preferences (of the people who think it is more important that it be easy than that it be culturally neutral, how strongly do they think that), it is ineffective at assessing the overall community support. People who feel strongly about the existence of an image filter will find that skew their preferences of the other questions. People who generally hold strong opinions hold more sway than those who hold an opinion but are wary of selecting the "0" or "10" "nuclear options". Compare with the "pain scale" - "on a scale of one to ten, how much does it hurt?". Answers to this question are useless because someone used to broken legs will rate his twisted ankle a three, while someone who rarely gets injured will rate their heartburn a 10. The heartburn isn't worse pain or a medically worse condition, but the "scale" has poorly-defined endpoints and undefined other points. [11]
  • Poor data analysis
    The available data is very limited, consisting primarily of the answers to each question with no crosstabs or other breakdowns. Perhaps that's because there were no crosstabs available due to missing questions, but you should really be providing the ability to filter results by user activity level {inactive, active, admin, developer}, origin wiki, and overall opinion on the image filter, as well as on each question you asked. You can do that with a simple webpage with a bit of javascript or you can do that by providing a csv or sqlite database containing all of the above information. But if you want your users to be able to analyze the community opinion on whether this feature should exist and if it does, what it should look like, then that's essential information.

So I hope I've explained why there are a few problems with the data presented here. Releasing the whole data set would be a good start, but unfortunately once you've asked the wrong question, you've got the wrong answer and the only way to get answers that reflect the community's opinion on the feature as a whole (and, if the feature is implemented, the opinion on certain subfeatures) is to try again with the right questions. 67.85.5.224 22:59, 5 September 2011 (UTC)[reply]

Thank you for this concise and thorough critique. What do you think about our ability to gauge the importance of a change to readers who do not edit, based on this data? Trying again may be valuable for a few reasons -- better questions and answers, better data gathering techniques, and a clarified and revised proposal based on feedback so far. SJ talk | translate   23:28, 5 September 2011 (UTC)[reply]
Thank you! This about settles the question of whether a new and proper referendum is required. The only thing left to do now is to get the WMF to acknowledge and respond to such rational and powerful analysis. Looking at the referendum main talk page, that will be the hardest part. Actually, since most of the board members are intellectually capable people, I can hardly imagine that they wouldn't have noticed any of the many fatal flaws in their exact methodology and their general approach which have already been pointed out. The one question to ask now is whether the board will actually go ahead with their plans despite being caught redhanded. --213.168.110.95 23:55, 5 September 2011 (UTC)[reply]
You are right that the people involved are all great and capable folks, but that doesn't mean that we have much expertise within the Foundation in running effective surveys. This may change as it becomes a regular practice. Among other things, we should be expanding and relying on the expertise of the research committee and other community groups throughout the lifecycle of such polls. Here the referendum committee helped a great deal, and included professional statisticians, but was not involved with the design of the poll itself (as I understand it); just communication, translation, technical aspects and reports. SJ talk | translate   03:52, 6 September 2011 (UTC)[reply]
And why is it, exactly, that the Foundation didn’t feel the need to ask anybody who knows how to poll? Knowing that you don’t have much expertise and then not asking people who have doesn’t look very capable to me. -- Carbidfischer 05:19, 6 September 2011 (UTC)[reply]
That I don't know. I assumed we would have that sort of implementation support. It may be wrong to say that the Foundation don't have such expertise on staff, as it engages in a lot of community research; but I don't believe that those people were involved in designing the language/form of this survey. SJ talk | translate  
But you were involved, if we are to believe Sue Gardner. Didn’t it occur to you, at any time, that it might be a good idea to ask anybody who actually knows a thing or two about polling? Wikimedia isn’t exactly some kind of little neighborhood association, it styles itself as a global movement, so it would sound reasonable to me to, like, at least try to pretend to be professional when it comes to matters as important as these. -- Carbidfischer 21:47, 6 September 2011 (UTC)[reply]
The errors extend far beyond the referendum itself, into core aspects of the filter version envisioned by the board. Will the wealth of input regarding concerns and possible alternatives provided at Talk:Image filter referendum/en be taken into account from here on out? --87.78.137.221 05:20, 6 September 2011 (UTC)[reply]
Yes, that was the primary intent of this poll. (Point of correction: the Board did not envision any particular version of a filter. The Board's resolution doesn't request a specific implementation -- simply a feature mentioning the value in empowering individual readers to choose whether or not to see an image.) SJ talk | translate   21:20, 6 September 2011 (UTC)[reply]
I was talking about the version presented in the referendum, whoever designed it. --87.78.50.127 23:04, 6 September 2011 (UTC)[reply]
You (as Board member) voted in favor of a resolution (passed unanimously) which asked for a "personal image hiding feature" to be "developed and implemented" that "will enable readers to easily hide images hosted on the projects that they do not wish to view, either when first viewing the image or ahead of time through preference settings." [12] In other words, you have agreed with the rest of the Board to make the development of a "personal image hiding feature" a priority software project for WMF. When you say that the Board was "simply [mentioning] the value in empowering individual readers to choose whether or not to see an image", that suggests to me a statement of values and proposed actions, like in the BLP resolution and the openness resolution. But the controversial content resolution goes beyond that by allocating WMF resources to a specific software project. That's certainly the prerogative of the Board, but I'd like to request clarity that the decision to build a software feature is a decision that has been made by you and the other Board members, well before the poll took place. At this point, the open question isn't whether a software feature will be built (unless the Board changes its mind on its earlier decision), but how. Correct?-Eloquence 05:28, 7 September 2011 (UTC)[reply]
I clarified my comment above. I mean that no particular implementation was indicated by that feature request. And a request to develop does not mean "at all costs". "Whether" and "how" are not entirely separable - a feature that does not significantly empower readers, or that empowers some while disempowering others, might not be worth the investment, or might be worse than the status quo. Similarly, a feature that requires tremendous technical effort and delays many other important projects might not be worth the opportunity cost. If those were the only two options, even those in favor of such a feature in theory might reconsider the merit of developing one in practice.
So taking concerns and alternative suggestions into account seems to me a necessary prerequisite to deciding how to proceed. SJ talk | translate   07:42, 7 September 2011 (UTC)[reply]
To put it short: I think that the people WMF are not stupid. But for me this referendum has two meanings. Either the WMF is willingly ignoring the community and tries to betray it, or I'm wrong. --Niabot 00:17, 6 September 2011 (UTC)[reply]
In this case, you are wrong. But I am glad that you are constantly sharing your ideas, suggestions, and reactions in public with everyone. The fact that you interpreted it this way is important. SJ talk | translate   04:28, 6 September 2011 (UTC)[reply]
If he isn’t wrong, then I’d like to hear an explanation for Wikimedia acting as if they were the first organisation to ever decide about anything. You don’t ask experts, you don’t communicate with your workforce, you try to gain information out of a poll created as if there were no experiences with polling at all ... Let me get this straight: You aren’t betraying us, you’re just unbelievably incompetent? -- Carbidfischer 21:47, 6 September 2011 (UTC)[reply]
Despite the complicated analysis above, for me there were two main mistakes. Firstly, calling it a "referendum" when it clearly wasn't, and failing to fix this despite a stream of people pointing out the mistake virtually from hour one. Secondly, the fudging of the central question of whether people wanted the feature at all with this use of the word "important". 86.160.211.238 14:08, 6 September 2011 (UTC)[reply]
I'm sad to see that our anonymous statistician complains about "poor data analysis" based in the incomplete data analysis. Perhaps his notion of analysis is so limited that it can always be done instantly? Perhaps he does not know what the words "preliminary" and "interim" mean? Perhaps he didn't think it necessary to read the first sentence of this preliminary report?
IMO it is impossible to judge the quality of the data analysis until the final report is published. The final report is described in the very first sentence on this page as something the committee "will publish a full report once final analysis is done", i.e., what you're looking at is not the full or final report. WhatamIdoing 15:40, 6 September 2011 (UTC)[reply]
You can judge it from the questions alone that the poll was not well prepared. Of course you (WhatamIdoing) would be pleased to hear that the result is postive. You have promoted this propaganda all the time. Now someone with expertice says that the poll indeed asked the wrong questions und that the results are pretty much useless. Of course you can't agree and must attack the writer. Nice job as usual. LOL --Niabot 15:46, 6 September 2011 (UTC)[reply]
I agree wholeheartedly. WhatamIdoing, all you have contributed to the discussion over the past few weeks is fallacies (strawmen, red herrings, non sequiturs galore) and an obstinate refusal to even acknowledge the points made by people who do not share your pro-filter fervor.
Interestingly, you (WhatamIdoing) never even admitted that you are not merely a supporter of an image filter, but particularly of the utterly unworkable approach of a filter based on filter categories. It would be best if you abstained from the discussion from now on, seeing as you are not able and/or willing to contribute in a constructive manner. --87.78.50.127 23:13, 6 September 2011 (UTC)[reply]

Related comments[edit]

I think this quibbling is a little ridiculous. The interpretation of the vote, regardless of its defects, is clear enough. Only a minority, albeit it a sizable minority, of Wikipedians who voted totally opposed the filter. I am very disappointed it was not more: I hoped for an overall result like the deWP poll. I don't think any of the objections to the poll would change that basic fact. But I think this minority has made the point that this is something we care about passionately, while most of the people who do support it are not quite so passionate about it. I don't think any further analysis of the poll would change that either, & I thank Risker and SJ for making it clear that these two sides of it are both generally understood. DGG 03:42, 7 September 2011 (UTC)[reply]
Only a minority, albeit it a sizable minority, of Wikipedians who voted totally opposed the filter -- No, that is precisely not something that can be read from the poll. A new, proper referendum is required to actually gauge the community's opinion on that and other crucial issues with the filter (such as: "should it be based on special filter categories?"). As it stands, the poll has achieved nothing beyond experience value. --87.78.20.214 06:19, 7 September 2011 (UTC)[reply]
I think DGG is right. Reality is doubtless disappointing it is to the filter's opponents, but the reality is that most of the community doesn't oppose the filter. Only an idiot would simultaneously oppose the filter's existence and ask that the WMF make the filter's development and implementation a high-priority project, and our community is not composed of idiots.
I believe DGG is also correct in saying that there is far more passion among the opponents than among the supporters.
IMO the most reasonable interpretation of the results is that, on balance, most people consider it to be a medium- or normal-priority project, not something to be rushed or given priority over other important projects. WhatamIdoing 17:14, 7 September 2011 (UTC)[reply]
You are just hellbent on the filter being implemented, and on it being implemented with special filter categories. For the time being, I disregard your input about anything other than why you want a category-based filter so desperately that it leads you to this neverending staccato of fallacies and intellectual dishonesty. --213.168.119.239 20:42, 7 September 2011 (UTC)[reply]
I did not say that only a minority of Wikipedians totally opposed the filter. I said that only a minority of Wikipedians who voted did not oppose the filter. I know that I, for one, waited till the last minute to vote because I opposed the entire process of deciding first and asking the community afterwards. I think enough Wikipedians did oppose it, and strongly enough, including apparently the majority in one of the national Wikipedias, that it is reasonable to hope that the board will realize their error. A strategy of denial will not help convince them of that. DGG 21:56, 7 September 2011 (UTC)[reply]
What the? Me and DGG are in agreement over something? Quick, say something I can oppose before the universe implodes! — Coren (talk) / (en-wiki) 23:34, 7 September 2011 (UTC)[reply]
For sure denial will make them see, because no denial means to them that everything is fine, that they can decide a fundamental question without asking the communities. They all talked the last months so nice about why we loose so many contributers, how we can get more and hold them, but then the opinion of us is not even asked on one of the major questions for the projects: do we want any kind of censorship, even if individual and opt-in? The Board and the Foundation have no mandat to decide this without us. And this poll ... it is a joke, a bad joke. And noone can say if a majority of wikipedians want this filter, because they didn't ask us IF we want it, but they just asked some questions around it to give us the feeling to get asked, while the decision that it will come was already made. That is cheating, that is making us all fools. That is not the respect we contributers deserve. On such a fundamental question the people who write the Wikipedia have to get asked - and not just some questions to get fooled, but the real one and that is: Do we want this?! I would like to get to know by the Board memebers how the hell they could think they have a mandate to decide this without asking us. Something is rotten in the state of Denmark. Something is rotten in the Wikimedia/Wikipedia. I just can appael to the people who amde those decisions to stop and think and then to ask again, but with honesty and the respect we deserve. I don't like to get fooled. And from for example the vote of the most active german wikipedians in the Meinungsbild it is to see that Board and Foundation are lossing teh support in the second biggest project. I ask: Can that be in the interest of Board and Foundation to loose the hold, and to cut the bounds with the communities by making fools out of us? I doubt that, but i'm not sure if they really still ahve contact to us or live in their own world. Julius1990 04:28, 8 September 2011 (UTC)[reply]
Don't feel betrayed or victimize yourself when you think your sole opinion is not listened. Before this proposal there were regular inquiries about image (and article) filter in Chinese Wikipedia village pump. The filter is proposed on demand. Even though this talk page seems to be dominated by opposing voices, it doesn't necessarily represent the WHOLE community's trend. More or less the proponents give up reasoning with you because there's no conclusion could satisfy both sides. Albeit in vain, I have to say, labeling the images does not mean we forsake the principle of neutrality or discriminate the labeled images. We have articles about crimes and terrorism which is negative to our POV, it doesn't mean we glorify those idea, we just admit their existence, that's why we include them, as this proposal acknowledges the nature of human for being turned off by objectionable images subjectively with clear and common elements such as nudity, violence, etc. -- Sameboat (talk) 07:05, 8 September 2011 (UTC)[reply]
If there are so amny supporters that teh Board and Foundation could be sure of teh support, why not asking IF we want this filter? Why deciding for delegitimation the complete filter by making up a fake poll? And the number of negative comments in teh comments ection and so on should give you to think that there is simply no consens. And you should ask yourself if this can be decided by just teh Board. Why do you negotiate that the Wikipedians have to make such a fundamental decision? Because you fear that your position than wouldn't succeed? That can't be the point to think about in a democratic process and a democratic project ... why wasn't here gone the democrativ way to legitimate this flter? Julius1990 07:15, 8 September 2011 (UTC)[reply]
Sameboat...I appreciate your directness. You seem a little exasperated by all the screaching by users who have problems with the filter. Many of them are exasperated because their real problems are not being addressed by anyone anywhere except in a really superficial way. I have posted worries I have about the referendum itself and the filter on many forums several times and they have mostly been ignored. While the referendum didn't come directly out of the blue, for many users it did seem that way. Like "woah...what's this referendum? And why is the presentation of it so one sided? And why are the questions this way? Why was the community not invited to have a comprehensive dialogue about it first? What are your responses to these various problems we have?" and to be honest, I haven't seen an apropriate response yet in any of these forums anywhere. All I have read is "come on guys, its opt in only, why would you care?" or "its just a referendum...theres lots to think about still" or "look...its the first time we've done something like this, and it wasn't really us who planned it". I don't think the users here are under the mentality of "reject any reason FOR the filter" as you say they do. They simply have not had the biggest and most profound questions they have...ever answered. And so far I am only talking about the referendum. As for the problems some users have about the filter...many users have grave concerns about "how the images will be categorised", "how on earth a consensus will ever be met on the system of categorization and how to go about placing each image in a category", "can this tool be used by 3rd parties for underhanded means of censorship", "Is this the first step of a slippery slope towards censorship". These questions have not been properly answered by anyone yet. Only single sentence responses brushing away these concerns. And some users are very much against the idea of a community colaborating together in a great project and forming articles both text and images through consensus and finding that the community will now be responsible for sorting out images that will now be cut out of view by some users (voluntarily or not). And the biggest question "whats the difference between filtering images and filtering text? Images are just as integral to an article as the text is". While I personaly don't find a few of these questions troubling...all of them are valid concerns and questions that should have been thought of...talked about and adressed by the board, the committee and the entire comunity first. And we STILL don't have any proper answers to these questions. Where are they? --Shabidoo 14:11, 8 September 2011 (UTC)[reply]
Shabidoo: I answered some of your questions on the Forum. You are right that many people are reacting to the presentation of the referendum, and mistakes made there. Many are also against any reason for the sort of filter that was been proposed, though willing to take part in a balanced discussion of it. [the discussion on the German Wikipedia has been a well-balanced and well-argued one, for instance, even though the community is overwhelmingly against any such feature.] A few of the questions you raise were asked and answered in the less visible discussion that was the public development of last year's report on controversial content. Most questions related to categorization haven't been asked or answered before, since it was not clear what design would be used for an image-hiding feature. Questions of 3rd party censorship are also much more specific and answerable in the case of a particular design - so they too could be answered now in a way they could not before.
Someone decided something for the whole wikipedia at a place most of us don't even know? Do you really believe us to accept that kind of decission? There are a lots of complaints by people who do not want to see some contents in the WP. Some do not want other people to read something about their religion, parties etc. Why should we listen to them? We write an encyclopedia. Someone who thinks we are a radio station, that needs to pleasure its consuments has not understood the encyclopedia. --Liberaler Humanist 14:34, 8 September 2011 (UTC)[reply]
For that to be true, you would have to believe that "the whole Wikipedia" does not include the Chinese Wikipedia or any of the other languages where this is routinely requested. You are the victim of a en:False consensus effect: most of German speakers hate the filter, and you believe that this means most people hate it.
For your claim to be true, you must also believe that most of us are ignorant of who owns Wikipedia. Wikipedia is not owned by its editors. Wikipedia is one of many projects owned by the Wikimedia Foundation. As explained in the very first sentences of the referendum, the "yes or no" decision about the filter was made by the people legally in control of the WMF. That is the meaning of the sentence, "The Board of Trustees has directed the Wikimedia Foundation to develop and implement a personal image hiding feature." Auf Deutsch, that's hat beschlossen (has decided), not merely suggested or considered or asked about. Napoleon Bonaparte decided to invade Russia; the WMF's board of trustees decided to have a personal image filter. You are free to believe that both decisions will have equally disastrous results (and they might), but if it surprises you to learn that the board has both the power and the legal right to make such decisions, then your ignorance is your personal problem, not one that is generally shared by everyone. WhatamIdoing 16:17, 9 September 2011 (UTC)[reply]
You should know that Wikipedia is hosted by the Foundation, but the content belongs to anyone, due to our licenses. If the WMF proceeds without any reaction whatsoever then you might loose many German contributers or constitute a fork of the project which does not implement the filter and hosts it's own media. It will also play a big role inside the next fund raising. If it doesn't step back, it will damage the project anyway. The project lives from it's authors, the content they created, and not from happy censors or some prude/loud minority.
To prevent that, the board has to communicate with the projects before making such decisions. This indicates to me that there is deep ditch between the the current head and the community, a general lack of communication. (Smashing kittens at the bottom of user pages, that get buried alive by annoyed user, doesn't help in this case). --Niabot 17:00, 9 September 2011 (UTC)[reply]
No: You actually own your contributions. Anyone can use your contributions because of the licensing arrangements, but your contributions do not belong to anyone except you. Wikipedia, however, is an entity that is distinct from our contributions to it, and Wikipedia belongs to the WMF. WhatamIdoing 23:21, 12 September 2011 (UTC)[reply]
WhatamIdoing, please stay cool. The board is a body responsible for the wellbeing of the Foundation and accountable to the community of editors -- very unlike Napoleon, in fact. It has the ability to make such decisions, generally in response to the needs of the projects. And it has the obligation to pay attention to what the community and the projects need. This survey, for all of its flaws, was an effort to better understand that. (@Niabot: for example, both sides of the controversial content issues are convinced that the other side is a vocal minority, and think that the other side "enjoys shock images" or "enjoys censorship". Those caricatures are untrue; readers and editors both seem fairly well divided.)
We're still working on the communication piece. Forging better links between Meta and local Wiki-projects is a good way to proceed there; as has happened in the past weeks with the discussions on de:wp. SJ talk | translate   21:55, 9 September 2011 (UTC)[reply]
Regardless of the legal situation, Wikipedia is morally owned collectively by the editors who have created the content. 86.176.215.130 20:40, 9 September 2011 (UTC)[reply]
That depends on your vision for the project. As a platform for knowledge-collaboration, Wikipedia is designed for every person on the planet. So in a sense the platform is morally owned collectively by everyone who needs access to information. One of the arguments for adding new image prefs for readers is that it would allow more of those readers use the projects in more parts of their lives. One of the arguments against is that providing new options to readers could be used to limit how they are allowed to use the projects against their will. At present we lack feedback loops between readers (most of whom don't know how to edit!) and editors, so it is hard to know how relevant such arguments are in practice. SJ talk | translate   21:55, 9 September 2011 (UTC)[reply]

::::::SJ...thanks for your last message. It was informative. SJ would you please tell me...or can you tell me is there anyone who can respond to any of those questions (which I have asked numerous times now)? Is there anywhere else I can ask or another forum where I can ask these questions in hope that they could be answered by someone in the foundation/board (I have read as much as I can and its still difficult to see what is which for the referendum). This isn't a case of one side versus other side, or assuming the filter is a tool of censorship, but simply valid concens some users have and questions I'm sure all users would like to know the answer to. These concerns that users have are not being adressed by anyone in the board/foundation. It would probably be very helpful to see someone...giving at least some details of how the board/foundation sees these potential problems 1. How the referendum was carried out, 2. does that affect how they see the results? 3. Do they entertain the possibility of another referendum? 4. Are there lessons to be learned from the referendum and will new ones in the future be more carefully carried out by consulting the community first? 5. How would they reply to the user above with his own analysis of the results and how hard it is to many any meaningful cross analysis with the questions and 1-10 scale answers. And questions about the filter. 1. What does the boad feel about the difficulty in finding any consensus on which images should be tagged in a neutural way? 2. What does culturally neutural mean...and what would be one concrete example? 3. What does the board think of the potential use of the filter by 3rd parties to limit images that users would see using computers connected to that network? 4. What do they think about the possibility of the filter being the first step in a slippery slope of providing more and more tools that cross the line of personal control vs. censorship? 5. Is the board entertaining a community wide discussion about the filter?

Even an answer to a couple of these questions would probably do LOTS to lessen the concerns or bad feelings they have towards the filter and make these users (and all users as well) feel like they are a part of the process of implementing (or not) this filter. --Shabidoo 22:53, 9 September 2011 (UTC)[reply]
I think the OP made some good points, but there is some work that could be done to address these using the data. For example, you could make correlation matrices pairwise for each of the four questions, so that we see how much the answer on one question predicts that for another. Or you could designate statistical "mountains" like they do with gene expression data, where you try to use neutral means to classify most of the voters into one of a few model votes that fairly closely resemble theirs, then ask what these model votes mean.
In general, I think that what the vote means, when contrasted with the discussion pages, is that creating an opt-in filter doesn't sound like much trouble to people who don't think about it very hard. Still, the median vote is a 6 - not exactly a resounding endorsement. Wnt 16:43, 11 September 2011 (UTC)[reply]
What evidence do you have 'creating an opt-in filter doesn't sound like much trouble to people who don't think about it very hard'. If you are making the assumption that just because some people don't feel the need to argue their view to the death means they haven't thought hard about their view, that assumption is inherently flawed, not to mention highly offensive. In fact comments like these are likely one of the reasons some people feel there is not much point arguing their view. Nil Einne 10:34, 14 October 2011 (UTC)[reply]

"Next steps" planning page[edit]

Image_filter_referendum/Next_steps/en – In case users here aren't aware of this planning page yet. It appears to be open to revisions and additions by anyone. Feel free to share your ideas. --Michaeldsuarez 19:19, 9 September 2011 (UTC)[reply]

What was the point of this referendum?[edit]

It obviously wasn't to ascertain whether a feature should be implemented or not. I honestly don't know where the B.O.Ds were going with this? "Hey guys. Here's a list of incredibly slanted questions to answer about a feature that you may not want. We don't actually give you that option though." Terrible. 68.81.71.188 17:59, 20 September 2011 (UTC)[reply]

The Board of Directors did not order the referendum; the Board told its employees to create one. The employees separately decided to survey the community. I believe that their goals were two:
  1. to find out how high a priority the filter should be (Should we take resources away from other projects to do this? Or put this on the back burner and get to it later?)
  2. to find out which features were important (Should it be possible for an unregistered reader to use this? Should it be 'locked', so you can turn it on for the kids and the kids can't turn it off?) WhatamIdoing 17:11, 21 September 2011 (UTC)[reply]
If that were (or is) the goal, then they did a really bad job.
  1. To see the priority of something you will always need to compare it to other things. Otherwise you don't have a relation how important "important" really is. Ask some people how important it is "to improve the editor" or to "implement an image filter". Then you will get a relative result. What we have now is a result with no meaning, since we don't have any data for comparison.
  2. The last question was never asked in the first place, since locking (opt-out) was out of discussion. It where general questions and again without any comparison. On top of that the questions where ambiguous. What does it mean to vote on the other questions if you deny the feature at all (absolutely not important = 0)?
Therefore the results of this poll are pretty much useless. It asked the wrong questions to begin with and an interpretation of the results is not truly possible. If a statistician made this poll, i doubt that he would still have it's job or at least a strong word with his chief. --Niabot 17:59, 21 September 2011 (UTC)[reply]
Well, perhaps you did not understand the questions.
  1. The WMF already knows what the absolute priorities of other projects are. They asked the community how important this should be on an absolute scale. They got an answer in the middle. Therefore, it is a mid-priority project on an absolute scale.
    Think of it this way: If I want to know how our heights compare, I do not need to ask you directly "Are you taller or shorter than I am?" I can just ask you "How tall are you in centimeters?" and compare that to how tall I am on an absolute scale. With those two pieces of information, even a child can figure out which of us is taller. The same applies to this: I do not need to ask "Is this more or less important to you than improvements to the Article Feedback Tool?" I can just ask "How important is this to you?" and compare that answer to how important the other project is.
  2. The last question was asked, just perhaps not in a way that was clear to you: that was the point behind the item "readers should be supported if they decide to change their minds." The result is that whenever parents complain in the future, they have proof that easy reversibility is strongly supported by the community, with only 2% of respondents saying that readers should not be able to change their minds easily. I think the phrasing of that question was quite clever. WhatamIdoing 16:24, 22 September 2011 (UTC)[reply]
  1. There is no such thing as an absolute scale if it's about importance of something. It strongly depends on the voter how strong a "5" is. For someone who wants that filter it's nearly an oppose and for the others the opposite. Thats why we have a split vote. It is by far not an absolute scale. You can only make relative opinion polls that way. But since we have nothing for comparison it's useless to define what a "5" could mean.
  2. The other questions strongly depend on the first question, but they don't ask how the voter would react under different assumptions for the result of the first question. Thats make them also pretty much useless.
  3. We obviously had many protest votes, which are also present inside the other questions (since no assumptions where proposed). This basically destroys the result in general.
Overall this poll is doomed, thanks to weak, not well thought out questioning. --Niabot 16:57, 22 September 2011 (UTC)[reply]
We do not agree on these points, and I do not think it likely that we will ever agree. WhatamIdoing 17:32, 23 September 2011 (UTC)[reply]
I would add anyone who has ever actual taken part in a telemarketing survey will know that they nearly always do ask importance on a absolute scale. They rarely ask 'how do you feel about X compared to Y' they ask 'how important is X to you' perhaps also 'how important is Y to you' if they want to check on multiple things. This includes ones where the results are never published but intended for internal use. It may be professional market research companies despite their statisticians etc have no idea what they're doing, or their customers only ever hire these companies to prove something to someone rather then to really get a gauge of the mood of the public, I can't say. I don't really understand Niabot's second point (different assumptions for the result of the first question). As Niabot has already acknowledged, it's clear one assumption of the poll is the feature is going to be implemented, it's just a matter of when. Therefore, it shouldn't matter much what your answer was for the first question. Whether or not you think a filter is an important feature to implement shouldn't change in itself what features you think are important in it if it does exist. (There may be a correlation in views but that's a different matter.) If someone doesn't think the feature is important to implemnt, they may not have thought a great deal about what features are important in the filter but once the questions are asked, they know to think about such details. Why would the foundation want to ask 'even though you think a filter is not important to implement, imagine it was important to implement, in such a case what features would be important in it' or 'even though you think a filter is important to implement, imagine it was not important to implement, in such a case what feature would be important in it'? Nil Einne 11:02, 14 October 2011 (UTC)[reply]