Am 16.09.2011 10:40, schrieb Peter Gervai: Where exactly has such a vote taken place? Just a few bits about the "Meinungsbild" in the de-WP: Only active authors are allowed to vote since the results of a "Meinungsbild" are binding, unlike the results of ordinary polls ("Umfage"). Of those 14% who did not oppose the filter, I did not really see much actual support for it either. The general tone of the people who did not vote against it was that they don't mind that such a tool should be introduced if there's really demand for it. The 86% rejection rate means that the feature will not be activated in the de-WP and that the WMF would be in a heap of trouble if they tried to force the second largest project to adopt something that the people who actually shape the project simply do not want. I believe it is also safe to assume that those 86% are not likely to do the dirty work of tagging pictures with categories to support the filter. Regards, Oliver

On Fri, Sep 16, 2011 at 11:23, David Gerard <dgerard(a)gmail.com> wrote: Well I am the universally official source for my own beliefs. But then you state that WMF will make it compulsory for all projects to activate the feature? (Citation is welcome, sure, but not required.) g

On Fri, Sep 16, 2011 at 11:31, David Gerard &lt;dgerard(a)gmail.com&gt; wrote: Please read what I wrote, you have quoted it. If I wanted to write "we have voted as" then I would have written just that. I didn't. Well I do not have the original poll handy but as far as I remember it it was about what we think would be good to have, what to would like to see implemented. I do not remember any question about making it compulsory. Do you? I am not sure what is the point debating this with _me_. (Apart from my person I mean.) I am not German. I am not active on DEWP. I voted for the feature, and I believe it's good to have it. You try to teach a lesson to me about your own troubles, but I really cannot help it. The only thing I can offer my views are the global poll about the feature, and yes, it wasn't a strong concensus. But even it it were I do not think we should change the otherwise very well working method of WMF *not* messing with local projects apart from the very basic principles like the five pillars. This feature isn't *that* important - this is my opinion, please save me from asking a citation. Ironically this was what I was talking about, and what you were rejecting. All I say is that if a local project vote not to use a feature then they shouldn't have to. If you disagree with that one you can simply say it, but do not try (and fail) to describe what I want to convince people about, please. <g>

On Fri, Sep 16, 2011 at 12:15, Tobias Oelgarte &lt;tobias.oelgarte(a)googlemail.com&gt; wrote: Well you have to see that such controversial features will never have huge consensus in such a large and diverse community. Even simple majority would be an awesome result. :-) g

Am 16.09.2011 12:22, schrieb Peter Gervai: In a poll that asks for importance and is divided more or less into two groups? Thats a very strange interpretation. Tobias

Yes, and as was mentioned previously this should be a very easy query to run. People are wondering why the data isn't already available. 2011/9/16 phoebe ayers &lt;phoebe.wiki(a)gmail.com&gt;

Am 16.09.2011 20:38, schrieb phoebe ayers: It took you three days to get out the first results, with most time spend in reading comments. Now i simply asked for a table from the database and it takes more then two weeks, while it could calm down the controversy about the poll and provide good arguments. The longer it takes the worser it gets. So you should hurry up on this simple matter. Just the raw data, a simple table Q1 | Q2 | Q3 | Q4 | Q5 | Project (> 20 voters or "other") | date (no time) with random order of rows would satisfy. We can make the analysis alone based on this data, provide the plots and so on. What we don't want, is a final conclusion without the anonymous raw data. Tobias

On 16 September 2011 11:23, David Gerard &lt;dgerard(a)gmail.com&gt; wrote: Meanwhile, over on Bugzilla… https://bugzilla.wikimedia.org/show_bug.cgi?id=30208 Am I the only one who thinks this is getting somewhat out of hand? Stuff like "I'm getting tired of your aggressive comments and borderline personal attacks (…) All you did at Wikimania was to publish a flyer full of proven lies to reinforce your mantra" (responding to things I wouldn't at all consider aggressive) is perhaps a sign that tempers are more than a little frayed. Michel

On Fri, Sep 16, 2011 at 11:40 AM, Peter Gervai &lt;grinapo(a)gmail.com&gt; wrote: Not so, the resolution was to implement, not merely develope. -- -- Jussi-Ville Heiskanen, ~ [[User:Cimon Avaro]]

On 16 September 2011 02:14, Fae &lt;fae(a)wikimedia.org.uk&gt; wrote: Yikes, this is a very fast-moving thread. I haven't read it all yet, but I wanted to jump in and say that yes, there has not yet been an official statement responding to the referendum results. There will be, but there isn't yet. Currently, the referendum team is still doing some analysis of the results -- there are some questions we are hoping to get answered around language breakdown. And I am currently reading lots and lots of write-in comments. If I had to guess, I would image there will be a statement within about two weeks. But that's not a commitment, just an estimate. Thanks, Sue -- Sue Gardner Executive Director Wikimedia Foundation 415 839 6885 office 415 816 9967 cell Imagine a world in which every single human being can freely share in the sum of all knowledge.  Help us make it a reality! http://wikimediafoundation.org/wiki/Donate

On Fri, Sep 16, 2011 at 12:39, emijrp &lt;emijrp(a)gmail.com&gt; wrote: Oh, you found the first useful purpose of the image filter!

On Fri, Sep 16, 2011 at 12:50, Tobias Oelgarte &lt;tobias.oelgarte(a)googlemail.com&gt; wrote: Then, make it opt-out :P

On Fri, Sep 16, 2011 at 13:45, Jussi-Ville Heiskanen &lt;cimonavaro(a)gmail.com&gt; wrote: It's about implementing image filter on images which have copyright problems in Germany (but not in US), not about nudity.

On Fri, Sep 16, 2011 at 16:27, Jussi-Ville Heiskanen &lt;cimonavaro(a)gmail.com&gt; wrote: Ah, you are right! I completely missed the point of the image filter because my obsession... It's about scared places of indigenous peoples of Australia and similar. Back to the initial "point": "Make it opt-out" was about users of German Wikipedia, not about the projects. Whatever the point of the filter is. If it's illegal to see apples in Germany, then they could impose the filter for all of the users and allow them to opt-out from not seeing apples.

Am 16.09.2011 16:47, schrieb Milos Rancic: You make really crucial mistakes in your argumentation. We have no legal problem to view the images. We have a legal problem to provide them. If you implement it as opt-out then you would rip off the rights of the readers. A complete fail. But again: We talk about the opt-in image filter feature and that the core of the second largest project isn't willed to accept it.

On Fri, Sep 16, 2011 at 7:55 PM, Tobias Oelgarte &lt;tobias.oelgarte(a)googlemail.com&gt; wrote: Take a deep breath and read up the chain. You started this thread, and I am doing my level best to try to stop it being derailed from the subject of the German vote and its repercussions -- -- Jussi-Ville Heiskanen, ~ [[User:Cimon Avaro]]

On Fri, Sep 16, 2011 at 18:52, Tobias Oelgarte &lt;tobias.oelgarte(a)googlemail.com&gt; wrote: Not opt-in, but opt-out :P Sorry for hijacking the thread with a joke :) I think that everything is so obvious that it's become tiresome to discuss it: * There is significant disproportion in position between editors with a couple of edits and the core of the community. * It's not likely that it would be ~85% against, but similar pool on English Wikipedia would likely finish with ~60% against. Hypothetical referendums on projects in many European languages would finish similarly to the referendum on German Wikipedia, as in this case macho-patriarchal culture, dominant in large parts of Europe, corresponds with libertarian positions, dominant among the core editors. * It's likely that staff and Board already know that correlation between the results of German Wikipedia referendum and global survey could be drawn to support previous two conclusions. Thus, they don't want to publish that part of data. * There is still significant minority of core editors who want the filter at any cost. * Board is divided and doesn't know what to decide. I would repeat the best possible solution to end this: Implement it on English Wikipedia -- you (those who want that filter) have some numbers which would support that action -- and leave the rest of the projects alone.

Hang on, I thought that http://wikimediafoundation.org/wiki/Values (which underpin the Mission) means that WMF is obliged by their own published bylaws to openly publish the data in question? Cheers, Fae

Am 16.09.2011 19:13, schrieb Milos Rancic: That still has to be proven. I asked for localized (project based) data from the poll to inspect if there are huge cultural differences or if there is a general bias towards the filter. This was more then two weeks ago and i reminded Philippe repeatedly to release this data. So far nothing was released and one excuse followed the other. Thats why i can't support or oppose your statement. But assuming that it would be true would be as false as to say that it is false. So i repeat my request again: "Philippe, can you hear me? Release the data as soon as possible, we need it". You would have to proof that your facts are indeed true. But if you accept it as a huge difference between cultures, how can you impose a filter for a culture that doesn't need it or wants it? How would you expect to find a good compromise in decisions on what to filter and what not? Do you intend to put an extremist conservative Arab and and the most liberal German inside the same room, close the door, go away, come back after two weeks and look if they could find a compromise about Yes or No? How should this work? The referendum showed that cultural neutrality is important for the voters. But how do you think to find a compromise between hell and heaven, without having hell and heaven inside the discussions at commons at earth? I doubt that. But if they do, I will call them "assholes for betrayal". Just to make it clear. It would also not suite the story onto who has access to the data and who has not. A "significant minority" is a curios choice of words. "A significant minority tries to abolish the constitution by any cost". Now ask yourself if you would follow their wishes. Thats the same sentence, you said, with different actors. Still happy with it? We don't know what the board thinks. It does not communicate with us (the authors), it did not react to the discussions at Meta, it did not answer serious questions and in general is somewhere between a legend and a forgotten ghost that no one can see, even if present. That would imply not to implement it on commons. Otherwise the the categorization/labeling/... could be misused by local providers inside regions that didn't intended to use this feature. Tobias

On Fri, Sep 16, 2011 at 08:19:05PM +0200, Andre Engels wrote: In this case there's this Evil Cat system we may need to set up. I don't want anyone to use that. Especially not evil people. sincerely, Kim Bruning -- Blofeld would have a field day!

On Fri, Sep 16, 2011 at 9:13 PM, Tobias Oelgarte < tobias.oelgarte(a)googlemail.com&gt; wrote: So how do you expect those censors to use this? -- André Engels, andreengels(a)gmail.com

I don't except it to work perfectly. Nothing is perfect by default. But even if it would perfectly we provide a simple tool (the filter labels/categories) to censors, to improve their doing, while we, the volunteers, would indirectly support them in doing so. For example: The head of a group (state, religions group, ...) of people is trying to censor Wikipedia, because it might damage it's position. What would be easier to comply at the mailing list that a filter for xyz is seriously needed. Now he can start to add images to this filter, calling for volunteers that have to obey to do so. At the end we represent the opinion from the head of the group (not the individuals, that fear the head), publish it as consent and help them to justify their position. What would someone living inside such a group think if the content is already labeled that way, that he should not look at it. Isn't it social pressure put on the free mind, especially if other members of the group are around?

Am 16.09.2011 22:53, schrieb Stephen Bain: Would we do that for text as well? Where is the fundamental difference between text and images? Both can be objectionable or offending to some readers/viewers. Is there a real difference? Wouldn't be a simple button to hide all images be enough to reach our goal, without the need to introduce categories?

Am 16.09.2011 21:57, schrieb Andre Engels: Just ask yourself what our Wikipedia interface would do. The server provide the images (HTML-Documents with <img> tag) along with labels. Depending on the settings of the user some kind of Javascript will hide the images. This "passed along" labels could simply be used to exclude the image as the whole, making the "show image" button disappear. Since Wikipedia serves more or less static pages, due to seriously needed caching, the labels will need to be passed that way. Now you should think about topic and try to understand why this opens for a new kind of censorship. Blocking Wikipedia as a whole is a problem for most providers. This will cause users to change the provider or to insist to have access to it. This is a pressure put onto the access provider. The provider itself isn't able to filter the image or the content, since this is a lot of working time and time costs money. But if we choose to label the content for no fee, we open a new field for partial censorship. The users could still access it, but they won't see anything. In the result there would be some complaints. But way less complaints as if Wikipedia wasn't present at all. A good compromise for a censor.

Am 16.09.2011 22:52, schrieb Stephen Bain: Do you believe that it would help you if the image page itself will also be blocked upon the same labeling or that it wouldn't be easy to remove this traces as well? Speaking as a censor, this would have high priority for me. Showing others that i try to hide something? I would be stupid if i didn't do it that way. Leaving traces of what i have done would cause me possible problems. Thats why a censor would not be so stupid to leave you a link to click on to view the image. He will remove the image together with it's surrounding HTML-Elements. Therefor you will have to make the request from a mirror. But where does your browser get the information where to get the image? It lies within the filtered HTML-document. Ooops sorry. It isn't there anymore because it was filtered. I already stated that it is trivial to remove the caption (the whole image and it's container elements. Your Javascript experts could write a working code for that in half a day or less. That would not be an great effort. Altering text is much more difficult, because it isn't labeled. It changes over time. An image is included in a repeating pattern. But now i must wonder about your argumentation. You said that the "risk is marginal" and it isn't a "high" priority task. Looking at the boards decision it is seen as an task of high importance. I'm wondering why it is that way. For the exactly same reasoning as why providers tend not to block Wikipedia entirely. Blocking such projects partially, only representing the suitable information is a much better choice. If you look at something at Wikipedia and have no such feature, then you won't need an excuse. But if there is an filter, then you need an excuse why you didn't use it.

On Fri, Sep 16, 2011 at 09:01:04PM +0200, Milos Rancic wrote: You're better than I am! I only got Phoebe, Sj, Kat, and Jimmy. How'd you get the rest? sincerely, Kim Bruning --

Am 16.09.2011 21:01, schrieb Milos Rancic: This would imply that the referendum indeed asked the wrong questions. If all would have equal values, then i must wonder about the strong difference in result. We have a referendum which points out that many are in favor of this feature (important) and we have a Meinungsbild at the German Wikipedia closed with 86% against the filter. This is a huge difference. If it is not based on the fact that cultures are so different, what would be the reason? The questions and the interpretation? One of the aspects why I'm so interested in per project raw data and overall participation (number of votes per project). It was just an example (a literal allegation). The current proposal (as represented in side the referendum) did not assume any cultural difference. My thoughts on this is, how we want to create filter categories which are cultural neutral. One common (easy to describe) example is nudity. What will be considered nude by an catholic priest and an common atheist, both from Germany. Will they come to the same conclusion if they look an swimsuits? I guess we can assume that they would have different opinions and a need for discussion. Would we need this discussion until now and for all images? No we did not. We discussed about the articles and would be a good illustration for the subject. But now we don't talk about if something is good illustration. We talk about if it is objectionable by someone else. We judge for others what they would see as objectionable. That is inherently against the rule of NPOV. That isn't our job as an encyclopedia. We present the facts in neutral attitude toward the topic. We state the arguments of both or multiple sides. A filter only knows a yes or no to this question. We make a "final" decision what people don't want to see. That is not our job! I would also need to lie, but some progress would be nice. We already represented different alternative models to Ting at Nürnberg WikiCon 2011. So far his reaction described exactly what you think. They don't know what to do in the current situation. What we could do (i might not speak for all German users, but for many) is to implement a very simplified approach. A simple button to hide all images or no image. If you think you might read about a topic that could be controversial, you could enable this function and display images which you are sure about to be not offended. You could represent articles to your children, without the fear that some image might slip through the filter. Additionally we would have no problems with NPOV, no additional work. It would be very easy to implement. We would not play in the hands of censors or possible censors and we would solve many of the issues. (Mohamed? No image no Mohamed. Children? No image, no child will see porn.) No need to invest valuable time in a feature and a big categorization progress, with little or big wars inside this new playground. The only thing this proposal can't do: It won't make "ugly words", words you don't want to read, disappear. But that was out of question from the beginning. I see the problem, that devil sticks within the details. It is not the filtering at the of a project. It is the whole infrastructure (not only the technical) which creates the problems. It's prejudice about content, it's labeling content as objectionable and so on. This will still happen, even if the German community would not participate or use/enable the filter. The basic thought progress behind this whole idea is what _I'm_ opposed to. I don't know where you got this information. But I would not wonder if it is as it is presented by you. At least in case of Ting and Jimbo you should have right. I learned with the time about Jimbo, his attitude towards topics and it's understanding. So i have no doubt that he would trade intellectual freedom against some more donations. I saw Ting's reaction at the WikiCon and i have to say that he wasn't really prepared and not happy with the drama. But saying that it is already decided no matter what the German community would say, was a punch in the face of the audience. I guess she will remember that. That is my personal main issue with the whole filter thing based on arbitrary non-neutral labeling of content and POV as the measure for judgment.

Am 16.09.2011 22:37, schrieb Andre Engels: The requirements where different. Thats right. But i don't believe that this would influence the overall outcome. I monitored some articles about sexual topics for a while (especially Hentai, Lolicon, Futanari, and so on) that had controversial images in them. Additionally i looked up the numbers of views per day and compared them with the number of complaints over time. In the last year we had one short request to remove the well known image from Lolicon, because it was considered porn (which it isn't). That was the first and only complain in the whole time, despite the fact that every of this articles has more then 1000 views a day. On EN the story is a little bit different. The articles have in average the double or triple amount of viewers, but you will get at least one complain in a month per article. This is still very very low. That means that one out of approximately 90.000 viewers will complain (including the constant repeating complainers). Assuming that only 1 of 100 viewers would leave a comment while being offended, then it is still approximately 1 in 1000 (0,1 %). This a vague fact on for how many users the filter would be an maybe helpful tool. Compared to the German Wikipedia it is still high in relation. (I did not count the zeros after the comma.) Given that, can we assume that the filter is really necessary or that it would be an valid explanation based on the fact that the polls had different requirements for voting? If you compare my simple (not waterproof) calculation to the result of the German poll you might think that this isn't true. Only 86% opposed the filter. Thats right, but you will also have to consider the public comments. There are only some that found it important that the filter will be implemented, since they see use in it. The most opposing votes (14%) are: Let them do it, it won't hurt, but not strictly in favor of the filter itself. And again (now i must repeat myself) i have to say, that we aren't able to manage such an amount of categories. Every category increases the work drastically and complicates it. But even if we are able to categorize millions of images, then we still have to represent them to the reader who is willed to filter the content. A dialog with options to sort out of hundreds or more categories seems not the a viable option. Who would really take his time to go trough all points as an IP-user who loses it's settings if switching the browser or the workplace? That would be anything but not user friendly. So you will have to make a compromise (i assume 10-20 categories), where you don't have this fine graded model. You proposed alone 7 categories for nudity. What about other topics? You will easily reach high numbers. That is an high effort for categoization/labeling as well as for the user. It would just not work that way and I don't believe that we can proceed under the assumption of having an infinite amount of labels. Judging that nothing is objectionable is our job as the writers of an encyclopedia. We are not hiding facts based upon the view that it might be objectionable for some users. This is the most basic rule that every encyclopedia followed, if not under the control of a dictatorial system. Thats exactly why library organizations or the editorial staff of encyclopedias are opposed to label content by non-neutral categories. For example have a look at http://www.ala.org/ala/issuesadvocacy/intfreedom/librarybill/interpretation… Which claims for more then 50 years that labeling after objectionableness is in violation with the basic rules of intellectual freedom. We have the same rules. One of them is NPOV, which covers this topic as well. Follow the link above and you will hopefully understand what this is about. Intellectual freedom means to represent knowledge in way that does respect the facts, but does not respect the audience and their wishes or likings. It means to represent knowledge, but not to bend before people who think that something might be objectionable. Not censoring yourself and making knowledge available to everybody, that is the definition of intellectual freedom. The POV starts when you draw a line between what is nude and what is not nude. Is someone wearing a half transparent swimsuit nude? (yes/no) Is it still (not) nude if it is a tiny bit thinner/thicker? (yes/no) You see that even this simple question can't be answered with a simple yes or no. No think about what you use to separate between nude content and non nude content. Do you have a strict catalog, with points to follow. Isn't this your point of view where nudity begins and where it ends? Do you think that your neighbor would come to exactly the same conclusion? He won't. It will not divide far away from each other. But this is also the simplest question. What about violence. Is a boxing fight inside a ring violence? Does an image from after the fight depict violence, even so there is blood on the floor and both boxers shake hands? Now it gets a little bit more complicated. Now we could move on to next topic and i could ensure you, that it won't get easier, only much harder. You could play the judge, but i can guarantee you, that won't be able to claim neutrality on this matters anymore. To be truthful to yourself, you will have to admit that you have choosen "nudity" as the example, because it is much easier to handle then every other topic. But still you found a lot of categories to divide between different grades of nudity.

On Sat, Sep 17, 2011 at 8:16 PM, David Levy &lt;lifeisunfair(a)gmail.com&gt; wrote: Thereby giving those who have objections nothing just because there are others who we can't give what they want. If we had the same attitude towards article creation, we would not have published Wikipedia until we had articles on all subjects we could think of. We don't say they're unreasonable, we say that we don't cater to it, or at least not yet. That may be non-neutral, but no more non-neutral than that one subject has an article and the other not, or one picture is used to describe an article and the other not, or one category is deemed important enough to be used to categorize our articles, books, words and images and another not. Or even clearer: that one language has a Wikipedia and another not. Wid we make a non-neutral choice that only certain languages (the ones for which Wikipedias exist) are valid languages to use for spreading knowledge? I'd say it is, provided there are people wanting to use the filter, and not minding the fact that in the beginning it will be far from perfect. I'd say if there are people actually wanting to use such a filter, then yes, I would think we might well get one. -- André Engels, andreengels(a)gmail.com

On Sun, Sep 18, 2011 at 12:20 AM, David Levy &lt;lifeisunfair(a)gmail.com&gt; wrote: Wikimedia *used* to hold the position that we wouldn't aid China to block images of the Tianamen Massacre, and went to great lengths to assure that chinese users of Wikipedia could evade blocks to viewing. I am not sure you are on a right track with regards to our traditions and values here. -- -- Jussi-Ville Heiskanen, ~ [[User:Cimon Avaro]]

On Sun, Sep 18, 2011 at 10:46 AM, Andre Engels &lt;andreengels(a)gmail.com&gt; wrote: a>> here. Okay. Is there a commitment on the part of the foundation that they will help people using our filtering scheme and the usual browser add-ons to Wmake it impossible to view material on wikipedia from schools with a religious orientation, but students of way past any reasonable age of consent get past the filters? Once you tag something, you lose control over what that tag is used for. We might be talking about primary schools, but we could just as easily be blocked by colleges and high schools, where limiting hormonal damage would be a joke. -- -- Jussi-Ville Heiskanen, ~ [[User:Cimon Avaro]]

Am 18.09.2011 14:02, schrieb Andre Engels: While such content is excluded/filtered out by the local school network? Then they have the choice to not view the image with active filter or to not view the images with disabled filter...

On Sun, Sep 18, 2011 at 11:45 AM, Tobias Oelgarte < tobias.oelgarte(a)googlemail.com&gt; wrote: On itself the one who tags the image, but we happen to have a system for that in Wikimedia. It is called discussion and trying to reach consent. Who decides whether a page is in a category? Who decides whether a page has an image? Who decides whether something is decribed on a page? All the same. I don't think it's that easy, and if it were, the best thing would be to make it harder to misuse rather than to throw away the child with the bathwater. -- André Engels, andreengels(a)gmail.com

2011/9/18 Oliver Koslowski &lt;o.nee(a)t-online.de&gt;de>: +1 At the beginning, I was quite neutral about a filter: I had no idea how it would work, and I wouldn't use it, but what if somebody else wants it? But after reading nearly all comments on this list, I think that the arguments for a filter do not hold water. The pratical implemention would be a nightmare, and the purpose not really within Wikimedia mission. The thread above on how to create categories for a filter is full of irrational assumptions, impracticable propositions, and impossible solutions. It seems it is time to drop the whole idea... Regards, Yann

Am 18.09.2011 15:58, schrieb Fred Bauder: Nothing wrong with hosting that picture (http://en.wikipedia.org/wiki/File:Muh-hund-original-rondellliten.JPG ). It's really about where it's used. Should it be used in the article on Muhammad? Heck no. But in the article (http://en.wikipedia.org/wiki/Lars_Vilks_Muhammad_drawings_controversy ) describing the controversy that arised around that very picture? Sure enough, because it serves an encyclopedic purpose there. Regards, Oliver

On Sun, Sep 18, 2011 at 3:02 PM, Oliver Koslowski &lt;o.nee(a)t-online.de&gt; wrote: I don't see a difference. I want to show images showing so-and-so, or I do not want to see them. It's all about saying whether images show so-and-so. It depends. If people want to do it, it is their choice how to use their volunteering time. If they don't, then bad luck to those using the feature. I do agree that there are dozens of things in Wikipedia/Wikimedia/Mediawiki that I'd rather see; I chose the secon-lowest rating in the referendum, and might well have chosen the lowest had I not expected that to be understood as "I am against this". I do think there are many better things to do with our time and other means. -- André Engels, andreengels(a)gmail.com

André Engels wrote: In our normal categorization scheme, "so-and-so" might equal "a beach" or "a woman wearing a swimsuit." Such classifications are straightforward, uncontroversial and not mutually exclusive. Under the proposed setup, "so-and-so" might equal "nudity." We'd need to decide whether the aforementioned photograph of a swimsuit-clad woman on a beach qualified, even if she merely appeared in the background (and therefore wouldn't be a factor in our normal categorization). It's reasonable to oppose the change on the basis that it would divert time better spent on other tasks, particularly given the likelihood that many non-supporters would participate in the process to prevent misuse. Additionally, if and when the WMF proudly announces the filters' introduction, the news media and general public won't accept "bad luck to those using the feature" as an excuse for its failure. David Levy

I understand that the details (well, quite big and relevant details) of this concept was the topic of the survey. So probably it has not been mapped out yet (because it was/is unknown), but that would be the next step. I also would like to make a sidenote: if the main argument of the German Wikipedians would be that this categorization an sich would be evil because it can be used by governments and ISP's etc, then I have to disappoint you: even if only one project would like to make the implementation of a filter possible for their readers, categorization would appear. Further, categorization of images will be happening likely on Commons (my guess) - so even if you opt out as German Wikipedia (although personally I think it would be more interesting to do a reader survey inside the German langauge visitors before deciding on that) it would not help that specific scenario. Lodewijk Am 19 de Setembro de 2011 09:47 schrieb David Gerard &lt;dgerard(a)gmail.com&gt;om>:

Zitat von Tobias Oelgarte &lt;tobias.oelgarte(a)googlemail.com&gt;om>: From what I understood the image filter will not have subjective criteria like "a little offensive", "very offensive", "pornography", but neutrally decidable criteria like "depicts nude female breasts", "depicts the face of Muhammad", "depicts mutilated dead body". If you select these criteria carefully there should be no need for any "sources" for your decision to put a file in the criterion's category. Either the image depicts the category topic or it doesn't. Marcus Buck User:Slomox

Am 19.09.2011 15:33, schrieb me(a)marcusbuck.org: We discussed this already and came to the conclusion, that you would need hundreds of these categories to filter out most of the "objectionable content". But that is neither manageable from our side nor manageable by the user. You run into a deadlock. Either we will end up having some rather subjective categories or we have whole lot of them, we can't manage (at least not under the assumption to be user-friendly or wasting a whole lot of resources for tiny group of readers).

A "dead human bodies" category that excludes mummies "because we're not idiots" is, by definition, not neutral. 2011/9/19 Stephen Bain &lt;stephen.bain(a)gmail.com&gt;

Fae wrote: Hasn't this already happened, albeit on a voluntary basis, and with free distribution? http://schools-wikipedia.org/

Does not meet the criteria previously discussed, I immediately find unrestricted depictions (and photographs) of bare breasted women and close up photographs of dead bodies. They made the mistake of putting culture before religious taboos and images that might scare children. * http://schools-wikipedia.org/wp/m/Mummy.htm * http://schools-wikipedia.org/images/150/15070.png.htm * http://schools-wikipedia.org/images/893/89336.jpg.htm * http://schools-wikipedia.org/images/925/92551.jpg.htm Cheers, Fae

Am 19.09.2011 18:08, schrieb Stephen Bain: If we implement an "hide all images" option, then we solved already 95% of all possible use cases mentioned before. Now we take on the doubtful work to categorize for even lower potential need? I support the "hide all" option. But if we have this feature, then, especially then, i see absolutely no need for any categorization. Then we create hobby project for censors with even less support from the community. I definitely can't follow your reasoning in this case.

Stephen Bain wrote: With such an option in place, why take on the task of labeling images on readers' behalf? As previously discussed, it would be a logistical nightmare and enormous resource drain. Furthermore, I don't regard "If you don't like our specific filters, you can just use the blanket one." as a remotely appropriate message to convey to the millions of readers whose beliefs we fail to individually accommodate. Or we can simply accept the idea's infeasibility and provide the non-broken image filter implementation alone. Note that disclaimers won't satisfy the masses. (If they did, we wouldn't be having this discussion.) As soon as the WMF introduces such a feature, large segments of the public will expect it to function flawlessly and become outraged when it doesn't. I foresee sensationalistic media reports about our "child protection filters that let through smut and gore." (I realize that we aren't advertising "child protection filters." Nonetheless, that's a likely perception, regardless of any disclaimers.) Firstly, I don't know why you've singled out genitals. People commonly regard depictions of other portions of the human anatomy (such as buttocks and female breasts) as objectionable. Secondly, "imagery subject to religious restriction" (which doesn't constitute a viable "category" in this context) includes "images of unveiled women." You state that "almost all will be in the first two categories," but I'm confident that we host significantly more images of unveiled women than images of human genitals and images of dead human bodies combined. Thirdly, there's been a great deal of discussion regarding other images to which people commonly object, such those depicting violence (whatever that means), surgical procedures, spiders and snakes, to name but a few. ...thereby arriving at an inherently subjective, binary determination that fails to meet many readers' expectations. Or we could simply provide that functionality alone, thereby enabling the same scenario. Indeed. David Levy

On 19 September 2011 15:50, Fae &lt;fae(a)wikimedia.org.uk&gt; wrote: How much is "mutilated"? A scratch? Ten scratches? A hundred scratches? St Sebastian? http://commons.wikimedia.org/wiki/File:Sebastia.jpg This is what I mean by "magical categories". "I know it when I see it" isn't enough. "Lots of us know it when we see it" is a blatant neutrality violation. You need to be able to answer these sort of questions if you advocate a filtering system involving any assessment of image content. - d.

On 19 September 2011 16:14, Stephen Bain &lt;stephen.bain(a)gmail.com&gt; wrote: I'd estimate the chances as pretty high that we're going to get a thorough exploration of every possible axis that's measured for a filter. So you're thinking to apply this only to photos, then? - d.

Marcus Buck wrote: The WMF outline cites "5–10 categories," with "sexual imagery" and "violent imagery" as examples: http://meta.wikimedia.org/wiki/Image_filter_referendum/en As Tobias Oelgarte noted, even with the type of specificity to which you refer, the omission of countless widespread objections (a non-neutral, discriminatory practice) would be unavoidable. Who's going to analyze millions of images (with thousands more uploaded every day) to tag the ones containing unveiled women? Tobias also pointed out that even if we _were_ to include every widespread objection, the resultant quantity of filter categories would be unmanageable. As David Gerard noted, terms like "mutilated" are highly subjective. At what level of injury does a murder victim qualify? Fae mentioned ancient Egyptian mummies. Do they count? Does an autopsy constitute "mutilation"? Is a dead man's circumcised penis "mutilated"? (Many people would argue that it is.) Do non-photographic images qualify? If so, this article contains several images that arguably should be filtered: http://en.wikipedia.org/wiki/Descent_from_the_Cross Speaking of artwork, Fae mentioned the depictions of "nude female breasts" contained therein. Do those count? What about photographs of breasts taken in medical contexts? Are those equivalent to those taken in sexual contexts? If not, how do we define a "sexual context"? Do you foresee clear consensus in these areas? This is what Fred Bauder means when he notes that "some depictions of such things are offensive or pornographic and some not at all." But this is entirely subjective. We can expect endless debates on these subjects and countless others (and probably tag wars as well). Conversely, the alternative image filter implementation on which I've harped would enable readers to decide for themselves on an individual, case-by-case basis: http://meta.wikimedia.org/wiki/Talk:Image_filter_referendum/en/Categories#g… or http://goo.gl/t6ly5 David Levy

On 9/19/11 11:24 PM, Kim Bruning wrote: Censoring Freedom incarnated... for a free encyclopedia.... ça ne manque pas de sel....

* Andre Engels wrote: They are given plenty, in fact there are all sorts of filters already in place that lead to people not being exposed to media they do not wish to be exposed to, starting with laws, what people regard as appropriate for encyclopedic content, and local cultural norms affecting the latter, to name just a few examples. They also get to see images they do not object to without additional effort, and they have the option to hide them all, and they can be careful which articles they load, avoiding those likely to feature media they do not want. Exposure to things you are uncomfort- able with, where the feeling is not overwhelmingly shared, like when you click a link to a foreign Wikipedia, is also giving them something, most likely something good (consider images that cannot legally be displayed publically in some jurisdiction; they will be so displayed in others in- cluding on Wikipedias where they are relevant, with no filtering for the users from the jurisdiction that banned it, up to where it's likely that a court that Wikimedia cares about orders the image to be taken down). You can consider the last point from the other direction: if you don't like to see media with human suffering, horror, graphic violence, so you filter them, what should you see when reading about Nazi concentration camps? Profile pictures of Nazi leaders, architecture, maps maybe, but please, not the prisoners? What about people who find it wrong to make it easy for others to scroll down a Wikipedia article without the reader being visually confronted with human suffering if there is a consensus to display it at all in the article in this manner? Or, for that matter, that in this context it is wrong to require the reader to tell the com- puter "Yes, please show me piles or rotting corpses of starved people!"? Note that it may be the user of the filter who thinks this, in which case not giving them a filter that would do this is addressing one of their needs aswell (while failing to address another need; giving them a context-aware filter that avoids this problem would work of course, but then the system would be harder to use making it worse, and so on). So we already do plenty so people are not overexposed to media that they reasonably do not wish to be exposed to (note that I use "wish" broadly, someone suffering phobiae for instance has more of a need than a wish to avoid certain kinds of media). To a point really where I am unsure what is left that can realistically be optimized even further, and I am some- what surprised the "referendum" had so many people voting that this is of the highest priority (whatever that means given that this wasn't com- pared to other things that should be of high priority), though since it was already decided to introduce a filter because there is a problem, it can be assumed that some of the votes are subject to confirmation bias. (I do not know people who would say they frequently encounter articles on Wikipedia featuring images that would disturb them no matter where they appear and would thus prefer to have those pictures hidden for them, though I do know people who would prefer that medical articles concerning humans feature images that go beyond nudity, like showing the symptoms of a disease, or a corpse that has been cut open, towards the end in a specially labeled section, and people who do not like insects much, and thus do not browse around articles on insects. Neither of the two examples leads to me thinking a filter as proposed is the solution.) These analogies are all invalid as individual preference is rejected as argument in all of these cases, while the filter is solely concerned with individual preference (rather: aggregated individual preferences). Tagging an image with tag X, knowing that this will lead to the image being hidden to users who chose to hide all X images, is a matter of whether these users want the image to be hidden, who are in a minority, because if the majority agrees an image should be hidden, it would not be shown at all, no need for a personal filter; with the added problem that the people affected by the tagging are the least likely to engage in discussions about the tagging; and the added problem that this is by its very nature a leaky system, as editors will consider whether some image is filtered for some in their decisions whether to include it to begin with (add a, to a minority, "controversial" image to an article that lacks controversial images and likely get reverted, for instance). -- Björn Höhrmann · mailto:bjoern@hoehrmann.de · http://bjoern.hoehrmann.de Am Badedeich 7 · Telefon: +49(0)160/4415681 · http://www.bjoernsworld.de 25899 Dagebüll · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/

Am 18.09.2011 02:45, schrieb Stephen Bain: You must be kidding me to describe that as "inherently neutral". You miss the point that articles are build upon reputable sources. Therefore the sources are the ones that state the different point of views. We quote that points and gather them. We only exclude viewpoints which did not pass an editorial process already. Categorizing the images is not the same as gathering viewpoints from sources. It is the viewpoint of the contributer. Given your argumentation we would only need to write about one opinion. Our opinion.

On Fri, Sep 16, 2011 at 07:13:26PM +0200, Milos Rancic wrote: Thanks for throwing us en.wikipedians under the bus dude! ;-) Might be nice to wait with enwiki implementation too until we have something that we know will not have Evil Repercussions. Else we end up forking enwiki, if my spidey^Wwiki senses don't betray me. O:-) sincerly, Kim Bruning -- I won't shove Evil Images down your throat, if you don't shove Evil Cats down mine. :-)

On 17 September 2011 09:17, Jussi-Ville Heiskanen &lt;cimonavaro(a)gmail.com&gt; wrote: We need people to try the technical basics of a fork, i.e. taking an en:wp dump, an images dump, a copy of MediaWiki and relevant extensions and making it functional and documenting the procedure. - d.

On 17 September 2011 10:16, John Vandenberg &lt;jayvdb(a)gmail.com&gt; wrote: If there isn't, there should be. (I'm now trying to work out how to get the images without using up all my bandwidth allowances ever.) - d.

David Gerard wrote: It's easy enough to get a VPS with unlimited bandwidth. It's a few terabytes of data, though, depending on what you're talking about. Thumbnails, current images, older versions of images, deleted images, math renderings, etc. The sanest solution probably involves mailing a hard drive to someone and then having them mail it back. MZMcBride

On Sat, Sep 17, 2011 at 03:35:44PM +0100, Peter Coombe wrote: I think we want to investigate forking commons too. If any of you fine folks could use a vps, what specs do you need? I prolly have a little space to make one. :-) Debian ok for starters? I'm not sure forking is a good idea right here, right now, but it's always useful to check that failing to an external backup is still tractable. sincerely, Kim Bruning --

That is an legal issue. We do that to comply with the law, since that image isn't in public domain under German jurisdiction. This has nothing to do with hiding perfectly legal content. Additionally an optional filter would not help to make it legal. Filter or no filter wouldn't change a thing. Two different topics, one wrong assumption. Tobias Am 16.09.2011 12:39, schrieb emijrp:

Am 16.09.2011 12:57, schrieb emijrp: It's called "Schutzlandprinzip" [1]. Sorry that there is no English article about this topic as a whole. But you might read * http://en.wikipedia.org/wiki/Berne_Convention_for_the_Protection_of_Literar… * http://en.wikipedia.org/wiki/Agreement_on_Trade-Related_Aspects_of_Intellec… * http://en.wikipedia.org/wiki/World_Intellectual_Property_Organization_Copyr… * http://en.wikipedia.org/wiki/Paris_Convention_for_the_Protection_of_Industr… [1] http://de.wikipedia.org/wiki/Schutzlandprinzip

You miss the point that German contributers are bound to German law. They have to play along the German- or DACH (German, Austria, Switzerland) law, if they don't want to be in personal trouble. Since we don't want the provider, the WMF, to be in trouble we also comply to US-law as well. Thats why we are bound to comply to both sets laws. For instance we can't rely on fair use, since there is no fair use in Germany. We have exceptions for press (we aren't) and for educational purpose (we might be, but not always). Additionally the project decided to enforce free content over content with very tight rational. Am 16.09.2011 20:04, schrieb Ryan Kaldari:

Your stupid! Am 16.09.2011 12:39, schrieb emijrp:

Very simple: There is the Schutzlandprinzip. [1] To bad that EN has no article for it. But you can read the following articles to understand why this example was stupid, and that we bound to this laws, even if the servers are at nirvana: * http://en.wikipedia.org/wiki/Berne_Convention_for_the_Protection_of_Literar… * http://en.wikipedia.org/wiki/Agreement_on_Trade-Related_Aspects_of_Intellec… * http://en.wikipedia.org/wiki/World_Intellectual_Property_Organization_Copyr… * http://en.wikipedia.org/wiki/Paris_Convention_for_the_Protection_of_Industr… [1] http://de.wikipedia.org/wiki/Schutzlandprinzip Am 16.09.2011 13:01, schrieb Strainu:

German laws matter for german citizend. It's still not clear what that means for the rest of us (except that de.wp took a decision) 2011/9/16 emijrp &lt;emijrp(a)gmail.com&gt;om>:

Direct this question at Commons. Why are images deleted that are perfectly legal after US-law but not after DACH-(German, Austria, Switzerland)-law _and_ the other way around? Ask them, and you will get your answer. I mention this, because it is common practice, that you might not heard about. You really should try to understand the matter of copyright and that it has nothing to do with _self_ censoring or the image filter. We (the German authors) are bound to German law, since the page is directed at an major German readership (Schutzlandprinzip). But we also have to take care of US-law, since the servers are hosted inside the US. The later applies for the content the WMF hosts, not for us German contributers. But to be nice, we consider US-law as well. Tobias Am 16.09.2011 13:22, schrieb emijrp:

Am 16.09.2011 13:50, schrieb Strainu: I call this example "stupid", since it was constructed to make a point, even so it was lack of knowledge to propose it that way and it creates an awkward subtopic with no relation to the initial problem (see headline) whatsoever. This was the blatantly assumption/imputation that the image filter would be fine, since we already self censor ourself; which isn't true. We just respect the copyright law, to protect the readers/reusers. (Print a book with that image and publish it inside Germany. If noticed by the original author/photographer you might be in trouble). Tobias

On Fri, Sep 16, 2011 at 3:22 PM, emijrp &lt;emijrp(a)gmail.com&gt; wrote: Most are. And if German Wikipedia community believes that in order for de.wp to be more useful it should be made redistributable under the law of most German-speaking countries, that would be a reasonable choice for them. --vvv