While surveys show that a small majority finds this option (marginally) acceptable, current best analysis suggests that this particular option may not be implementable within the intersecting frames of the wikimedia movement and the laws of physics[1].

David Gerard, 21/09/2011 14:00: Not to mention that they've proved the Peano axioms to be inconsistent, probably making Kim's note pure nonsense: Kim Bruning, 21/09/2011 12:41:

On Wed, Sep 21, 2011 at 5:00 AM, David Gerard<dgerard(a)gmail.com> wrote: I don't have any unicorns, but there are lots of ponies. I'd be happy to stick a horn on one and call her sparkles if that would help? User rating / categorization systems are like ponies. They are a familiar and commonplace way of organizing things. They can be used to filter some things and reduce the degree of surprise; however they will always have both a large false positive rate and a large false negative rate. No filter is going to fly or shit rainbows. The question is not where to find mythical beasts, but whether dressing up a horse so that it looks a little like a unicorn would actually be useful. And that depends on whether there is actual demand for such filters, and whether having a filter that is sort-of-okay some of the time would be helpful to the people who want filtering. -Robert Rohde

On Wed, Sep 21, 2011 at 11:10, WereSpielChequers &lt;werespielchequers(a)gmail.com&gt; wrote: You didn't understand me well. It's not about fork(s), it's about wrappers, shells around the existing projects. * en.safe.wikipedia.org/wiki/<whatever> would point to en.wikipedia.org/wiki/<whatever> * When you click on "edit" from en.safe, you would get the same text as on en.wp. * When you click on "save" from en.safe, you would save the text on en.wp, as well. * The only difference is that images in wikitext won't be shown like [[File:<something sensible>.jpg]], but as [[File:fd37dae713526ee2da82f5a6cf6431de.jpg]]. * safe.wikimedia.org won't be Commons fork, but area for image categorization to those who want to work on it. It is not the job of Commons community to work on personal wishes of American right-wingers. (Note: "safe" is not good option for name, as it has four characters and it could be used for language editions of Wikipedia; maybe safe.en.wikipedia.org could be better option.)

On 21 September 2011 14:14, Jussi-Ville Heiskanen &lt;cimonavaro(a)gmail.com&gt; wrote: Look, the magical flying unicorn pony and the rainbows it shits have been specified, and considerable donors' money *will* be spent on the task, and that's all there is to it. The volunteers will just have to shape up and participate.

On 21 September 2011 15:43, Milos Rancic &lt;millosh(a)gmail.com&gt; wrote: Ah, you didn't read the bit of the report that specifically said a filtering mechanism should be on the main site, not an ignorable bolt-on. Though there wasn't a rationale.

They want that censoring tool and I think that they won't be content until they get it. Thus, let them have it and let them leave the rest of us alone.

On Wed, Sep 21, 2011 at 18:00, David Levy&lt;lifeisunfair(a)gmail.com&gt; wrote: Don't worry! Any implementation of censorship project would lead to endless troll-fests which would be more dumb than Youtube comments. The point is just to kick out them out of productive projects. Imagine a place where Christian, Muslim,<religon3>,<religionN> fundamentalists will have to cooperate! That would be the place for epic battles of dumbness. We'll have in-house circus!

Milos Rancic wrote: You're comfortable with the Wikimedia Foundation hosting/funding an "in-house circus"?

You choose discussions about images in a "circus" outside the context they belong to? This won't be "circus", since we just reduced the amount of arguments from some to zero. If combatants argue about a topic without having a word left, isn't this called a battlefield?

An alternative proposal (which is *far* simpler and maintains neutrality) already has the public backing of WMF trustee Samuel Klein:

Milos Rancic wrote:

> Don't worry! Any implementation of censorship project would lead to > endless troll-fests which would be more dumb than Youtube comments. > The point is just to kick out them out of productive projects. Imagine > a place where Christian, Muslim, <religon3>, <religionN> > fundamentalists will have to cooperate! That would be the place for > epic battles of dumbness. We'll have in-house circus!

You're comfortable with the Wikimedia Foundation hosting/funding an "in-house circus"?

On Wed, Sep 21, 2011 at 4:06 PM, Milos Rancic &lt;millosh(a)gmail.com&gt; wrote: The real problem here is that if there was a real market for stupid sites like that, they would already be there. And they are not, which does seem to point to the conclusion that there isn't a real market for such sites. Doesn't it?

-- -- Jussi-Ville Heiskanen, ~ [[User:Cimon Avaro]] _______________________________________________ foundation-l mailing list foundation-l(a)lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l

The real problem here is that if there was a real market for stupid sites like that, they would already be there. And they are not, which does seem to point to the conclusion that there isn't a real market for such sites. Doesn't it?

On 21 September 2011 14:14, Jussi-Ville Heiskanen&lt;cimonavaro(a)gmail.com&gt;  wrote: Not really. There are basically no major WP-derivative sites of any kind in existence - the ones that exist are either plain dumps studded with ads, or very small-scale attempts to do something good and innovative. As far as I can tell, it's just very hard to get a fork or a significantly different derivative site up and running successfully; it requires a large investment on fairly speculative predictions. Given this, it's hard to say that the absence of a particular kind of derivative site is due to there being a lack of demand for that *kind* of site - there might be demand, there might not, we just can't tell from the available evidence. (To steal David's analogy, it's a bit like saying that unicorns can't be trained, as there are no trained unicorns. Of course, there are no unicorns at all, and their trainability is moot...)

And where would the problem be? If a user prefers to go to a Bowdlerised site like that, rather than wikipedia.org, where they will see the pictures unless they specifically ask not to see them, then that is their choice, and no skin off our noses. A.

How odd, checking Tobias' list, I tried http://www.safesearchkids.com/wikipedia-for-kids.html to look for "penis" and it recommended [[File:Male erect penis.jpg]] as the second match. I was expecting it to restrict me to the more rounded and educational encyclopaedia entries, not straight to the most challenging images without context. If the WMF were to "recommend" such a solution for schools or religious groups, we might run into some immediate complaints.

... I believe Saudi Arabia has sporadically blocked access to Wikipedia, and blocks access to porn sites at the Internet service provider level: http://en.wikipedia.org/wiki/Censorship_in_Saudi_Arabia http://www.andrewlih.com/blog/2006/07/27/wikipedia-blocked-in-saudi-arabia/

Tobias, That is not quite what I thought we were talking about, because these are set-ups made on an individual computer, rather than restrictions at the internet service provider level. For example, I would not have a problem with it if schools figured out a way to prevent access to controversial images on school computers. I might have a problem with it if no one in an entire country were able to view these images; hence my question. I thought that was what you were talking about. If there are countries/Internet service providers that restrict all of their citizens from accessing porn sites, searching for adult images on Flickr, or prevent them from performing Google searches with safe search switched off, then it would be reasonable to assume that they might make an effort to do the same for Wikipedia. There was a similar situation in Germany, when Flickr prevented all German users with a yahoo.de address from accessing adult Flickr material, because Germany has unusually strict youth protection and age verification laws. http://en.wikipedia.org/wiki/Flickr#Controversy However, that was done by the company itself, because they wanted to avoid legal liability in Germany, and not by German Internet service providers. People in Germany with a yahoo.com (rather than yahoo.de) e-mail address were still perfectly able to access adult Flickr material from within Germany, using German internet service providers. I believe Saudi Arabia has sporadically blocked access to Wikipedia, and blocks access to porn sites at the Internet service provider level: http://en.wikipedia.org/wiki/Censorship_in_Saudi_Arabiahttp://www.andrewlih… Wikipedia was also briefly blocked in Pakistan, because of the Mohammed cartoon controversy. So there might be a scenario where countries like Saudi Arabia and Pakistan figure out how to block access to adult images and images of Mohammed on Wikipedia permanently, using methods like the ones you describe, based on the personal image filter categories. That might be a concern worth talking about. Of course, it has to be balanced against the concern that these countries can block Wikipedia altogether. Regards,Andreas --- On Fri, 23/9/11, Tobias Oelgarte&lt;tobias.oelgarte(a)googlemail.com&gt; wrote: From: Tobias Oelgarte&lt;tobias.oelgarte(a)googlemail.com&gt; Subject: Re: [Foundation-l] Possible solution for image filter To: foundation-l(a)lists.wikimedia.org Date: Friday, 23 September, 2011, 8:33 Yes we are aware of such pages. Just search for "google safe version" and so on. At first you will find plugins from Google for browsers itself, that can be used to enable the filter as an default option. If you scroll down a bit, then you will find other pages that are using Google to perform so called "safe searches".[1] There is a room for such tools.[2] Google limited it somewhat by providing the feature trough browser plugins itself. But you still find many examples for such pages.[3] There is already a market for such tools. First someone could check them out to see if we really need to do categorization or if this software is already good enough. Secondly it's nearly a proven that we would make an addition to that market. [1] For example: http://www.uk.safesearchlive.com/ http://www.safesearchkids.com/wikipedia-for-kids.html (Interestingly it does safe-search for Wikipedia trough Googles image categorization) [2] https://addons.mozilla.org/de/firefox/addon/linkextend-safety-kidsafe-site/… Plugin for firefox that removes even the buttons to disable "safe search" from google pages. [3] Many Anti-Virus software includes googles "safe search" functionality http://forum.kaspersky.com/lofiversion/index.php/t145285.html ... Am 23.09.2011 02:46, schrieb Andreas Kolbe: _______________________________________________ foundation-l mailing list foundation-l(a)lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l _______________________________________________ foundation-l mailing list foundation-l(a)lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l

As I see it, if the personal image filter categories can be exploited by censors to restrict image access permanently and irrevocably, this could result in the following scenarios: 1. A country or ISP that currently does not censor access to Wikipedia switches to access without the categorised images, removing choice from users (net loss for free access to information; this might extend even to basic anatomical images of vulvas, penises etc.). 2. A country or ISP that currently blocks access to Wikipedia completely makes Wikipedia available again, but without access to the images covered by the personal image filter categories (net gain for free access to information). 3. A country or ISP that currently blocks access to all Wikimedia images restores access to all images outside the personal image filter categories (net gain for free access to information, but it would be useful to have confirmation as to how many ISPs currently block all Wikimedia images -- at the moment we only have an unsourced statement in http://en.wikipedia.org/w/index.php?title=List_of_websites_blocked_in_the_P… claiming that some Chinese ISPs do this). 4. A country or ISP that currently blocks access to Wikipedia completely, or currently blocks access to Wikimedia images globally, restores access, using the personal image filter as designed, i.e. leaving it at the user's discretion (net gain for free access to information, but I agree with you that this scenario is rather unlikely). We clearly should not assume that these net gains or net losses are all equal in magnitude, or that all these scenarios would be equally likely. We should also remember that this only addresses the consequences of countries or providers using the personal image filter categories in the way that you have warned would be possible, i.e. for complete censorship of these images. Such use of the categories for outright censorship is an important part of the picture, but it's not the whole picture, as there is also the perceived benefit of the personal image filter when it works as designed (i.e. giving the user a choice they don't have right now). Still, these are important matters to think about. I like the personal image filter idea as designed, but I'd be uncomfortable with 50 countries, say, using the opportunity to implement scenario 1. Andreas --- On Fri, 23/9/11, Tobias Oelgarte&lt;tobias.oelgarte(a)googlemail.com&gt; wrote: From: Tobias Oelgarte&lt;tobias.oelgarte(a)googlemail.com&gt; Subject: Re: [Foundation-l] Possible solution for image filter To: foundation-l(a)lists.wikimedia.org Date: Friday, 23 September, 2011, 12:03 I gave you a simple example on how easy it would be to use our categorization to implement a filter based upon those categories. The sources on that this actually happens are not rare if we look at china or Iran. The problem are many local providers over which you will seldom find a report. Many third world Internet users are bound to use a single local provider or the access depends at an organization. You said that we have to concern the point, that Wikipedia might be blocked entirely if we don't have such a feature. This argument is weakend by the fact that the filter (as intended) can just be ignored by user. This rises the doubt, that the feature would be strong enough for "censors needs" and therefore might not be reason against blocking Wikipedia completely. But lets also think the other way around. Many of this "potential censors" aren't blocking Wikipedia entirely since this would most likely result in pressure against the decision to "take down" Wikipedia. Blocking only selected content is the way censors prefer. It is done in a much greater amount of countries. For example even in Taiwan or South Korea. If we provide the categories then this is exactly one of the things what could be used to extend censorship without the pressure to take down Wikipedia entirely. It is much more acceptable. An option that is not present at the moment. To be fair: We have no numbers on that. It is speculation and it might go the one way or the other way. But should we take that risk? Currently we are promoting free access to information and knowledge. If a filter like this has a 50:50 chance to improve or worsen things, then we might raise the question: Is it worth the effort or should we search for better solutions? Greetings Tobias Am 23.09.2011 12:38, schrieb Andreas Kolbe: _______________________________________________ foundation-l mailing list foundation-l(a)lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l _______________________________________________ foundation-l mailing list foundation-l(a)lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l

Wikipedia was also briefly blocked in Pakistan, because of the Mohammed cartoon controversy. So there might be a scenario where countries like Saudi Arabia and Pakistan figure out how to block access to adult images and images of Mohammed on Wikipedia permanently, using methods like the ones you describe, based on the personal image filter categories.? That might be a concern worth talking about.

Of course, it has to be balanced against the concern that these countries can block Wikipedia altogether.

On Fri, Sep 23, 2011 at 2:38 PM, Kim Bruning &lt;kim(a)bruning.xs4all.nl&gt; wrote: Funny how nobody is mentioning the Virgin Killer farce. The organisation that made that happen might just ask British ISP's to *silently* use the filter-tags to prevent British viewers to ever know there was such tagged content at all. And just to prove I am not being anti-British here, the Finnish police did even worse things, but the margin of this e-mail is too narrow to explain them.

On 24 September 2011 23:00, Phil Nash&lt;phnash(a)blueyonder.co.uk&gt; wrote: The IWF situation is analogous to the present one. The government of the day called for "something to be done!" (the requirement spec for magical flying unicorn ponies), and the ISPs nodded and smiled and set up something called the IWF that pretended to supply magical flying unicorn ponies. And everyone was happy. That is, the IWF is the sort of organisation that can exist only as long as it doesn't affect people's lives, e.g. doesn't hit the headlines. What happened then was that the IWF had a rush of blood to the head, misunderstood their purpose (to pretend to do something impossible) and thought they needed to actually do something to implement the magical flying unicorn ponies requirement. So they blocked Wikipedia. And everyone noticed. And now they are well-known and are widely regarded as dangerous cretins. A cautionary tale for those hoping to implement filters. - d.

On Wed, Sep 21, 2011 at 11:06 PM, Milos Rancic &lt;millosh(a)gmail.com&gt; wrote: Well, while we're tarring large groups of people with the same brush, it's not their job to bend to the desires of European anarcho-libertarians either.

You didn't understand me well. It's not about fork(s), it's about wrappers, shells around the existing projects. * en.safe.wikipedia.org/wiki/<whatever> would point to en.wikipedia.org/wiki/<whatever> * When you click on "edit" from en.safe, you would get the same text as on en.wp. * When you click on "save" from en.safe, you would save the text on en.wp, as well. * The only difference is that images in wikitext won't be shown like [[File:<something sensible>.jpg]], but as [[File:fd37dae713526ee2da82f5a6cf6431de.jpg]]. * safe.wikimedia.org won't be Commons fork, but area for image categorization to those who want to work on it. It is not the job of Commons community to work on personal wishes of American right-wingers. (Note: "safe" is not good option for name, as it has four characters and it could be used for language editions of Wikipedia; maybe safe.en.wikipedia.org could be better option.)

I still can't the a rational difference between images included in articles by the will of the community and text passages included by the will of the community.

It's much easier to note offensive text fragments before reading them than to note offensive images before seeing them. But I guess the more fundamental issue is: there are, I assume, people who have requested this feature for images. There are either no or only very few who have requested it for text.

Beyond the vandalism problem, I have dealt with complaints on OTRS relating to weight, such as emphasis on stories about paedophiles on school articles. However it is hard to imagine how a filter would deal with this and I suspect the majority of our community would not want to start hiding paragraphs that include difficult words such as "paedophile", particularly when they are likely to be accurate and verifiable even for articles likely to be accessed by young readers.

Truthfully, i see not different approach to include images and text passages. Both are added, discussed, removed, re-added the same way as text is. Now i heard some say that text is written by multiple authors and images are only created by one. Then i must wonder that we are able to decide to include one source and it's arguments written by one author, while it seams to be a problem to include the image of one photographer/artist. There really is no difference in overall progress.

What is the advantage of that compared with the feature as it was originally proposed? All you've done is made the URL more complicated. You'll still need to use user preferences to determine which images are getting hidden, so why can't you just have an "on/off" user preference as well rather than determining whether the filter should be on or off based on the URL?

On Wed, Sep 21, 2011 at 19:10, Thomas Dalton &lt;thomas.dalton(a)gmail.com&gt; wrote: * People should have possibility to choose the set of images which they don't want to see.

* As it's not the main site, but wrapper, it could have turned off images offensive to anyone, so everybody would be able to see the site without having to log in. It could lead to "no images" by default, but that's not my problem.

* They could experiment, as nobody would care about the site. As Tobias mentioned below, if some text is offensive to someone, they could add it into the filter.

* Most importantly, that won't affect anything else. Except, probably, ~$1M/year of WMF budget for development of censorship software and censorship itself, as they will say that they lack of people to censor images and that they need employees to do that. Although it would be more useful to give that ~$1M/year for access to Wikipedia from African countries, I think that it's reasonable price for having people who want censorship content. Bottom line is that News Corp will pay all of that and much more by giving us free access to Fox News.