I get the idea that there are theoretical reasons why image filters can't work, and I share the view that the proposal which was consulted on needs some improvement. An individual choice made at the IP level was a circle that looked awfully difficult to square.
But since Flickr has already proven that something like this can work in practice, can we agree to classify Image filters as one of those things that work in practice but not in theory? Then we can concentrate on the practical issue of if we decide to implement this, how do we do it better than Flickr has?
NB I would not want us to implement this the way Flickr has http://www.flickr.com/help/filters/#258 And not only because I'm not totally convinced that our community would share their view that Germany is the country that needs the tightest restrictions.
Hugs
WereSpielChequers
PS My niece absolutely wants that magical flying unicorn pony for the winter solstice, especially if it s***s rainbows. Would you mind telling me where I can order one?
Message: 7 Date: Wed, 21 Sep 2011 13:00:18 +0100 From: David Gerard dgerard@gmail.com Subject: Re: [Foundation-l] Possible solution for image filter To: Wikimedia Foundation Mailing List foundation-l@lists.wikimedia.org Message-ID: <CAJ0tu1H+haRDH_Hjj_tvmbsL8bxu+h=h1JhV0L3dZEkSS84_JQ@mail.gmail.com
Content-Type: text/plain; charset=UTF-8
On 21 September 2011 11:41, Kim Bruning kim@bruning.xs4all.nl wrote:
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].
The board resolution specifies a magical flying unicorn pony that shits rainbows. A wide-ranging survey has been conducted on the precise flight patterns and the importance of which way round the rainbow spectrum goes. These tiresome people who keep calling this "impossible" just do not understand that the high-level decision for a magical flying unicorn pony that shits rainbows has been set in stone.
- d.
WSC, I'm not sure Flickr is a great example. The restricted images policies only work because they employ a full time set of "police officers" with the massively unpopular duty to delete images and close accounts (including those that have paid in advance for their "pro" account) using their own personal judgement using a classification system that has no clear definition, independent appeal or assessment process.
WMF *could* force Commons to become such a policed service (perhaps by paying admins a salary; yea where can I sign up?) but we would probably not consider it part of an open community from that point on.
Cheers, Fae
On 21 September 2011 16:37, WereSpielChequers werespielchequers@gmail.com wrote:
But since Flickr has already proven that something like this can work in practice, can we agree to classify Image filters as one of those things that work in practice but not in theory? Then we can concentrate on the practical issue of if we decide to implement this, how do we do it better than Flickr has?
They do it by crowdsourcing a mass American bias, don't they?
An American POV being enforced strikes me as a problematic solution.
(I know that FAQ says "global community". What they mean is "people all around the world who are Silicon Valley technologists like us - you know, normal people." This approach also has a number of fairly obvious problems.)
PS My niece absolutely wants that magical flying unicorn pony for the winter solstice, especially if it s***s rainbows. Would you mind telling me where I can order one?
I suggest you pass a board resolution demanding one. When people protest, tell them to just shut up and do their jobs.
- d.
On 21 September 2011 16:53, David Gerard dgerard@gmail.com wrote:
They do it by crowdsourcing a mass American bias, don't they?
An American POV being enforced strikes me as a problematic solution.
(I know that FAQ says "global community". What they mean is "people all around the world who are Silicon Valley technologists like us - you know, normal people." This approach also has a number of fairly obvious problems.)
I mentioned this a couple of weeks ago, I think, but this effect cuts both ways.
We already know that our community skews to - as you put it - "people all around the world who are technologists like us". As a result, that same community is who decides what images are reasonable and appropriate to put in articles.
People look at images and say - yes, it's appropriate, yes, it's encyclopedic, no, it's excessively violent, no, that's gratuitous nudity, yes, I like kittens, etc etc etc. You do it, I do it, we try to be sensible, but we're not universally representative. The community, over time, imposes its own de facto standards on the content, and those standards are those of - well, we know what our systemic biases are. We've not managed a quick fix to that problem, not yet.
One of the problems with the discussions about the image filter is that many of them argue - I paraphrase - that "Wikipedia must not be censored because it would stop being neutral". But is the existing "Wikipedian POV" *really* the same as "neutral", or are we letting our aspirations to inclusive global neutrality win out over the real state of affairs? It's the great big unexamined assumption in our discussions...
... many of them argue - I paraphrase - that "Wikipedia must not be censored because it would stop being neutral". ...
- Andrew Gray
Please categorize anyone who thinks the projects are anything like "neutral" as a <redacted>. One need only do, say, a simple count of photos of girls in skimpy bikinis versus lads in skimpy speedos that happen to illustrate articles, to come to an obvious conclusion that consensus of the majority is not the same thing as neutrality. Though I am not supportive of filtering, I agree that however you argue it, this point of view has little to do with the policy issues.
Cheers, Fae
Am 21.09.2011 18:41, schrieb Andrew Gray:
On 21 September 2011 16:53, David Gerarddgerard@gmail.com wrote:
They do it by crowdsourcing a mass American bias, don't they?
An American POV being enforced strikes me as a problematic solution.
(I know that FAQ says "global community". What they mean is "people all around the world who are Silicon Valley technologists like us - you know, normal people." This approach also has a number of fairly obvious problems.)
I mentioned this a couple of weeks ago, I think, but this effect cuts both ways.
We already know that our community skews to - as you put it - "people all around the world who are technologists like us". As a result, that same community is who decides what images are reasonable and appropriate to put in articles.
People look at images and say - yes, it's appropriate, yes, it's encyclopedic, no, it's excessively violent, no, that's gratuitous nudity, yes, I like kittens, etc etc etc. You do it, I do it, we try to be sensible, but we're not universally representative. The community, over time, imposes its own de facto standards on the content, and those standards are those of - well, we know what our systemic biases are. We've not managed a quick fix to that problem, not yet.
One of the problems with the discussions about the image filter is that many of them argue - I paraphrase - that "Wikipedia must not be censored because it would stop being neutral". But is the existing "Wikipedian POV" *really* the same as "neutral", or are we letting our aspirations to inclusive global neutrality win out over the real state of affairs? It's the great big unexamined assumption in our discussions...
You describe us as geeks and that we can't write in a way that would please the readers. Since we are geeks, we are strongly biased and write down POV all day. If that is true, why is Wikipedia such a success? Why people read it? Do they like geeky stuff?
Don't you think that we would have thousands of complaints a day if your words would be true at all? Just have a look at the article [[hentai]] and look at the illustration. How many complaints about this image do we get a day? None, because it is less then one complain in a month, while the article itself is viewed about 8.000 times a day.[1] That would make up one complainer in 240.000 (0,0004%). Now we could argue that only some of them would comment on the issue. Lets assume 1 of 100 or even 1 of 1000. Then it are still only 0,04% or 0,4%. That is the big mass of users we want to support get more contributers?
Am 21.09.2011 um 19:04 schrieb Tobias Oelgarte tobias.oelgarte@googlemail.com:
Don't you think that we would have thousands of complaints a day if your words would be true at all? Just have a look at the article [[hentai]] and look at the illustration. How many complaints about this image do we get a day? None, because it is less then one complain in a month, while the article itself is viewed about 8.000 times a day.[1] That would make up one complainer in 240.000 (0,0004%). Now we could argue that only some of them would comment on the issue. Lets assume 1 of 100 or even 1 of 1000. Then it are still only 0,04% or 0,4%. That is the big mass of users we want to support get more contributers?
Your assumtion is wrong. The 8.000 daily are neither neutral nor representative for all users. Put the picture on the main page and You get representative results. We had that in Germany.
* Kanzlei wrote:
Your assumtion is wrong. The 8.000 daily are neither neutral nor representative for all users. Put the picture on the main page and You get representative results. We had that in Germany.
That's missing the point. Putting an image on the front page is putting it out of context, so you get complaints about an image appearing there where people do not mind the image appearing in the article, and people do not get to decide whether to open an article that might feature some content they are uncomfortable with and consequently do have their mind in a "I do not know what to expect" state.
Am 21.09.2011 19:36, schrieb Kanzlei:
Am 21.09.2011 um 19:04 schrieb Tobias Oelgartetobias.oelgarte@googlemail.com:
Don't you think that we would have thousands of complaints a day if your words would be true at all? Just have a look at the article [[hentai]] and look at the illustration. How many complaints about this image do we get a day? None, because it is less then one complain in a month, while the article itself is viewed about 8.000 times a day.[1] That would make up one complainer in 240.000 (0,0004%). Now we could argue that only some of them would comment on the issue. Lets assume 1 of 100 or even 1 of 1000. Then it are still only 0,04% or 0,4%. That is the big mass of users we want to support get more contributers?
Your assumtion is wrong. The 8.000 daily are neither neutral nor representative for all users. Put the picture on the main page and You get representative results. We had that in Germany.
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
You said that my assumption is wrong. We can repeat this for hundreds of articles and you would get the same result. Now proof that this assumption, which is sourced (just look at it) is wrong or say what is wrong with my assumption (in detail).
On 21 September 2011 11:10, Tobias Oelgarte tobias.oelgarte@googlemail.com wrote:
Am 21.09.2011 19:36, schrieb Kanzlei:
Am 21.09.2011 um 19:04 schrieb Tobias Oelgartetobias.oelgarte@googlemail.com:
Don't you think that we would have thousands of complaints a day if your words would be true at all? Just have a look at the article [[hentai]] and look at the illustration. How many complaints about this image do we get a day? None, because it is less then one complain in a month, while the article itself is viewed about 8.000 times a day.[1] That would make up one complainer in 240.000 (0,0004%). Now we could argue that only some of them would comment on the issue. Lets assume 1 of 100 or even 1 of 1000. Then it are still only 0,04% or 0,4%. That is the big mass of users we want to support get more contributers?
Your assumtion is wrong. The 8.000 daily are neither neutral nor representative for all users. Put the picture on the main page and You get representative results. We had that in Germany.
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
Can you point me towards that poll?
Thanks, Sue
* Sue Gardner wrote:
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
Can you point me towards that poll?
http://de.wikipedia.org/wiki/Wikipedia:Meinungsbilder/Beschr%C3%A4nkung_der_...
On 21 September 2011 12:37, Bjoern Hoehrmann derhoermi@gmx.net wrote:
- Sue Gardner wrote:
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
Can you point me towards that poll?
http://de.wikipedia.org/wiki/Wikipedia:Meinungsbilder/Beschr%C3%A4nkung_der_...
Thanks, Björn. That's so interesting: I hadn't known about that poll.
Can someone help me understand the implications of it?
Does it mean basically this: deWP put the Vulva article on its front page, and then held a poll to decide whether to i) stop putting articles like Vulva on its front page, because they might surprise or shock some readers, or ii) continue putting articles like Vulva on the front page, regardless of whether they surprise or shock some readers. And the voted supported the latter.
If I've got that right, I assume it means that policy on the German Wikipedia today would support putting Vulva on the main page. Is there an 'element of least surprise' type policy or convention that would be considered germane to this, or not?
I'd be grateful too if anyone would point me towards the page that delineates the process for selecting the Article of the Day. I can read pages in languages other than English (sort of) using Google Translate, but I have a tough time actually finding them :-)
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!
Am 21.09.2011 21:52, schrieb Sue Gardner:
On 21 September 2011 12:37, Bjoern Hoehrmannderhoermi@gmx.net wrote:
- Sue Gardner wrote:
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
Can you point me towards that poll?
http://de.wikipedia.org/wiki/Wikipedia:Meinungsbilder/Beschr%C3%A4nkung_der_...
Thanks, Björn. That's so interesting: I hadn't known about that poll.
Can someone help me understand the implications of it?
Does it mean basically this: deWP put the Vulva article on its front page, and then held a poll to decide whether to i) stop putting articles like Vulva on its front page, because they might surprise or shock some readers, or ii) continue putting articles like Vulva on the front page, regardless of whether they surprise or shock some readers. And the voted supported the latter.
If I've got that right, I assume it means that policy on the German Wikipedia today would support putting Vulva on the main page. Is there an 'element of least surprise' type policy or convention that would be considered germane to this, or not?
I'd be grateful too if anyone would point me towards the page that delineates the process for selecting the Article of the Day. I can read pages in languages other than English (sort of) using Google Translate, but I have a tough time actually finding them :-)
Thanks, Sue
At first we had some basic discussion which topic might be suitable for the main page. That was the offspring for idea to put the excellent article "vulva" together with a depiction (photograph) on the main page to see what would be the reaction. There was quite some reaction, but not so much as we expected. The opinions where fairly balanced. After some other topics with "may be objectionable content" followed in the meantime the discussion was going forward, leading to the decision (initiated by a group of users who opposed that every topic should be treated equally) to create a Meinungsbild (the linked one). The result was very clear and one of the main arguments where: "How do we draw a line between objectionable and not objectionable content, without violating NPOV?"
After that we did not represent one shocking article after the other. We just let them come and if the article itself is well written he will have it's chance to be put on the main page (it has to be an "excellent" or "worth reading" article, after the article quality rating system [1]) . The decision will be made in an open progress (even so it looks like a poll, it isn't) found at: http://de.wikipedia.org/wiki/Wikipedia_Diskussion:Hauptseite/Artikel_des_Tag...
[1] http://de.wikipedia.org/wiki/Wikipedia:Kandidaturen_von_Artikeln,_Listen_und...
* Sue Gardner wrote:
Does it mean basically this: deWP put the Vulva article on its front page, and then held a poll to decide whether to i) stop putting articles like Vulva on its front page, because they might surprise or shock some readers, or ii) continue putting articles like Vulva on the front page, regardless of whether they surprise or shock some readers. And the voted supported the latter.
The poll asked whether there should be formalized restrictions beyond the existing ones (only good articles can be proposed). Voters decided against that and to keep the status quo instead where it is decided on a case-by-case basis which articles to feature on the main page without additional formalized selection criteria that would disqualify certain articles. Put differently, they decided that if someone disagress that a certain article should not be featured, they cannot point to policy to support their argument.
If I've got that right, I assume it means that policy on the German Wikipedia today would support putting Vulva on the main page. Is there an 'element of least surprise' type policy or convention that would be considered germane to this, or not?
Among editors who bothered to participate in the process, featuring the article at all was not particularily controversial, but there was a rather drawn out discussion about which, if any, image to use. I have read much of the feedback at the time and my impression is that this was not very different among "readers", most complaints were about the image they had picked (and possibly some about images in the article itself).
Keep in mind that continental europe's attitude towards sex is quite different than north america's. I read this the other day and found it quite illustrative, "While nine out of 10 Dutch parents had allowed or would consider sleepovers once the child was 16 or 17, nine out of 10 American parents were adamant: “not under my roof.”".
I'd be grateful too if anyone would point me towards the page that delineates the process for selecting the Article of the Day. I can read pages in languages other than English (sort of) using Google Translate, but I have a tough time actually finding them :-)
http://de.wikipedia.org/wiki/WD:Hauptseite/Artikel_des_Tages
Am 21.09.2011 23:53, schrieb Bjoern Hoehrmann:
- Sue Gardner wrote:
Does it mean basically this: deWP put the Vulva article on its front page, and then held a poll to decide whether to i) stop putting articles like Vulva on its front page, because they might surprise or shock some readers, or ii) continue putting articles like Vulva on the front page, regardless of whether they surprise or shock some readers. And the voted supported the latter.
The poll asked whether there should be formalized restrictions beyond the existing ones (only good articles can be proposed). Voters decided against that and to keep the status quo instead where it is decided on a case-by-case basis which articles to feature on the main page without additional formalized selection criteria that would disqualify certain articles. Put differently, they decided that if someone disagress that a certain article should not be featured, they cannot point to policy to support their argument.
That isn't true. Since the policy states that all terms are treated equal (NPOV) there is only a discussion if the date might be suitable (topics with correlation to a certain date get precedence). Otherwise it is decided if the quality (actuality and so on) is suitable for AotD, since there might be a lot of time between the last nomination for good articles and the versions might differ strongly due to recent changes. If a topic is offensive or not does not play any role. Only quality matters. This rule existed from the beginning and it did not change.
If I've got that right, I assume it means that policy on the German Wikipedia today would support putting Vulva on the main page. Is there an 'element of least surprise' type policy or convention that would be considered germane to this, or not?
Among editors who bothered to participate in the process, featuring the article at all was not particularily controversial, but there was a rather drawn out discussion about which, if any, image to use. I have read much of the feedback at the time and my impression is that this was not very different among "readers", most complaints were about the image they had picked (and possibly some about images in the article itself).
Keep in mind that continental europe's attitude towards sex is quite different than north america's. I read this the other day and found it quite illustrative, "While nine out of 10 Dutch parents had allowed or would consider sleepovers once the child was 16 or 17, nine out of 10 American parents were adamant: “not under my roof.”".
That illustrates very well why the german community would not share the same view. Additionally it clarifies that a global approach for filtering isn't possible to be implemented the right way. We really put something like ice and fire in the same box and want them to come to the same conclusion. It will just happen to be something like a battle. But a result, a compromise? Impossible by design.
I'd be grateful too if anyone would point me towards the page that delineates the process for selecting the Article of the Day. I can read pages in languages other than English (sort of) using Google Translate, but I have a tough time actually finding them :-)
http://de.wikipedia.org/wiki/WD:Hauptseite/Artikel_des_Tages
* Tobias Oelgarte wrote:
The poll asked whether there should be formalized restrictions beyond the existing ones (only good articles can be proposed). Voters decided against that and to keep the status quo instead where it is decided on a case-by-case basis which articles to feature on the main page without additional formalized selection criteria that would disqualify certain articles. Put differently, they decided that if someone disagress that a certain article should not be featured, they cannot point to policy to support their argument.
That isn't true. Since the policy states that all terms are treated equal (NPOV) there is only a discussion if the date might be suitable (topics with correlation to a certain date get precedence). Otherwise it is decided if the quality (actuality and so on) is suitable for AotD, since there might be a lot of time between the last nomination for good articles and the versions might differ strongly due to recent changes. If a topic is offensive or not does not play any role. Only quality matters. This rule existed from the beginning and it did not change.
What I meant to say is: "if someone disagrees with featuring a certain article, they cannot point to policy that restricts which subjects can be featured to support their argument" as there is none and editors de- cided against introducing any.
Am 22.09.2011 00:42, schrieb Bjoern Hoehrmann:
- Tobias Oelgarte wrote:
The poll asked whether there should be formalized restrictions beyond the existing ones (only good articles can be proposed). Voters decided against that and to keep the status quo instead where it is decided on a case-by-case basis which articles to feature on the main page without additional formalized selection criteria that would disqualify certain articles. Put differently, they decided that if someone disagress that a certain article should not be featured, they cannot point to policy to support their argument.
That isn't true. Since the policy states that all terms are treated equal (NPOV) there is only a discussion if the date might be suitable (topics with correlation to a certain date get precedence). Otherwise it is decided if the quality (actuality and so on) is suitable for AotD, since there might be a lot of time between the last nomination for good articles and the versions might differ strongly due to recent changes. If a topic is offensive or not does not play any role. Only quality matters. This rule existed from the beginning and it did not change.
What I meant to say is: "if someone disagrees with featuring a certain article, they cannot point to policy that restricts which subjects can be featured to support their argument" as there is none and editors de- cided against introducing any.
Now we speak the same language. Sorry if i misunderstood your first wording. ;-)
Am 21.09.2011 21:28, schrieb Sue Gardner:
On 21 September 2011 11:10, Tobias Oelgarte tobias.oelgarte@googlemail.com wrote:
Am 21.09.2011 19:36, schrieb Kanzlei:
Am 21.09.2011 um 19:04 schrieb Tobias Oelgartetobias.oelgarte@googlemail.com:
Don't you think that we would have thousands of complaints a day if your words would be true at all? Just have a look at the article [[hentai]] and look at the illustration. How many complaints about this image do we get a day? None, because it is less then one complain in a month, while the article itself is viewed about 8.000 times a day.[1] That would make up one complainer in 240.000 (0,0004%). Now we could argue that only some of them would comment on the issue. Lets assume 1 of 100 or even 1 of 1000. Then it are still only 0,04% or 0,4%. That is the big mass of users we want to support get more contributers?
Your assumtion is wrong. The 8.000 daily are neither neutral nor representative for all users. Put the picture on the main page and You get representative results. We had that in Germany.
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
Can you point me towards that poll?
Thanks, Sue
foundation-l mailing list foundation-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
Gladly. You will find it under: "Restrictions of topics for article of the day" http://de.wikipedia.org/wiki/Wikipedia:Meinungsbilder/Beschr%C3%A4nkung_der_...
It started some time after the "vulva" was presented at the main page. After the poll we even presented a topics like Futanari [1] on the main page at November 10th 2010 [2]. The reaction can be described with "no reaction at all". It was just as if it was any other article. Some left some praise at the discussion, some others made some corrections and so on. There simply wasn't such a thing as an uproar or any complaints. Now the article had 3k views a day and not one comment on removing images or something else since that date. Thats one of the reasons why I'm wondering if the "offensive image problem" is even exists, for the German Wikipedia. But if i look at the discussion pages at EN it's basically the same. There are more complaints, but also at least the triple amount of viewers per day.
[1] http://de.wikipedia.org/wiki/Futanari [2] http://de.wikipedia.org/wiki/Wikipedia:Hauptseite/Artikel_des_Tages/Zeittafe...
Tobias
Am 21.09.2011 um 20:10 schrieb Tobias Oelgarte tobias.oelgarte@googlemail.com:
Am 21.09.2011 19:36, schrieb Kanzlei:
Am 21.09.2011 um 19:04 schrieb Tobias Oelgartetobias.oelgarte@googlemail.com:
Don't you think that we would have thousands of complaints a day if your words would be true at all? Just have a look at the article [[hentai]] and look at the illustration. How many complaints about this image do we get a day? None, because it is less then one complain in a month, while the article itself is viewed about 8.000 times a day.[1] That would make up one complainer in 240.000 (0,0004%). Now we could argue that only some of them would comment on the issue. Lets assume 1 of 100 or even 1 of 1000. Then it are still only 0,04% or 0,4%. That is the big mass of users we want to support get more contributers?
Your assumtion is wrong. The 8.000 daily are neither neutral nor representative for all users. Put the picture on the main page and You get representative results. We had that in Germany.
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
This poll was not representative for wikipedia readers, but only for some German wikipedia editors. Scientifically research found that Germa editors are not representative for German speaking people but far more environmetal-liberal-leftists than avarage Germans. The poll was even not representative for German editors because only a few voted.
You said that my assumption is wrong. We can repeat this for hundreds of articles and you would get the same result. Now proof that this assumption, which is sourced (just look at it) is wrong or say what is wrong with my assumption (in detail).
See above
foundation-l mailing list foundation-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
On 21 September 2011 21:20, Kanzlei kanzlei@f-t-hofmann.de wrote:
This poll was not representative for wikipedia readers, but only for some German wikipedia editors. Scientifically research found that Germa editors are not representative for German speaking people but far more environmetal-liberal-leftists than avarage Germans. The poll was even not representative for German editors because only a few voted.
233 would be a *large* turnout on en:wp. What is a large turnout on de:wp?
Your arguments look to me like fully-general counterarguments against *any* on-Wikipedia poll whatsoever, no matter the structure or subject. What would you accept as a measure of the de:wp community that would actually be feasible to conduct?
- d.
Am 21.09.2011 22:37, schrieb David Gerard:
On 21 September 2011 21:20, Kanzleikanzlei@f-t-hofmann.de wrote:
This poll was not representative for wikipedia readers, but only for some German wikipedia editors. Scientifically research found that Germa editors are not representative for German speaking people but far more environmetal-liberal-leftists than avarage Germans. The poll was even not representative for German editors because only a few voted.
233 would be a *large* turnout on en:wp. What is a large turnout on de:wp?
Your arguments look to me like fully-general counterarguments against *any* on-Wikipedia poll whatsoever, no matter the structure or subject. What would you accept as a measure of the de:wp community that would actually be feasible to conduct?
- d.
A so called "Meinungsbild" (opinion poll) is the tool of choice to make basic decisions for the project. Admins and authors are bound to such decisions. It usually needs 2/3 of the users to agree with a proposal (formally correctness) and 2/3 of the users actually voting for and not against the proposal. There may be variations depending on the questioning.
On Wed, Sep 21, 2011 at 11:58 PM, Tobias Oelgarte tobias.oelgarte@googlemail.com wrote:
Am 21.09.2011 22:37, schrieb David Gerard:
On 21 September 2011 21:20, Kanzleikanzlei@f-t-hofmann.de wrote:
This poll was not representative for wikipedia readers, but only for some German wikipedia editors. Scientifically research found that Germa editors are not representative for German speaking people but far more environmetal-liberal-leftists than avarage Germans. The poll was even not representative for German editors because only a few voted.
233 would be a *large* turnout on en:wp. What is a large turnout on de:wp?
Your arguments look to me like fully-general counterarguments against *any* on-Wikipedia poll whatsoever, no matter the structure or subject. What would you accept as a measure of the de:wp community that would actually be feasible to conduct?
- d.
A so called "Meinungsbild" (opinion poll) is the tool of choice to make basic decisions for the project. Admins and authors are bound to such decisions. It usually needs 2/3 of the users to agree with a proposal (formally correctness) and 2/3 of the users actually voting for and not against the proposal. There may be variations depending on the questioning.
I am mildly amused -- and please accept this as an attempt to defuse the acrimony -- by the fact that the German vulva article makes no mention of a symbol of it being very commonly used to spoil ballots here in Finland. Losing only to a celebrated comic talent called "Spede" and to Donald Duck.
Connecting voting and vulvas...
* David Gerard wrote:
233 would be a *large* turnout on en:wp. What is a large turnout on de:wp?
Most Meinungsbilder have between 100 and 300 editors participating and the 300s are seen regularily. Participation maxes out at around 500 so "large" probably begins somewhere in the 300s. This largely matches the number of participants in admin elections, to offer a comparison.
Am 22.09.2011 05:15, schrieb Bjoern Hoehrmann:
- David Gerard wrote:
233 would be a *large* turnout on en:wp. What is a large turnout on de:wp?
Most Meinungsbilder have between 100 and 300 editors participating and the 300s are seen regularily. Participation maxes out at around 500 so "large" probably begins somewhere in the 300s. This largely matches the number of participants in admin elections, to offer a comparison.
You should took into account that this are open polls. One issue with open polls is participation. If a poll is on the edge (50:50 situation), you will always have much more votes then in a poll that looks already decided after a few days. Thats why polls which are going strongly in one direction usually have a lesser number of participants.
Am 21.09.2011 um 22:37 schrieb David Gerard dgerard@gmail.com:
On 21 September 2011 21:20, Kanzlei kanzlei@f-t-hofmann.de wrote:
This poll was not representative for wikipedia readers, but only for some German wikipedia editors. Scientifically research found that Germa editors are not representative for German speaking people but far more environmetal-liberal-leftists than avarage Germans. The poll was even not representative for German editors because only a few voted.
233 would be a *large* turnout on en:wp. What is a large turnout on de:wp?
Your arguments look to me like fully-general counterarguments against *any* on-Wikipedia poll whatsoever, no matter the structure or subject. What would you accept as a measure of the de:wp community that would actually be feasible to conduct?
233 is a large amount for a poll on de:wp. But it was no democratic poll, because the manner by which the poll was conducted was not democratic. A democratic and representative poll has to be equal, common and private. The poll was not common because not every user entitled to vote was noticed about the poll,
(example for a more democratic poll was the poll from the foundation in question bildfilter: it was on an anonymous server and I was notified by email that I was entitled to vote),
it was not private, because everybody can see who choose what. And finally it was not equal, because there was no means to exclude the possibility of sock puppet voting (Which is very common and very easy as far as I know - I know an unpunished such voting).
- d.
foundation-l mailing list foundation-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
Am 22.09.2011 08:07, schrieb Kanzlei:
Am 21.09.2011 um 22:37 schrieb David Gerarddgerard@gmail.com:
On 21 September 2011 21:20, Kanzleikanzlei@f-t-hofmann.de wrote:
This poll was not representative for wikipedia readers, but only for some German wikipedia editors. Scientifically research found that Germa editors are not representative for German speaking people but far more environmetal-liberal-leftists than avarage Germans. The poll was even not representative for German editors because only a few voted.
233 would be a *large* turnout on en:wp. What is a large turnout on de:wp?
Your arguments look to me like fully-general counterarguments against *any* on-Wikipedia poll whatsoever, no matter the structure or subject. What would you accept as a measure of the de:wp community that would actually be feasible to conduct?
233 is a large amount for a poll on de:wp. But it was no democratic poll, because the manner by which the poll was conducted was not democratic. A democratic and representative poll has to be equal, common and private. The poll was not common because not every user entitled to vote was noticed about the poll,
(example for a more democratic poll was the poll from the foundation in question bildfilter: it was on an anonymous server and I was notified by email that I was entitled to vote),
it was not private, because everybody can see who choose what. And finally it was not equal, because there was no means to exclude the possibility of sock puppet voting (Which is very common and very easy as far as I know - I know an unpunished such voting).
Every poll will be visible at the Autorenportal [1] under "Aktuelles" (current issues). So everyone can inform himself and decide if he wants to vote. We decided to have public polls since everyone should be able to discuss about the arguments and to leave comments. We have a policy for that. This is our model.
You must be an asshole to claim that we have many sock puppets inside this votes. It's an open attack against the community.
* User must be logged in * He must be active for at least two month (poll announcement and duration time is shorter) * He must have at least 200 edits inside the article namespace and more then 50 edits in the last 12 month.
On 22 September 2011 02:21, Tobias Oelgarte tobias.oelgarte@googlemail.com wrote:
You must be an asshole to claim that we have many sock puppets inside this votes. It's an open attack against the community.
Please let's try not to demonize and insult each other.
These are hot issues, and I know it's tempting to believe that people who disagree with us are stupid or evil, but they're not. We are all here for the same reason -- we want to make the world a better place by making information freely available for everyone. Everybody here is smart and thoughtful and committed to that mission, and the conversations work better when we engage each other with that in mind.
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!
Am 21.09.2011 22:20, schrieb Kanzlei:
Am 21.09.2011 um 20:10 schrieb Tobias Oelgartetobias.oelgarte@googlemail.com:
Am 21.09.2011 19:36, schrieb Kanzlei:
Am 21.09.2011 um 19:04 schrieb Tobias Oelgartetobias.oelgarte@googlemail.com:
Don't you think that we would have thousands of complaints a day if your words would be true at all? Just have a look at the article [[hentai]] and look at the illustration. How many complaints about this image do we get a day? None, because it is less then one complain in a month, while the article itself is viewed about 8.000 times a day.[1] That would make up one complainer in 240.000 (0,0004%). Now we could argue that only some of them would comment on the issue. Lets assume 1 of 100 or even 1 of 1000. Then it are still only 0,04% or 0,4%. That is the big mass of users we want to support get more contributers?
Your assumtion is wrong. The 8.000 daily are neither neutral nor representative for all users. Put the picture on the main page and You get representative results. We had that in Germany.
Yes we put the "vulva" on the main page and it got quite some attention. We wanted it this way to test out the reaction of the readers and to start a discussion about it. The result was as expected. Complains that it is offensive together with Praises to show what neutrality really is. After the discussion settled, we opened a Meinungsbild (Poll) to question if any article/image would be suitable for the main page (Actually it asked to not allow any topic). The result was very clear. 13 supported the approach to leave out some content from the main page. 233 (95%) were against the approach to hide some subjects from the main page.
This poll was not representative for wikipedia readers, but only for some German wikipedia editors. Scientifically research found that Germa editors are not representative for German speaking people but far more environmetal-liberal-leftists than avarage Germans. The poll was even not representative for German editors because only a few voted.
This needs a big *CITATION NEEDED*. We have the opposite examples like the article "Futanari", which i mentioned before.
You said that my assumption is wrong. We can repeat this for hundreds of articles and you would get the same result. Now proof that this assumption, which is sourced (just look at it) is wrong or say what is wrong with my assumption (in detail).
See above
foundation-l mailing list foundation-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
foundation-l mailing list foundation-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
On 21 September 2011 18:04, Tobias Oelgarte tobias.oelgarte@googlemail.com wrote:
One of the problems with the discussions about the image filter is that many of them argue - I paraphrase - that "Wikipedia must not be censored because it would stop being neutral". But is the existing "Wikipedian POV" *really* the same as "neutral", or are we letting our aspirations to inclusive global neutrality win out over the real state of affairs? It's the great big unexamined assumption in our discussions...
You describe us as geeks and that we can't write in a way that would please the readers. Since we are geeks, we are strongly biased and write down POV all day. If that is true, why is Wikipedia such a success? Why people read it? Do they like geeky stuff?
...no, that's really not what I said.
We've known for ten years that Wikipedia editors have systemic biases, and we've tried to avoid them by insisting on NPOV. This is one of the reasons we've been successful - it's not the only one, but it's helped.
But being neutral in text is simple. You give both sides of the argument, and you do it carefully, and that's it. The method of writing is the same whichever side you're on, and so most topics get a fair treatment regardless of our bias.
We can't do that for images. A potentially offensive image is either there, or it is not. We can't be neutral by half including it, or by including it as well as another image to balance it out - these don't make sense. So we go for reasonable, acceptable, appropriate, not shocking, etc. Our editors say "this is acceptable" or "this is not acceptable", and almost all the time that's based on *our personal opinions* of what is and isn't acceptable.
The end result is that our text is very neutral, but our images reflect the biases of our users - you and me. That doesn't seem to be a problem to *us*, because everything looks fine to us - the acceptable images are in articles, the unacceptable ones aren't.
People are saying we can't have the image filter because it would stop us being neutral. If we aren't neutral to begin with, this is a bad argument. It doesn't mean we *should* have the image filter, but it does mean we need to think some more about the reasons for or against it.
Am 22.09.2011 00:07, schrieb Andrew Gray:
On 21 September 2011 18:04, Tobias Oelgarte tobias.oelgarte@googlemail.com wrote:
One of the problems with the discussions about the image filter is that many of them argue - I paraphrase - that "Wikipedia must not be censored because it would stop being neutral". But is the existing "Wikipedian POV" *really* the same as "neutral", or are we letting our aspirations to inclusive global neutrality win out over the real state of affairs? It's the great big unexamined assumption in our discussions...
You describe us as geeks and that we can't write in a way that would please the readers. Since we are geeks, we are strongly biased and write down POV all day. If that is true, why is Wikipedia such a success? Why people read it? Do they like geeky stuff?
...no, that's really not what I said.
We've known for ten years that Wikipedia editors have systemic biases, and we've tried to avoid them by insisting on NPOV. This is one of the reasons we've been successful - it's not the only one, but it's helped.
But being neutral in text is simple. You give both sides of the argument, and you do it carefully, and that's it. The method of writing is the same whichever side you're on, and so most topics get a fair treatment regardless of our bias.
We can't do that for images. A potentially offensive image is either there, or it is not. We can't be neutral by half including it, or by including it as well as another image to balance it out - these don't make sense. So we go for reasonable, acceptable, appropriate, not shocking, etc. Our editors say "this is acceptable" or "this is not acceptable", and almost all the time that's based on *our personal opinions* of what is and isn't acceptable.
Given that this would be true. Do you expect us to categorize images for the filter in a right way, so that we are able to define what is offensive or not? Do we have now the option to hide an image or not, while being able to be neutral in judgment? Isn't it just the same? Did anything change, despite the fact that we are now making global, image based (not article based) decisions to show or hide an image?
The end result is that our text is very neutral, but our images reflect the biases of our users - you and me. That doesn't seem to be a problem to *us*, because everything looks fine to us - the acceptable images are in articles, the unacceptable ones aren't.
If a statement is included in the article is based upon the decision of the authors. If some authors disagree they will have to discuss. If one author inserts an image in the article that he does find usable and another disagrees, don't we also discuss about it? What is the difference between the decision to include a fact or an image inside an article?
People are saying we can't have the image filter because it would stop us being neutral. If we aren't neutral to begin with, this is a bad argument. It doesn't mean we *should* have the image filter, but it does mean we need to think some more about the reasons for or against it.
I personally choose images only based on the fact if they illustrate the topic. That means that an offensive image will without doubt get precedence over an not offensive alternative image if it depicts the subject better. Thats a very simple way. Just leave out moral aspects and use the images to describe the topic. If two images have the same educational value then we could start to discuss if other aspects (quality, moral, etc.) might apply. But I'm not willed to exchange a correct depiction of a subject against and imperfect depiction on moral grounds. That means to represent the truth, pleasing or not, and not to represent pink easter bunnies on soft green with a charming sunset in the background.
Andrew Gray wrote:
People are saying we can't have the image filter because it would stop us being neutral. If we aren't neutral to begin with, this is a bad argument.
An image filter feature isn't inherently non-neutral, but one reliant upon special categories (as described in the WMF outline) would be.
Unlike our normal categories (which objectively and non-exclusively describe images' subjects in great detail), a manageable setup would require broad, objection-based (i.e. extremely subjective) classifications.
Additional non-neutrality enters the equation when we provide categories for some "potentially objectionable" images and not others (implicitly validating belief x and discriminating against holders of belief y). As previously discussed, this isn't realistically avoidable.
Even if it were feasible to include an "unveiled women" category, who would analyze the millions of images (with thousands more uploaded every day) to tag them accordingly? That's merely a single example.
David Levy
Am 21.09.2011 17:37, schrieb WereSpielChequers:
I get the idea that there are theoretical reasons why image filters can't work, and I share the view that the proposal which was consulted on needs some improvement. An individual choice made at the IP level was a circle that looked awfully difficult to square.
But since Flickr has already proven that something like this can work in practice, can we agree to classify Image filters as one of those things that work in practice but not in theory? Then we can concentrate on the practical issue of if we decide to implement this, how do we do it better than Flickr has?
NB I would not want us to implement this the way Flickr has http://www.flickr.com/help/filters/#258 And not only because I'm not totally convinced that our community would share their view that Germany is the country that needs the tightest restrictions.
Hugs
WereSpielChequers
PS My niece absolutely wants that magical flying unicorn pony for the winter solstice, especially if it s***s rainbows. Would you mind telling me where I can order one
Using flickr as an example is an bad example. At first there thousands if not millions of images with false categorization, meaning that the filter is ineffective. Just do a quick search on your own and you will find the examples. Secondly flickr does not advocate knowledge. It has a completely different mission.
PS: Just implement the filter and you will see that unicorn-rainbow-brick-argumentation falling from the sky, where you pushed it.
On Wed, Sep 21, 2011 at 04:37:31PM +0100, WereSpielChequers wrote:
But since Flickr has already proven that something like this can work in practice, can we agree to classify Image filters as one of those things that work in practice but not in theory? Then we can concentrate on the practical issue of if we decide to implement this, how do we do it better than Flickr has?
The secret ingredient is the frame in which you work.
David's Magical Flying unicorn ponies work very well thank you in the frame of fairytale fiction.
Flickr's system works -to some degree- within intersection of the frames of the flickr community, commercial objectives, and the laws of physics.
Having a filter for wikimedia projects might be very hard, without violating some of our basic beliefs along the way.
How many folks here have actually tried to figure out a solution to this puzzle given the frame/constraints?
It's really tricky! I keep getting stuck.
sincerely, Kim Bruning
On 21 September 2011 18:41, Kim Bruning kim@bruning.xs4all.nl wrote:
David's Magical Flying unicorn ponies work very well thank you in the frame of fairytale fiction.
Originally from discussion of an actually impossible project at work. "Look, we've specified a magical flying unicorn pony because we think they'll be popular and sell well. Stop calling it 'impossible' and just do your job." I'm sure you can picture precisely how spectacularly well the projects specified in this manner turn out.
- d.
On Wed, Sep 21, 2011 at 07:42:18PM +0100, David Gerard wrote:
On 21 September 2011 18:41, Kim Bruning kim@bruning.xs4all.nl wrote:
David's Magical Flying unicorn ponies work very well thank you in the frame of fairytale fiction.
Originally from discussion of an actually impossible project at work. I'm sure you can picture precisely how spectacularly well the projects specified in this manner turn out.
Presumably your occupation has nothing to do with fairytale fiction? ;-)
sincerely, Kim Bruning
On 21 September 2011 18:58, Kim Bruning kim@bruning.xs4all.nl wrote:
On Wed, Sep 21, 2011 at 07:42:18PM +0100, David Gerard wrote:
On 21 September 2011 18:41, Kim Bruning kim@bruning.xs4all.nl wrote:
David's Magical Flying unicorn ponies work very well thank you in the frame of fairytale fiction.
Originally from discussion of an actually impossible project at work. I'm sure you can picture precisely how spectacularly well the projects specified in this manner turn out.
Presumably your occupation has nothing to do with fairytale fiction? ;-)
Unfortunately not - it deals with things that might actually affect people's well-being.
This makes projects based on ill-conceived premises that fundamentally require a magical category (or two, or three) to work more than a little ... problematic.
It turns out that thinking you can get a magical result by demanding it and passing resolutions doesn't work, no matter how politically expedient it may be to demand it.
The magical demand presently under discussion isn't on the same level, but is equally magical and is just as unlikely to actually work, no matter how politically expedient it was to demand it.
Dear WMF Board: you have passed a resolution demanding magic. If you don't realise this, there's no hope for any of you.
- d.
wikimedia-l@lists.wikimedia.org