I can think of other concerns.
The main one is that of our competence to form judgements. On some things we can: though nudity would seem something obvious, deciding on the various degrees of it is not: I do not think we are likely to agree on whether any particular nude image is primarily sexual,m or primarily non-sexual. If we tried to be precise, we would degenerate towards a situation like some legal codes which state exactly what portions of a female breast may be displayed in a particular context, or in just what way something must be covered to make it non-nude.
Further, though I consider it essential that Commons should include appropriately educational sexual and even pornographic content, I do not think it should concentrate on that; important though education about human sexuality is, there are other things to educate about also, many just as much dependent upon images. A project to minutely categorize pages on sexuality would concentrate much of the volunteer effort on this portion of the contents. We need some editors who want to work on this field primarily, but we do not need everybody.
David Goodman, Ph.D, M.L.S. http://en.wikipedia.org/wiki/User_talk:DGG
On Sun, May 9, 2010 at 6:26 PM, Excirial wp.excirial@gmail.com wrote:
*That is why it was addressed to FOSI and cc'ed to some parties that might have clue about such systems. The copy to foundation-l was a courtesy message. You are welcome to discuss censorship and your opinion about it, but I would appreciate it even more if people actually talked about rating systems.*
Very well, lets see if i can write up some something more on the point then:
*Definition and purpose*: The purpose of such a system would be to allow certain content to be filtered from public view. The scope of such a project is discussable, and dependent upon the goals we wish to reach.
*Rating System*: In order to decide a contents category and offensiveness there has to be a method to sort the images. Multiple options are available:
*Categorization:* Categories could be added to images to establish the subject of an image. For example, one image might be categorized nudity, the other might be categorized as sexual intercourse and so on. The categorization could be similar to the way we categorize our stub templates
- we could create a top-level filter for "Nudity" and create more specific
categories under that. That way it is possible to fine-tune the content one might not wish to see. *Rating:* Another method is rating each image. Instead of using a category tree we might use a system that allows users to set a level of explicitness or severity for each image. An image which shows non sexual nudity would be rated lower then an image which shows a high level of nudity. Note that such a system would require a clear set of rules as a rating might be subject to ones personal idea's and feelings towards a certain subject.
*Control mechanism*:There are various levels at which we can filter content:
*Organization wide:* An organization wide filter would allow an organization to block content based upon site-wide settings. Techically this would likely prove to be the more difficult option to implement as it would require both local and external changes. There are multiple methods to execute this though. For example a server may rely a certain value to Wikipedia at the start of each session detailing the content that should not be forwarded over this connection. Based on such a value the server could be programmed in such a way that images of a certain category won't be forwarded, or would be replaced by placeholders. The advantage of this method is that it allows organizations such as schools to control which content should be shown, therefor possibly negating complete blocks of Wikipedia. The negative is that it takes away control from the user. *Par-user:* A second method is allowing par-user settings. Such a system would be easier to build and integrate as it only requires changes on wikipedia's side. A seperate section could be made under "My preferences" which would include a set of check boxes where a user could select which content he or she prefers not to see. Images falling under a certain category could be replaced with the images alt text or with an image stating something akin to "Par your preferences, this image was removed". *Hybrid*: A hybrid system could integrate both systems. A user might override or increase organization level settings if he or she has personal preferences.
*Possible concerns* *Responsibility and vandalism: *One risk with rating systems is that they might be abused for personal goals, akin to article vandalism. Therefor there should be some limit on who can rate an image - anonymous rating could change images in such a way that they may be visible or invisible to people who might or might not want this. *Volunteer interest:* Implementing such a system would likely require a lot of volunteer activity. Not only has every image to be checked and rated, we would also have a backlog of over 6 million images to rate. Therefor we should have sufficient volunteers who are interested in such a system. *Public interest*: Plain and simple: Will people actually use this system? Will people be content with their ability to filter, or will they still try to remove images they deem offensive? Also: How many editors would use this system? *Implementation area*: Commons only? Local and commons?
That is all i can think of for now. I hope it is somewhat more constructive towards the point you were initially trying to relay :) ~Excirial
On Sun, May 9, 2010 at 11:26 PM, Derk-Jan Hartman d.j.hartman@gmail.comwrote:
This message was an attempt to gain information and spur discussion about the system in general, it's limits and effectiveness, not wether or not we should actually do it. I was trying to gather more information so that we can have an informed debate if it ever got to discussing about the possibility of using ratings.
That is why it was addressed to FOSI and cc'ed to some parties that might have clue about such systems. The copy to foundation-l was a courtesy message. You are welcome to discuss censorship and your opinion about it, but I would appreciate it even more if people actually talked about rating systems.
DJ
On 9 mei 2010, at 15:24, Derk-Jan Hartman wrote:
This message is CC'ed to other people who might wish to comment on this
potential approach
Dear reader at FOSI,
As a member of the Wikipedia community and the community that develops
the software on which Wikipedia runs, I come to you with a few questions.
Over the past years Wikipedia has become more and more popular and
omnipresent. This has led to enormous problems, because for the first time, a largely uncensored system has to work in the boundaries of a world that is largely censored. For libraries and schools this means that they want to provide Wikipedia and its related projects to their readers, but are presented with the problem of what some people might consider, information that is not "child-safe". They have several options in that case, either blocking completely or using context aware filtering software that may make mistakes, that can cost some of these institutions their funding.
Similar problems are starting to present themselves in countries around
the world, differing views about sexuality between northern and southern europe for instance. Add to that the censoring of images of Muhammad, Tiananman square, the Nazi Swastika, and a host of other problems. Recently there has been concern that all this all-out-censoring of content by parties around the world is damaging the education mission of the Wikipedia related projects because so many people are not able to access large portions of our content due to a small (think 0.01% ) part of our other content.
This has led some people to infer that perhaps it is time to rate the
content of Wikipedia ourselves, in order to facilitate external censoring of material, hopefully making the rest of our content more accessible. According to statements around the web ICRA ratings are probably the most widely supported rating by filtering systems. Thus we were thinking of adding autogenerated ICRA RDF tags to each individual page describing the rating of the page and the images contained within them. I have a few questions however, both general and technical.
1: If I am correctly informed, Wikipedia would be the first website of
this size to label their content with ratings, is this correct?
2: How many content filters understand the RDF tags 3: How many of those understand multiple labels and path specific
labeling. This means: if we rate the path of images included on the page different from the page itself, do filters block the entire content, or just the images ? (Consider the Virgin Killer album cover on the Virgin Killer article, if you are aware of that controversial image http://en.wikipedia.org/wiki/Virgin_Killer)
4: Do filters understand per page labeling ? Or do they cache the first
RDF file they encounter on a website and use that for all other pages of the website ?
5: Is there any chance the vocabulary of ICRA can be expanded with new
ratings for non-Western world sensitive issues ?
6: Is there a possibility of creating a separate "namespace" that we
could potentially use for our own labels ?
I hope that you can help me answer these questions, so that we may
continue our community debate with more informed viewpoints about the possibilities of content rating. If you have additional suggestions for systems or problems that this web-property should account for, I would more than welcome those suggestions as well.
Derk-Jan Hartman
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