On Thu, Dec 1, 2011 at 8:11 PM, Jussi-Ville Heiskanen cimonavaro@gmail.com wrote:
... The "downstream use" objection was *never* about downstream use of _content_ but downstream use of _labels_ and the structuring of the semantic data. That is a real horse of a different colour, and not of straw.
Tom thinks that this horse is real, but it has bolted. I agree with Tom that it is very simple for a commercial filter provider, or anyone else who is sufficiently motivated, to find most naughty content on WP and filter it. Risker said she had experienced something like this. Universities and schools have this too.
I would prefer that we do build good metadata/labels, but that we (wikimedia) do not incorporate any general purpose use of them for filtering from readers. Hiding content is the easy way out. The inappropriate content on our projects is of one of two types:
1. inappropriate content that is quickly addressed, but it is seen by some people as it works its way through our processes. Sometimes it is the public that sees the content; sometimes it is only the community members who *choose* to patrol new pages/files while on the train.
2. content which is appropriate for certain contexts, is known to be problematic but concensus is that the content stays, however readers stumble on it unawares.
The former cant be solved.
The latter can be solved by labelling but not filtering. If you are on the train and a link is annotated with a tag "nsfw", you can not click it, or be wary about the destination page.
-- John Vandenberg