On 3/27/06, Neil Harris <neil(a)tonal.clara.co.uk> wrote:
Unfortunately, that's easier said than done --
"child-safe" and
"worksafe" are concepts that are impossible to define in a way that
everyone can agree on. What one parent or community regards as
acceptable may be unacceptable in another; what a parent wants their
ten-year-old child to be able to see will probably differ from what they
want their sixteen-year-old child to be able to see, and so on.
I agree completely that defining such concepts in a global way is not
possible. However, concepts such as "covered nipples", "men in
shorts"
or "photographs of genitalia" are objective. Individuals with
appropriate software can then filter as they see fit.
For some examples of edge cases: consider pictures of
men wearing
shorts, which are regularly banned by the censors in some of the more
conservative Middle-Eastern states: do we mark all articles showing
images of uncovered arms or legs as "unsafe"? How about pictures of
No, we mark them "men with uncovered arms or legs".
women with uncovered hair? Do we mark the
[[Holocaust]] article, which
is extremely upsetting, as "unsafe" for children to read? How about
Ditto.
[[death]], which is upsetting for very small children?
What about
pictures of [[Bahá'u'lláh]], which observant Bahá'ís prefer not to see
in public, or even in their own homes?
Currently, we have no "offensiveness" markup whatsoever. If and when
we implement such a system, someone can create a tag called "pictures
of Bahaullah", which can be used and filtered against as appropriate.
I recommend reading RFC 3675 for a full and detailed
discussion of all
the issues involved: its authors conclude that broad-brush attempts at
content filtering as "ill considered [...] from the legal,
philosophical, and particularly, the technical points of view."
I agree. Which is why I'm not proposing broad-brush content filtering,
but instead fine-brush content *tagging*.
Rather than attempting to define "safe" and "unsafe" categories, we
should instead concentrate on assigning all Wikipedia articles to
meaningful fine-grained descriptive categories, without any implied
judgment that a category is "safe" or "unsafe" for any given viewer.
Downstream users who want to filter Wikipedia's content can then use
this information to make their own choices.
Ok, I now see that you were actually replying to my original message
where I glibly used the terms "kidsafe" and "worksafe". My bad.
Steve