[Foundation-l] Letter to the community on Controversial Content

David Levy lifeisunfair at gmail.com
Tue Oct 11 23:58:52 UTC 2011

Andreas Kolbe wrote:

> I would use indicators like the number and intensity of complaints received.

For profit-making organizations seeking to maximize revenues by
catering to majorities, this is a sensible approach.  For most WMF
projects, conversely, neutrality is a fundamental, non-negotiable

> Generally, what we display in Wikipedia should match what reputable
> educational sources in the field display. Just like Wikipedia text reflects
> the text in reliable sources.

This is a tangential matter, but you're comparing apples to oranges.

We look to reliable sources to determine factual information and the
extent of coverage thereof.  We do *not* emulate their value

A reputable publication might include textual documentation of a
subject, omitting useful illustrations to avoid upsetting its readers.
 That's non-neutral.

> That does not mean that we should not listen to users who tell us that they
> don't want to see certain media because they find them upsetting, or
> unappealing.

Agreed.  That's why I support the introduction of a system enabling
users (including those belonging to "insignificant" groups) to filter
images to which they object.

> I would deem them insignificant for the purposes of the image filter. They
> are faced with images of women everywhere in modern life, and we cannot cater
> for every fringe group.

The setup that I support would accommodate all groups, despite being
*far* simpler and easier to implement/maintain than one based on
tagging would be.

> At some point, there are diminishing returns, especially when it amounts to
> filtering images of more than half the human race.

That such an endeavor is infeasible is my point.

> We need to look at mainstream issues (including Muhammad images).

We needn't focus on *any* "objectionable" content in particular.

> That would involve a user switching all images off, and then whitelisting
> those they wish to see; is that correct? Or blacklisting individual
> categories?

Those would be two options.  The inverse options (blacklisting images
and whitelisting entire categories) also should be included.

And it should be possible to black/whitelist every image appearing in
a particular page revision (either permanently or on a one-off basis).

> This would be better from the point of view of project neutrality, but would
> seem to involve a *lot* more work for the individual user.

Please keep in mind that I don't regard a category-based approach as
feasible, let alone neutral.  The amount of work for editors (and
related conflicts among them) would be downright nightmarish.

> It would also be equally likely to aid censorship, as the software would have
> to recognise the user's blacklists, and a country or ISP could then equally
> generate its own blacklists and apply them across the board to all users.

They'd have to identify specific images/categories to block, which
they can do *now* (and simply intercept and suppress the data

David Levy

More information about the foundation-l mailing list