[Foundation-l] Image filter brainstorming: Personal filter lists

Erik Moeller erik at wikimedia.org
Mon Nov 28 10:46:49 UTC 2011

On Mon, Nov 28, 2011 at 10:21 AM, David Gerard <dgerard at gmail.com> wrote:
> Unfortunately, the issue is not dead.

That's correct; nobody from WMF has said otherwise. What's dead is the
idea of a category-based image filter, not the idea of giving
additional options to readers to reversibly collapse images they may
find offensive, shocking, or inappropriate in the context in which
they're viewing them (e.g. at work). However, Sue has made it clear
that she wants the WMF staff to work with the community to find a
solution that doesn't mean strong opposition. Her presentation on the
issue in Hannover begins with this slide:


My personal view is that such a solution will need to take into
account that actual current editorial practices and perceptions in our
projects vary a great deal, as did the image filter poll results by
language. As I pointed out before, projects like Arabic and Hebrew
Wikipedia are currently collapsing content that's not even on the
radar in most of these discussions (e.g. the 1866 painting L'Origine
du monde in Hebrew Wikipedia), while German Wikipedia put the vulva
photograph on its main page. A solution that pretends that this
continuum of practice can be covered with a single approach, one which
doesn't give a lot of flexibility to readers and editors, is IMO not a
solution at all.

I'm not convinced that the "collapse image one-by-one" approach to
develop a filter list is very valuable in and of itself due to lack of
immediate practical impact and likely limited usability. The idea of
making it easy to build, import and share such lists of images or
image-categories would move the process of categorization into a
market economy of sorts where individual or organizational demand
regulates supply of available filters. This could lead to all kinds of
groups advertising their own filter-lists, e.g. Scientology, Focus on
the Family, etc. From there, it would be relatively small step for
such a group to take its filter list and coerce users to only access
Wikipedia with the filter irreversibly in place.

While third parties are already able to coerce their users to not see
certain content, creating an official framework for doing so IMO puts
us dangerously close to censors: it may lead to creation of regimes of
censorship that did not previously exist, and may be used to exercise
pressure on WMF to change its default view settings in certain
geographies since all the required functionality would already be
readily available.

My personal view on this issue has always been that one of the most
useful things we could do for readers is to make NPOV, well-vetted and
thorough advice too users on how to manage and personalize their net
access available to them. Wikipedia is only one site on the web, and
whatever we do is not going to extend to the rest of the user's
experience anyway. There are companies that specialize in filtering
the Net; we could point people to those providers and give advice on
how to install specific applications, summarizing criticism and praise
they have received.

On the other hand, such advice would be pretty removed from the
experience of the reader, and l do think there are additional
reasonable things we could do. So I'm supportive of approaches which
give an editing community additional flexibility in warning their
readers of content they may find objectionable, and give readers the
ability to hide (in the general or specific case) such content. As I
said previously, this wouldn't create a new regime of filter lists or
categories, merely a broad community-defined standard by which
exclusion of some content may be desirable, which could vary by
language as it does today.

Kim, I just read the conversation on your talk page. In general, I
agree that more research into both the current practices of our
editing communities as well as reader expectations and needs would be
valuable. Right now we have some anecdotal data points from the
projects, Robert's original research which mostly focuses on
establishing definitions and principles, and the image filter poll
results. I think the latter are useful data if carefully analyzed, but
they do mingle low-activity users who are chiefly readers with the
core editing community in ways that don't give us tremendously clear
information by group. The poll also referred to a filtering concept
that's now been rejected.

At the same time, I do think that we shouldn't hesitate to build some
cheap prototypes to make abstract ideas more understandable. I think
to advance our understanding, as well as the state of the
conversation, through both additional pointed research, as well as
discussion of some interactive prototypes, without spending tremendous
amounts of time and money on either, feels like a response that's
commensurate to the scale and importance of the issue.

Erik Möller
VP of Engineering and Product Development, Wikimedia Foundation

Support Free Knowledge: http://wikimediafoundation.org/wiki/Donate

More information about the foundation-l mailing list