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

Sat Nov 26 14:41:51 UTC 2011

On Thu, Nov 24, 2011 at 14:59, Tobias Oelgarte
<tobias.oelgarte at googlemail.com> wrote:
> I'm a little bit confused by this approach. On the one side it is good
> to have this information stored privately and personal, on the other
> side we encouraging the development of filter lists and the tagging of
> possibly objectionable articles. The later wouldn't be private at all
> and even worse then tagging single images. In fact it would be some kind
> of additional force to ban images from articles just to keep them in the
> "clean" section.
>
> Overall i see little to now advantage over the previously supposed
> solutions. It is much more complicated, harder to implement, more
> resource intensive and not a very friendly interface for readers.
>

Err, think of it with an analogy to AdBlock. You can have lists stored
privately (in Adblock: in your browser settings files, in an image
filter: on the WMF servers but in a secret file that they'll never
ever ever ever release promise hand-on-heart*) and you can have lists
stored publicly (in Adblock: the various public block lists that are
community-maintained so that you don't actually see any ads, in an
image filter: on the web somewhere). And you can put an instruction in
the former list to transclude everything on a public list and keep it
up-to-date.

Given it works pretty well in Adblock, I don't quite see how that's a
big deal for Wikimedia either. Performance wise, you just have it so
the logged in user has a list of images they don't want to see, and
you have a script that every hour or so downloads and caches the
public list, then when they call to retrieve the list for the purposes
of seeing what's on it, it simply concatenates the two. This seems
pretty straightforward.

And if the WMF doesn't do it - perhaps because people are whinging
that me being given the option to opt-in and *not* see "My
micropenis.jpg" is somehow evil and tyrannical and contrary to
NOTCENSORED - it could possibly be done as a service by an outside
group and then implemented on Wikipedia using userscripts. The
difference is that the WMF may do it in a slightly more user-friendly
way given that they have access to the servers.

* That's less sarcastic than it sounds.

-- 
Tom Morris
<http://tommorris.org/>