[Foundation-l] Personal Image Filter results announced

Kim Bruning kim at bruning.xs4all.nl
Wed Sep 7 23:45:58 UTC 2011


On Mon, Sep 05, 2011 at 06:25:05AM -0700, phoebe ayers wrote:
> The difference lies in our role as active editors (vs the librarian
> role as curators), making active choices; a reference work is a
> different kind of project from a library. It also lies in a difference
> in intent -- what the ALA speaks out about is labeling that is
> intended to restrict access. None of our labeling intends to restrict
> access to anything for anyone.


I guess this is where we get to the point where I disagree with
you Phoebe :)

We both agree that restricting access is evil. 

I think you believe there is a way in which we can make a labelling
scheme for filtering that is not intended to restrict access.

I believe that filtering is -per definition- a form of restricting
access. The proposed filter itself is fairly benign.However, the same
labels that are used on wikipedia to help good people to restrict
themselves being exposed to bad pictures, can equally be used by bad
people to restrict access to good pictures.

I have the impression you believe in the good in people. :)
I do too. Rotten apples are very rare!

In this case though, I think it only takes just one rotten apple to ruin
everyone's day. So we need to plan to ensure that there is
no way the rare rotten apple can subvert our work.

I know you believe that this is possible. We have a smart
community, surely someone can come up with a working solution.

I'm not so sure. My experience is that filters and their databases
tend to have all kinds of unintended side effects and collateral
damage. I've never seen it go right. Wikipedia would be the first
time that it ever did. I'm not saying it's entirely impossible.
Just that apparently it is very hard. And if we accidentally miss
something, it's going to ruin our day, our month or even our
year.

If we succeed, we anger our friends, and our enemies will
only clamor slightly less loudly. I'm not sure we will reach many
new people. I have seen some reports, but none answered that particular
question afaik. (Have I missed anything?)

If we happen to fail in the wrong way, one worst case scenario is
that our mission becomes doomed.  (If I were evil, I'd know exactly
how to make that happen)

So it's a high risk, low reward kind of play, in my personal
assesment. The board has said that they want this. I think they
surely must have a different risk assesment. :-)

So that explains some of my practical reasons for being somewhat
skeptical -not of the filter- but of the category system behind it.

sincerely,
	Kim Bruning




More information about the wikimedia-l mailing list