[Foundation-l] Porchesia atonement

George Herbert george.herbert at gmail.com
Tue Oct 3 20:58:23 UTC 2006

On 10/3/06, David Gerard <dgerard at gmail.com> wrote:
> On 03/10/06, Birgitte SB <birgitte_sb at yahoo.com> wrote:
> > --- David Gerard <dgerard at gmail.com> wrote:
> > > The problem is how to come up with a sources
> > > criterion that can't be hoaxed.
> > > There's enough trouble with stupid AFD nominations
> > > on en:wp by people
> > > who couldn't find the subject on Google and presumed
> > > it therefore
> > > didn't exist.
> > > Will we forbid print sources unless a scan is lodged
> > > with the WMF?
> > Obviously not (Copyright?).  I think you are beind
> > sarcastic, but it is hard to be ceratain.
> It is somewhat reductio ad absurdum. But there are those on en: who
> seriously advocate that a reference can only be good if it's easy for
> a normal person (presumably in the US) to find.
> >  "How to
> > come up with a sources criterion that can't be hoaxed"
> > is only a the problem with the proposed solution of
> > requiring sources.  I think this is large enough
> > problem that requiring sources should be thrown out of
> > consideration for this particular problem.
> Coming up with a rigid rule that would catch this hoax without causing
> ridiculous quantities of collateral damage will not be easy.
> [[:en:Wikipedia:Reliable sources]] is marked "guideline" but phrased
> didactically, so when applied robotically - and people do apply it
> robotically - is disastrous in practice, gutting articles and causing
> the sort of PR disasters over living bios it was written didactically
> so as to avert.

Requiring sources is only partly helpful, as well.

It takes near-zero effort to set up multiple apparently widely separated
websites with coverage of a subject.

The amount of effort required to write up a faked book and self-publish it
out via one of the print-on-demand publishers, and then get it into Amazon,
isn't really all that much either.

Afraid that people will call you on that?  How much effort does it take to
make a fake page scan of a book claiming to have been printed prior to the
advent of ISBN numbers?  Oh, your library doesn't have a copy?  Not
suprising, only 5,000 copies were printed, ...

Intend to do it as an April 1 related event?  You can probably get reliable
news sources and personalities who are verifyably known to the world to
assist, if you find the ones with senses of humor.

If we have, or ever have in the future, people who are seriously interested
in hoaxing Wikipedia there is practically very little we can possibly do to
prevent it.  We can set a barrier for entry which is high enough to keep the
amateurs and vandals deterred... and I like the ideas people are posting for
additional review projects and hasn't-been-edited-for-a-month lists and
such.  I was thinking about this a couple of days ago; for example, on
en.wikipedia we have around a thousand admins and active trustable editors.
With a million articles, we could scan everything in a year, at a rate of 3
articles a person a day.  A few minutes of fact checking (and, while you're
there, putting in some references if it doesn't already have them, etc) each
day would go a long ways towards helping catch low-effort hoaxes.

I had been thinking that we could create an automatic list, which people
could then be randomly assigned articles out of to go check, and then
confirm back that they'd checked it somehow (put a tag on the talk page?).
Or break it down by topic, so people don't get article subjects they can't
reasonably fact check other than via google.

-george william herbert
george.herbert at gmail.com

More information about the foundation-l mailing list