[WikiEN-l] Shock site bot

[Poor, Edmund W]

> I recently recieved this message from a user:
> "I'm a regular wikipedia user although i don't have an 
> account here. I think this site is great and it really helps 
> me with my college work. But I recently heard of these people 
> that were talking about wikipedia that they were all 
> programming a hack for it. So after a little while I found it 
> was a spider to hunt down all the pages links and change them 
> to shocks site links or something along those lines. I didn't 
> know who to tell so I just thought I'd tell an administrator 
> as they might know who to tell or what to do. Just giving an 
> advanced warning so you might be able to do something to 
> protect this wonderful resourse. Apparently they permenantly 
> change their ip address using some thing (a bit beyond me). 
> Something like that. I just didn't know what to do. I hope I 
> didn't embaress myself here. Thanks for your time."
> 
> Brett

I'd like to suggest a standard where admins can block any account on
"suspicion of being a bot". This would be an assumption of guilt pending
proof of innocence, so let's talk about it.

Ordinarily, we assume good faith. But if, AFTER making that initial
assumption (and leaving the door open for anyone - human or bot) to come
in and edit - if we become suspicious, we ought to be able to "stop them
for questioning".

Here's how it might work.
1. Admin gets suspicious of a pattern of edits.
2. (optional step, try to engage user in any number of ways, e.g.,
article talk pages, edit summaries in reverts, etc.)
3. Block account
4. At this point, user / bot can only post to their / its user talk
page.
5. Post a message on the user talk page which (in your opinion) only a
human being could respond to.
6. If they refuse / fail to answer, the presumption stands. 

Drawbacks:
A. User might be shy about interacting. For example, just wants to
correct spelling - but not talk to other Wikipedians. Maybe they're
embarassed that their English isn't perfect, or they have Asperger's or
are autistic.
B. Evil admin might abuse this policy (gaming the system) to challenge
someone that they KNOW is a legitimate user.

Advantages:
C. Shoot first, ask questions later - saves time in an emergency
D. No real harm done to legitimate user: they say the bold "You have
messages" notice and can click on the link; and simply say, "Jeez,
relax, I'm not a bot."

Uncle Ed