[Wikipedia-l] A related ban issue

Sun Sep 8 02:04:07 UTC 2002

Rosa Williams wrote:
> 
> What with the recent discussion of banning "problem" users, I thought I'd
> bring this up for discussion/re-discussion.
> 
> Our policy on banning people for vandalism is (as I interpret what I've
> read) that we restrict it to "repeated and sustained" non-useful alterations
> of articles.
> 
> However, it's September, the high school and college students are back with
> their free school accounts, and inevitably the amount of drive-by vandalism
> seems to be on the increase. Several of us constantly check new edits by
> unknown contributors, and even then, we're missing vandalism that only turns
> up later when paging through via "Random Page" or otherwise coming across an
> article. As the number of articles goes up, the chance of locating such
> vandalism goes down.
> 
> I've tried a few approaches to ameliorating this.   I regularly check "this
> user's contributions" for vandals, and even sometimes for unfamiliar IP's
> (*thank* you folks for adding that code feature!) I do keyword searches for
> common obscenities, et cetera.  (No, Cunctator, I don't remove them if
> they're obviously part of the article.)  And, of course, I haunt the "Recent
> Changes" page. But I think it's getting harder to keep up.
> 
> I would like to suggest we add "obviously malicious vandalism" to reasons
> for an immediate (if temporary) IP ban: a single "Ths page is stupid"
> should be, in my opinion, enough to ban the address. This saves us from
> having to spend time on the next five instances of vandalism from that
> contributor, which could be better spent searching for other graffiti or
> *gasp* actually adding content.

My understanding of how ip banning and common use of reassignment
of IP numbers leads me to this concern:

If too many casual or hit and run type vandals are banned that we 
are likely banning the next users, not the vandal.  This could be
counterproductive if it occurs in conjunction with recruiting efforts
or methods under discussion in other threads.

For high school or college users to begin relying on the Wikipedia
as a resource timely access is required due to homework deadlines,
typically on the order of days or hours, not weeks.   Encountering
frequent blocks due to local vandals on the same pool of IP addresses
is likely to encourage the view that Wikipedia is unreliable, not
that inappropriate local use is causing the problem.  If the user
becomes aware that he/she is being punished for another's misdeeds
this could form an even worse impression.

> 
> Sure, one person's vandalism is another person's newbie goof.  I would agree
> that if there's any reasonable possibility that a change was just a newbie
> goof or something similar, we should err on the side of caution and not ban.
> But in the really obvious cases - "PHREAK WUZ HERE!!"  "Louis IV  was a
> dirty frog"  "f*ck you all", and similar - I honestly think we should go
> ahead and administer a slapdown in the form of a temporary IP ban.  If they'
> re just drive-by vandals, they'll lose interest that much faster; if it is a
> serious vandal, they'll at least have to go to the trouble of getting a new
> IP# for each new instance of vandalism.

What period of time for routine banning would you (anyone) suggest 
as an estimate of the initial proper tradeoff between potential 
denial of service to legitimate users and the attention span of 
casual or hit and run type vandalism?

Are you aware that denial of service is often the goal of
low level crackers or "script kiddies"?

This type of banning could actually become an incentive 
or invitation to certain types of vandal mentalities if
structured and managed carefully.

> 
> Yes, there's the possibility that someone may be too quick on the gun and
> ban someone who might, in the fullness of time, have become a useful
> contributor. But me, I think... do we really *want* a contributor who is
> starting off on the level of adding "This is so gay" to a page?  

Yes.  If they (some threshold percentage to be determined later
with empirical data) become productive contributers we should 
eventually show a profit.  

I agree it is a long term investment.  I agree it is possible to
be losing more initially than we are gaining until more effective 
methods are found.  

If the decision (for now) is to stick with quick, known, short
term returns then I propose we plan to periodically reassess the 
policy.   We should document the periodic review process so that 
interested members of the community know when, where, how to 
appropriately express their current views.

>The time
> just bringing them up to speed hardly seems worth it. 

It seems worth it to me.  If there is sufficient interest in
mutual self education then this can be left to those who choose
to volunteer for it.   If we decide that we have no time for
this in the stacks then perhaps an alternate medium for
"remedial" students and volunteer contributors can be established.

Directing newcomer's to an NPOV editing tutorial for practice
if interested may be more effective that directing them to the
current draft of the policy which is only lightly endorsed by
members of the community and is at least subtle for many people
if not actually "obviously" confusing.

Elsewhere I have proposed an ECP (engineering change proposal)
or code walkthrough type of approach that would require three
"vandals" to agree with each other that the "vandalism" is
an appropriate change.   This should slow down and break up
vandalism, without potential denial of service, in several ways.

1.  More effort is required to create the fake accounts and
engage in the vandalism.  It may be easier to delete 3 accounts
and revert the damage than it is for the creator of the vandalism.
Perhaps a delay on account creation with an appropriate explanation
would slow down vandals while not discouraging new dropings unduly.

2.  Any "vandals" editing in good faith are likley to encounter
resistance from cohorts.  They can spend some time bickering
among themselves on the talk page or elsewhere determining 
what is or is not appropriate.   Any defectors to our published
guidelines should be welcomed.

Perhaps an appropriate modification of my proposed approached
could be combined with the page freezing proposed elsewhere
for occasional testing.   When a freeze is invoked, changes
could only proceed once the requirements of the ECP process
are met.  The parties to the controversy now have an incentive
to agree or move on, not all Wiki authority has been stripped
away for excessive controversiality.  Only unilateral editing
of the controversial subject page undergoing excessive flip
flop editing or "edit war".

>If they're really
> *serious* about becoming a real contributor, they'll just have to wait for
> the ban to expire or appeal to the list.

It is my impression that we get more casual contributors that
become serious with increasing contribution and recognition
of the long term value, than that people who show up intending 
initially to be seriously committed long term contributors.

In other words, we have a buy in process that works if
people do not face too high a bar to begin contributing or do
not suffer burn out attempting to do too much work themselves.

Regards,
Mike Irwin