Mailing List Archive

Erik Moeller wrote:
> If we set a cookie in the browsers of all Wikipedia visitors, anonymous or
> not, we could the assign them random global user IDs. Instead of banning users
> by IP, we could ban them by GUID, which would eliminate the risk of
> accidentally banning legitimate contributors.
>
> While the majority of users have cookies enabled, a minority does not, so
> "soft bans" as I like to call them would not work for them. Other users might
> be smart enough to turn cookies off to avoid the ban. But I consider both
> beyond the technical understanding of most vandals, so I think soft bans might be
> quite efficient.
>
> What do you think?

This sounds like a reasonable tool, one among many.

We have yet to deal with a really persistant vandal. 99.9% of them just need to be
slightly discouraged.

--Jimbo

Erik Moeller wrote:
>
> If we set a cookie in the browsers of all Wikipedia visitors, anonymous or
> not, we could the assign them random global user IDs. Instead of banning users
> by IP, we could ban them by GUID, which would eliminate the risk of
> accidentally banning legitimate contributors.
>
> While the majority of users have cookies enabled, a minority does not, so
> "soft bans" as I like to call them would not work for them. Other users might
> be smart enough to turn cookies off to avoid the ban. But I consider both
> beyond the technical understanding of most vandals, so I think soft bans might be
> quite efficient.
>
> What do you think?
>

Seems like a step in a right direction to me.

If we must ban some people it would be nice
to be banning only the actual vandals.

Regards,
Mike Irwin

At 2002-11-10 20:56 +0100, Erik Moeller wrote:
>If we set a cookie in the browsers of all Wikipedia visitors, anonymous or
>not, we could the assign them random global user IDs. Instead of banning users
>by IP, we could ban them by GUID, which would eliminate the risk of
>accidentally banning legitimate contributors.
>
>While the majority of users have cookies enabled, a minority does not, so
>"soft bans" as I like to call them would not work for them. Other users might
>be smart enough to turn cookies off to avoid the ban. But I consider both
>beyond the technical understanding of most vandals, so I think soft bans might be
>quite efficient.
>
>What do you think?

Sounds like an idea worth considering.

Some other suggestions: Ban a whole range of
IP-addresses, except for people that did behave
in the past. You could for example give out
(coded) successive cookies for each range of
256*256 IP addresses. Whenever someone misbehaves
in that range, you ban the whole range, except
the people with a cookie with a lower number.
Because the numbers are coded it's hard to
guess a coded number that represents a lower
number.

Another idea is to ban not only based on the
IP-address-range but also on the kind of browser
and OS etc. This means that everybody in the
whole range would be banned unless they use
another browser/version or OS.

Another method could be to notify all former
users in the range by email and ask them to
contact a supervisor if they want to use
Wikipedia during the banning period.

By the way, I think people should not be allowed
to edit articles on Wikipedia when they don't
have logged in. That way you also have a better
handle on vandals.

It seems to me that there are probably many
more ways to tackle this problem.

Greetings,
Jaap

Jaap van Ganswijk wrote:
[cut]
> By the way, I think people should not be allowed
> to edit articles on Wikipedia when they don't
> have logged in. That way you also have a better
> handle on vandals.
[cut]

Er zal dan wel minder vandalisme zijn maar dan leg je de lat wel een
stuk hoger voor mensen om mee te werken. Hoe eenvoudig de registratie
ook is, het zal mensen tegenhouden. En er zijn anonieme medewerkers die
zeer waardevolle bijdragen leveren. Het zou zeer jammer zijn om die uit
te sluiten.


There will by less vandalism but you make it more difficult for people
to work at wikipedia. How simple the recording also is, it will stop
men. And there are anonymous users who provide very good contributions.
It would be very unfair to exclude those.

On Monday 11 November 2002 04:00 am, wikitech-l-request@wikipedia.org wrote:
> If we set a cookie in the browsers of all Wikipedia visitors, anonymous or
> not, we could the assign them random global user IDs. Instead of banning
> users by IP, we could ban them by GUID, which would eliminate the risk of
> accidentally banning legitimate contributors.
>
> While the majority of users have cookies enabled, a minority does not, so
> "soft bans" as I like to call them would not work for them. Other users
> might be smart enough to turn cookies off to avoid the ban. But I consider
> both beyond the technical understanding of most vandals, so I think soft
> bans might be quite efficient.
>
> What do you think?
>
> Regards,
>
> Erik

This seems like a good idea. However many browsers now have the ability to
easily block cookies so to combat this we would have to require everyone to
accept cookies in order to edit - which I don't think is too Draconian.

It would be great to have this /in addition to/ the current IP blocks but
please keep this functionality -- it is needed to at least temporarily block
the more technically savvy.

-- Daniel Mayer (aka mav)

Erik Moeller wrote:

>If we set a cookie in the browsers of all Wikipedia visitors, anonymous or
>not, we could the assign them random global user IDs. Instead of banning users
>by IP, we could ban them by GUID, which would eliminate the risk of
>accidentally banning legitimate contributors.

>While the majority of users have cookies enabled, a minority does not, so
>"soft bans" as I like to call them would not work for them. Other users might
>be smart enough to turn cookies off to avoid the ban. But I consider both
>beyond the technical understanding of most vandals, so I think soft bans might
>be quite efficient.

>What do you think?

I don't think that this is particularly *soft*,
just less error prone that banning by IP.
That is, it's less prone to errors of one type (banning the innocent)
and more prone to errors of the other type (not banning the guilty).

Will it work?
We could test the situation by giving anonymous users such cookies
and seeing how often the same IP comes back without the same cookie.
Then we could try to determine if any of these cases are the same person.
If not, or if rarely, then this method will have a good chance of working,
on the technically unsavvy.


-- Toby