[OSM-talk] RR8 - Possible International Vandal (assistance required in various countries)

David Earl david at frankieandshadow.com
Wed Sep 2 00:10:55 BST 2009


On 01/09/2009 23:40, Tom Hughes wrote:
> On 01/09/09 23:02, Pieren wrote:
> 
>> I'm just questionning myself if I will continue to contribute to OSM
>> if the admins are not able to react faster to something which looks
>> like the worst form of vandalism.
> 
> As an admin with the technical ability to do these things I'm perfectly 
> capable of reacting quickly once I believe I have legitimate authority 
> to act. I do not however plan to appoint myself as judge, jury and 
> executioner in these matters.
> 
> What we are lacking is not people to take action, but mechanisms and 
> people to quickly investigate and make decisions on what action should 
> be take in an appropriate transparent, democratic and legitimate manner.
> 
> A group of people demanding that "something by done" on a mailing list 
> does not, in my mind, constitute a legitimate authority for me to act.
> 
> After all no matter how "obvious" this or any other case might be in 
> general terms I can't possibly personally determine who is right in any 
> particular case, nor can I set any sort of quorum as to how big the 
> lynch mob on the mailing lists needs to be to trigger action because I 
> can't know if they are people with a legitimate grievance or are indeed 
> just a lynch mob.

Indeed, that puts you in an invidious position.

I had a look at a few of his edits, and while a B road in the UK 
midlands was simply invention, a change to the A9 in the far north of 
Scotland was indeed correct (though incomplete). A node for Keflavik 
airport in Iceland was created in one of its car parks (when it already 
had one nearby).

On the general point, I wonder whether a mechanism to lock the objects 
affected by a particular changeset from further change might be useful 
so that when we have a suspected case we can freeze its objects until we 
can make a decision, so that it can be simply reverted rather than 
ending up with conflicting later changes (often people trying to make 
manual corrections)?

Automatically locking further changes by a particular user may also be 
useful once suspected - more so than banning them, as a determined 
individual will just open a new account.

We've seen two or three users now making calculated, plausible edits 
that are actually wrong, which is a much harder case to deal with than 
wholesale crude deletion or scribbling because it needs local knowledge.

Ireland people, why don't you just revert the changesets you personally 
know to be wrong yourselves - any user can do this, and Frederick's perl 
script is publicly available in svn and isn't hard to set up and use.

David





More information about the talk mailing list