[OSM-talk] Potlatch and the destruction of good work (N6 Ireland)

Frederik Ramm frederik at remote.org
Wed Jul 25 19:50:04 BST 2007


> I'm not the first to suffer from this, and it can't be avoided in an
> open community project, but I have to add my voice to the others
> recently who have complained at the great ease with which an anonymous
> user (who it's hard to ask to stop) can cause damage to the existing
> data set that's difficult to reverse.

I am very much in favour of contributions without having to open an 
account first, but this is not currently possible.

The "anonymous" user of whom you speak has entered an E-Mail address and 
clicked on a link sent to that address - he was just too lazy or to busy 
with other things to note the "make my edits public" link somewhere.

I suggest:

(1) All new users should automatically be "public".

(2) All existing users should be given the following notice:

"Dear user, we are phasing out non-public accounts. Please make your 
edits public. If you don't, we will close your account on <date>; you 
can then open a new account but we only hand out "public" accounts. The 
data you have already contributed will remain anonymous in that case."

(3) If people are against that, then the Web page needs to at least 
contain a mechanism "send message to user who last edited way/node id 
#.....". That way those people would be reachable even if they remain 

> Personally, I'd favour a world where you can't edit anonymously, at
> least using Potlatch. 

Anonymous editing helps to reduce the barrier of entry, and I'm very 
keen on that. Granted, there's potential for trouble, but if we (like 
Wikipedia) store the IP address in lieu of a login name in these cases, 
you will have a good chance to "group" edits.

> The alternative to safeguards against rash edits is for us all to keep
> copies of all bits of the map that we've touched, and even then, in a
> densely mapped area, restoring broken stuff will be very difficult. We
> need to work out how to manage with "too many cooks".

Yes, finding good ways of (partially) rolling back areas without 
creating tons of conflicts is a very challenging task but we'll not get 
away without it.

> Finally, I need some advice on how to go about fixing things without
> causing further strife. Somewhere out there is a mapper who thinks
> it's valid to tag a roadway that doesn't exist yet. I could take the
> counter-position and simply wipe it, but that invites him to put it
> back again and maybe break more of the existing data in the process.
> Not good.

node place=hamlet name="please don't break my roads" ;-)

No, honestly, this is something we also need - a way to stick "editing 
notes" like post-its onto our maps so that others editing in the same 
area can be alerted to something you want them to know.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00.09' E008°23.33'

More information about the talk mailing list