[OSM-talk] anonymous edits

[Martijn van Exel]

On Thu, May 26, 2011 at 10:09 PM, Frederik Ramm <frederik at remote.org> wrote:
> Hi,
>
> Martijn van Exel wrote:
>>
>> That is exactly why I started this thread - to see how (un)acceptable
>> it is to do (semi) anonymous edits.
>
> An important reason against anonymous edits is accountability. We want to be
> able to contact someone and ask them: Why did you add that? What did you
> mean by it? Etc.

That was my initial thought when the idea came up. Of course you can
trace an edit back to a twitter user, but that's one step away from a
community member. It's not 100% anonymous, but at least someone who
came to the OpenStreetMap web site and made an account (however little
effort that takes) at least had a look at OpenStreetMap.

But then consider this. We have 400k+ "community members" of whom, if
the ratio has not changed, about 80% never made any edits. Of the 20%
remaining, about 80%% is inactive (no editing in the last two months)
leaving only about 6% of those 400k+ active mappers. This tells me two
things:
1. We need more ways to engage new mappers, to retain them once they
show an interest in OSM. The proposed twitter scheme could be one of
them.
2. 'Having an OSM account' does not really tell me anything about the
involvement of that particular person. Or of their ability to be
contacted / held accountable, for that matter.

The one thing I do see as a real issue is the legal one. All twitter
POI contributions would be added through one account - at least
initially, see below. This account has to be made by a real person and
that person is legally bound to the license and CT. Will this person
or legal entity need permission from the people actually using the
twitter scheme to be in conformance with the license / CT? How does
Wheelmap.org handle this, if at all?
>
> In order not to burden the mapper, your twitter bot would somehow have to
> establish bidirectional communiactions between mappers and twitter-ers. If
> you think you cannot expect the twitter-er to set up an OSM account, then in
> the same vein you cannot expect the mapper to set up a twitter account. You
> must make sure that a message sent by a mapper to your twitter bot actually
> reaches the twitter-er who is the source of the data. This is probably not
> easy.

I can see that happening. After a few additions, the scraper
application may send a tweet encouraging the user to create an
account. He would then be pointed to a web site that
1. would associate his twitter account with the application through
twitter's oauth;
2. guide the user through the steps of creating an OSM account;
3. associate the OSM account to the application and to the twitter account.
>From then on the application could post the POI additions through the
user's account, and allow for more edits. And we would have attracted
one more user that already showed an interest in making real
contributions.

>
>> I was surprised to see the
>> wheelmap construct but I'm sure that was discussed here before it was
>> implemented. Was it?
>
> Don't geht the "wheelmap visitor" thing wrong; the *only* thing that this
> visitor can do is to set one specific tag to one of three specific values on
> an already-existing object of a certain type. So, no free-form tagging, no
> creation of new objects - almost zero risk of vandalism or copyright
> violation. But even there we have already had problems where an unknown
> "wheelmap visitor" changed something that others found worthy of discussion.

Fair enough, that is fairly well constrained. On the other hand, it is
also 'more anonymous' in that there is no way the edit can be traced
back to a real person, whereas the contributions through the twitter
scheme could at least be traced back to a user account.

> [...]

Best,
-- 
Martijn van Exel
http://about.me/mvexel