[talk-au] A way to go.

Ian Sergeant inas66+osm at gmail.com
Mon Nov 14 22:28:06 GMT 2011


I started a thread on talk a while ago about how to handle this situation,
but it went nowhere, with the thread (as usual) hijacked to talk about
licencing issues unrelated to the practicality of implementation.

http://www.mail-archive.com/talk@openstreetmap.org/msg39790.html

My suggestion was a mod to the API to allow an earlier version of an object
to be modified and saved back the database, essentially ignoring or hiding
an intermediate version where there intermediate version contained nonCT
data.

This solves several problems of this type, but in the simplest case, if an
object has been modified to bulk add a single tag, a subsequent editor can
choose to ignore/hide that revision, editing the previous one.

I appreciate that there are probably cases where using a previous version
where nodes have been added that are shared by other objects, etc, would be
too complex to be handled automatically.  In this case, however, the API
could just refuse the request, and the editor is no worse off.

There are many many objects that are marked as tainted for almost trivial
changes made to the data. Better to allow nonCT data to be removed/hidden
by an editor when they are focussed on editing the object, than removing it
later in an automated way.

Ian.

On 14 November 2011 17:25, Nick Hocking <nick.hocking at gmail.com> wrote:

> Richard Wrote......
> """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
>
> On Sun, Nov 13, 2011 at 8:10 PM, Nick Hocking <nick.hocking at gmail.com>
> wrote:
> > PAST
> >
> > 1) decliner traces a way into OSM
> > 2) acceptor surveys road and copies in street sign info and turn
> restriction
> > info into OSM
> > 3) decliner bot-a adds maxspeed info
> > 4) decliner bot-b fixes incorrect maxspeed tag
> > 5) decliner adds numlanes info
> > 6) acceptor realigns way to smooth it out
> > 7) acceptor  unabbreviates st to Street
> > -------------------------------------------
> > FUTURE
> > I think that what should happen now is
> > a) Acceptor deletes way in its entirety.
> > b) Acceptor traces way from Bing imagery.
> > c) acceptor adds in info from 2) and 7) (info from 3 4 5 6 is lost)
> > The way is now (IMO) squeaky clean to be included under any OSM license.
>
> Some have argued that 3) and 4) would be fair game to promote to
> future OSM as well.  The argument goes like this.  Purely mechanical
> edits can not attract copyright as they are not an original creation;
> they are merely the implementation of a rule.
> Recreating the work of such a 'bot is trivial.  Such ways are
> artificially inflate the tainted way count.
>
>
> """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
>
> If the DWG and the LWG are happy with this then maybe they could
> alter the owenership of type 3 and 4 edits to be an anonymous accepting
> contributor.
>
> If the LWG or DWG are not happy with this then maybe they could revert the
> effect of
> the type 3 and 4 edits without affecting any edits either side of them.
>
> This would make the whole situation a lot clearer and enable us to easily
> determine
> which roads to remap or trace&name-copy.
>
> Nick
>
>
>
> _______________________________________________
> Talk-au mailing list
> Talk-au at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-au
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-au/attachments/20111115/0adfd88e/attachment.html>


More information about the Talk-au mailing list