[OSM-talk] Deleting data

Ian Sergeant inas66+osm at gmail.com
Sat Oct 19 07:46:15 UTC 2013


I think it is self-evident that correct data shouldn't be deleted from
OSM.  I'd be surprised if anyone actually disagrees with that.

However, frequently as an editor I have to make decisions as to the
correctness of data in the database.   In the areas I work there is a
lot of data that is just plain wrong, and part of correcting data is -
sometimes - deleting it.

I look for several cues to assist my decision making whether data is
correct.  If I'm familiar with the mapper and their work.  Whether the
mapper has entered information other than what can be determined from
imagery.  The source tagging on feature and changeset.  The date of
the imagery.  The date the object was added to the database.  If the
mapper is local.

It's a complex assessment.  Notes and source tagging really do assist
it finding what is correct when imagery and the database clash.
Extensive surveying is ideal, and I'd never dream of changing recently
well surveyed data to replace with data traced from imagery.
However, sometimes that's not the clear choice that is presented, and
quite often the changeset and object source tags are no help in
determining that.


On 19 October 2013 16:05, Russ Nelson <nelson at crynwr.com> wrote:
> This is wrong, and it's got to stop. Nobody should be deleting data
> that somebody else entered unless they have actually BEEN to the
> place, failed to see any trace of the mapped entity, and are an expert
> at identifying the mapped entity. This should be the case no matter
> the entity: whether fire hydrants, buildings, intersections, park
> benches, railroads, canals, whatever.

More information about the talk mailing list