[OSM-talk] OSM relation ID property in Wikidata

andrzej zaborowski balrogg at gmail.com
Fri May 3 21:32:36 UTC 2013


On 3 May 2013 23:22, Frederik Ramm <frederik en remote.org> wrote:
> * some objects whose ID had not changed and who had been created by someone
> who rejected the license were nonetheless kept if it could be shown that
> they had been changed in a major way since;
>
> * some objects that had been freshly created by people agreeing with the
> license change, but that were more or less copies of other objects from
> non-agreers, were removed even though they had a different ID.

The redaction bot code doesn't generally do that and if there were
such cases then they were an insignificant minority compared to those
where the change of the object's ID meant it was considered an
entirely different object.  It's strange that you'd negate that.

>
> I will not discuss this sub-thread further; object IDs are not stable and
> nothing we did during the license change is suitable as a counter argument.

The fact that the IDs are not persistent had been pointed out several
times during the process and both you and the LWG have said (this is
quite clearly stated their meeting minutes) that this isn't an issue
big enough to bother.  There were some statistics posted on the list
and on IRC (from Simon Poole) stating it affected 0.1 to 1% of the
database which is in the millions of objects range.

Cheers



More information about the talk mailing list