[OSM-legal-talk] feedback requested

Robert Whittaker (OSM) robert.whittaker+osm at gmail.com
Tue Dec 27 18:30:43 GMT 2011


On 24 December 2011 19:32, Frederik Ramm <frederik at remote.org> wrote:
> I have prepared changes to the OSMI map that allow me to
> * treat any tags contributed by a non-agreeing mapper as harmless if
>  these tags are not present any more in the current version

Are you sure that this is a good idea?

I can think of lots of times where I've improved the tagging of an
object by correcting / amending existing tagging using the information
contained in that previous tagging (and no other specific knowledge of
the object) to inform what the new tagging should be. This could
include correcting obvious typos, correcting common mis-taggings and
normalising multiple values to the recommended version contained in
the wiki. I'm sure others will have made similar edits. For example,
one might make the following 'obvious' changes whilst editing:

* name=Tesco's --> name=Tesco (correcting an obvious typo in a tag value)

* designated=public_footpath --> designation=public_footpath
(correcting an obvious typo in a tag key)

* note=signed as a public footpath --> designation=public_footpath
(converting a note tag into recommended tagging)

* shop=yes, type=coffee shop --> amenity=cafe, cuisine=coffee_shop
(improving tagging in line with the wiki recommendations)

In all these cases, the pertinent information in the new tagging can
be derived entirely from the old tagging, without the use of any other
source. So unless there's an explicit source tag for the new tagging,
I think we would have to play it safe and regard the new tagging as a
derived work of the old tagging. Therefore if the original tags were
added by a non-agreer, and an agreeing mapper made the type of change
above, I don't think it can be argued that the object is now
definitely clean. But if I've understood your proposed system, if
those changed / altered tags were the only tags added by a non-agreer,
the object would be automatically seen as clean.

(The first example I gave above may not get marked as clean as there's
still a name=* tag in the later version -- it's not clear from what
you've written if this "tags are not present any more" means the full
tag (key--value pair), or the key value only. But even if the first
example isn't included, I think the scheme will still give the wrong
result for the other three examples.)

Robert.

-- 
Robert Whittaker



More information about the legal-talk mailing list