[Osmf-talk] ODbL and Geocoding

[Simon Poole]

Am 14.02.2016 um 15:08 schrieb Christoph Hormann:
> My point here is that the idea of allowing cross license 
> feature-attribute recombination is something that - to my knowledge - 
> was previously not discussed in the sense it should be generally 
> allowed without share-alike.  I recognize the problem of drawing the 
> line (as i pointed out in my comments) but still it is IMO important 
> that the community discusses this on a broader scale.
>
> Note i do not have a firm standpoint in this matter myself, much less i 
> know others' views on the matter.  Which is exactly why i think a 
> broader discourse is important.  Sadly of course the matter is complex 
> enough to make it difficult to have a qualified open discussion that is 
> not dominated by specific interests.
>

The specific issue is that we (I hope everybody) would want the
guideline to be consistent and not be riddled with special cases.

One of the classical examples is combining proprietary restaurant
reviews with OSM data (we actually got this as an enquiry to
legal-questions some time back so it is actually real), my gut feeling
is that there is a rough consensus that this is an use case that we
would want to be possible (not much different than medical records and
similar examples that have been mentioned).

Now we typically associate a review with a longish piece on the place in
question stored separately from any OSM data, but naturally in reality
the review might just be a star rating that could be stored in an
key-value pair with an OSM object. I, at least, don't see a fundamental
reason why we should treat the one scenario different from the other. Or
why a long moan about road quality should be treated differently that a
third parties equivalent of smoothness=very_horrible :-).

As a consequence the proposed guideline doesn't differentiate either and
applies the same "all or nothing" principle equally, to both data that
we subjectively feel is "outside" of OSM, just as to things were we
could take the position this should really be "in" OSM.

Now the problem with being consistent is being consistent, and there may
be use cases covered by the guideline were we will all feel uneasy, but
I believe the proposal strikes a reasonably good balance between
supporting share-a-like and enabling combination with proprietary data
within reason.

Simon

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <http://lists.openstreetmap.org/pipermail/osmf-talk/attachments/20160215/588d3333/attachment.sig>