[OSM-legal-talk] Updated geocoding community guideline proposal

[Simon Poole]

Am 27.07.2014 23:52, schrieb Alex Barth:
> 
> On Fri, Jul 25, 2014 at 11:39 AM, Simon Poole <simon at poole.ch
> <mailto:simon at poole.ch>> wrote:
> 
>     If you apply this to your above example, the addresses would be subject
>     to SA (however no further information), and while potentially one could
>     infer that these are likely the addresses of the store locations, no
>     further information would needed to be disclosed*.
> 
> 
> So I think I follow: in a database of store locations [1], where
> coordinates have been added through OSM-based geocoding, only the
> coordinates (latitude/longitude pairs) from OpenStreetMap are subject to
> share alike. The store names, street names, house numbers, etc. wouldn't
> be subject to share alike, they didn't come through the OSM-based
> geocoder - nor any coordinates that haven't been added through the
> OSM-based geocoder.

Actually in the model what was used to extract the coordinates from OSM
-would- be subject to share alike. In the example the resulting
address/coordinate dataset.

So you would end with a database of geocoded addresses, subject to the
ODbL that then could be used to locate your proprietary data.

As I've said, the advantage of the model is it only depends on
recognizing a couple of principles

- the Fairhurst Doctrine
(http://wiki.openstreetmap.org/wiki/Open_Data_License/Metadata_Layers_-_Guideline),
or whatever you want to call it
- and likely allowing
https://wiki.openstreetmap.org/wiki/Open_Data_License/Community_Guidelines/Fall_Back


and not on making a determination if the way the data is used is still
compatible with it being a produced work.

A note on the later, the LWG guidance on produced works is deliberately
liberal because we wanted to include products delivered to the end user
and not primarily intended for further processing and distribution in
that category, even if they nominally could be considered databases.

> 
> While this reading is better than the uncertainty we have now it is not
> practical beyond well informed users. To appropriately handle geocoding
> under this practice, a geocoder application would not only have to
> expose on a granular level where data was sourced from [2] - but a
> geocoder user would have to store this information in a granular way to
> be able to release data appropriately.

I think you will find that this would be expected even in your model
(given that the attribution requirements are essentially the same). If
you would differentiate each individual result or just have a blanket
statement that is always displayed is likely just a matter of taste.

Simon

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 553 bytes
Desc: OpenPGP digital signature
URL: <http://lists.openstreetmap.org/pipermail/legal-talk/attachments/20140729/9e227cec/attachment.pgp>