[OSM-talk] Draft Geocoding Guideline

Thu Jul 20 18:35:23 UTC 2017

Thank you to everyone who looked over the draft Geocoding Guideline that
Simon circulated last month. The comments were very helpful.

The license working group (LWG) discussed the feedback at our last meeting,
and we've updated the draft to incorporate the following suggestions:
 - Expanding the "substantial extract" limitation to cover ref tags, house
numbers, and similar tags
 - Making the attribution obligations more specific by referencing Section
4.3 of the ODbL

These changes are now reflected in the wiki: https://wiki.openstreetmap.
org/wiki/Draft_Geocoding_Guideline

We also considered other suggestions:

 - There was a suggestion to remove the definitions of Direct Hit and
Indirect Hit. While we agree that the underlying rule is the same for both
types of results, we do believe there is a practical difference: more
Indirect Hits (typically interpolated results) will generally be necessary
to infer OSM data than Direct Hits. We think spelling this out will help
future users understand and apply the guideline.

 - There were a couple suggestions to further narrow the substantial
extract limitation ("all or substantially all" -> "majority",
sub-city-level geographic limitation). We think the combination of the
current geographic, tag, and proportion limitations are sufficient to deter
levels of abuse that would meaningfully compete with or pull contributions
from OSM, while the original reasons for the approach in the draft --
clarifying and facilitating typical, non-problematic geocoding use of OSM
-- remain. There was a related question about comparison of the draft to
the Substantial Guideline: in contrast to the Substantial Guideline, the
draft Geocoding Guideline rule applies only where the data in geocoding
results is limited to specific tags and excludes even results that are
small subsets of features within a geographic area (i.e. any group of
primary features). In contrast, the Substantial Guideline addresses when it
would be acceptable to extract feature data in its entirety, or all the
features in a geographic area. So it’s natural that the geographic
restrictions are different. We discuss our thinking on this a bit in the
FAQ on the wiki as well.

Again, we've greatly appreciated the comments to date. We'll be running
this second consultation phase on the revised draft through the next two
weeks.

Thanks to all the LWG members who've contributed thoughts and comments, and
to Simon for his leadership of the working group as always.

-Michael

On Mon, Jun 5, 2017 at 6:04 AM, Martin Koppenhoefer <dieterdreist at gmail.com>
wrote:

>
> 2017-06-02 10:32 GMT+02:00 Christoph Hormann <osm at imagico.de>:
>
>> > Individual Geocoding Results that are based on an Indirect Hit
>> > contain no OSM data and so are free of any obligations under the
>> > ODbL
>>
>> ....
>> IMO it would make sense to remove this distinction because the guideline
>> makes no significant difference between these two cases.  And even if
>> indirect hits contain no OSM data they are clearly derived from OSM
>> data.
>>
>
>
> +1
>
> More questions concerning:
>
>
> "A collection of Geocoding Results is not a substantial extract of the OSM
> database provided:
> ...
> 2. the collection is not a systematic attempt to aggregate all or
> substantially all Primary Features of a given type (as defined in the
> Collective Database Guideline) within a geographic area city-sized or
> larger."
>
>
>
> 2.1 "the collection is not a systematic attempt to aggregate all or
> substantially all Primary Features of a given type"
>
> what if it is not about Primary Features but certain properties (e.g.
> query in a grid for many points for the language of the name tag of nearby
> features and then construct language areas by creating hulls around points
> with the same language. With this guideline it would not fall under ODbL,
> but IMHO would clearly by a derivative database.
>
>
> 2.2 "all or substantially all Primary Features"
>
> If I throw away an arbitrary 5-10% of the results of all places in the
> world, it would be ok to store name and coordinates in a db and use it
> without any license obligation? Isn't "all or substantially all" quite
> exclusive? Why not for example "a majority"?
>
>
> 2.3 "geographic area city-sized or larger. "
> How big is a "city", which definition of "city" is used? If I use the
> method to get coordinates for all addresses of Manhattan, this would
> clearly be less than "city size"? Or 2 thirds of Rome?
>
> What is the reason we don't use the introduced "substantial" criteria as
> defined here, for geocoding as well:
> https://wiki.osmfoundation.org/wiki/Licence/Community_
> Guidelines/Substantial_-_Guideline
>
> In particular "The features relating to an area of up to 1,000 inhabitants
> which can be a small densely populated area such as a European village or
> can be a large sparsely-populated area for example a section of the
> Australian bush with few Features." which is way smaller than "city sized".
>
> Cheers,
> Martin
>
>
>
>
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
>

-- 
This is a private email.  Please check with me before forwarding, as it may
include information that's confidential or protected by the attorney-client
privilege.  If you feel like this email was sent to you by mistake, please
delete it and let me know. Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20170720/faee529c/attachment.html>