[Tagging] addr:street=* combined with place=square, name=*

Johannes Singler johannes at singler.name
Tue Aug 14 15:27:06 UTC 2018


I understand that it is useful to use addr:place for neighborhoods, 
hamlets, and isolated dwellings etc.  But here, it is a quite regular 
street address, just that the referenced feature is not a highway, but a 
square (we could limit it to place=square).  So why should this be ruled 
out categorically?  It does not read addr:highway, does it?

I think OSM Inspector should check that there is *some* entity close by 
that matches the street name, to avoid spelling mistakes etc.  In 
another case, the street name actually references a park, e.g here
So should I reference that with addr:park?  Or map the park as a place, 
or as a highway?  Rather not, eh?

So I propose to be more flexible here.  Too many "false positives" in 
the QA tools are frustrating to the users, and shadow the real mistakes.


> Hi
> I'd rather use addr:place="Square Name" in that case. In don't agree
> that addr:place is 'intended for larger objects like "villages,
> islands, territorial zones"'. I also use addr:place e.g. for
> settlements (place=neighbourhood) or hamlets, if there is no street
> with the addresses' name (example: [^1]).
> [^1]: <http://tools.geofabrik.de/osmi/?view=addresses&lon=7.59448&lat=47.54290&zoom=18&overlays=buildings,buildings_with_addresses,postal_code,entrances_deprecated,entrances,no_addr_street,street_not_found,place_not_found,misformatted_housenumber,nodes_with_addresses_defined,nodes_with_addresses_interpolated,interpolation,interpolation_errors,connection_lines,nearest_points,nearest_roads,nearest_areas,addrx_on_nonclosed_way>
> Regards
> Markus
> On Mon, 13 Aug 2018 at 21:05, Toggenburger Lukas
> <Lukas.Toggenburger at htwchur.ch> wrote:
>> Hi
>> I'm the main author of the address view of Geofabrik's OSM inspector: http://tools.geofabrik.de/osmi/?view=addresses , a QA tool for OSM, whose sourcecode you can find at https://github.com/ltog/osmi-addresses/
>> Some time ago I received the following issue and subsequent pull request:
>> - https://github.com/ltog/osmi-addresses/issues/111
>> - https://github.com/ltog/osmi-addresses/pull/115
>> The submitter johsin18 proposes the following:
>> Given a (node|way) with addr:street=theName and a (node|way) with place=square, name=theName, the first object should logically be tied to the second. Correspondingly, osmi-addresses should recognize this and not display it as an error as it is currently the case, e.g. at: http://tools.geofabrik.de/osmi/?view=addresses&lon=7.59448&lat=47.54290&zoom=18&overlays=buildings,buildings_with_addresses,postal_code,entrances_deprecated,entrances,no_addr_street,street_not_found,place_not_found,misformatted_housenumber,nodes_with_addresses_defined,nodes_with_addresses_interpolated,interpolation,interpolation_errors,connection_lines,nearest_points,nearest_roads,nearest_areas,addrx_on_nonclosed_way
>> osmi-addresses currently expects either
>> addr:street=* used in combination with highway=*, name=*
>> or
>> addr:place=* used in combination with place=*, name=*
>> Both myself and the current maintainer of osmi-addresses (=Nakaner) are unsure if this proposed change would be appreciated by the larger public or not. We are therefore seeking your opinion.
>> Best regards
>> Lukas
>> _______________________________________________
>> Tagging mailing list
>> Tagging at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/tagging

More information about the Tagging mailing list