Guten abig,
I was looking at the data for Luzern Bahnhof. I noticed some weird polygons, but maybe it is expected, maybe it is wrong. Can someone share their opinion?
https://www.openstreetmap.org/relation/17482554 - this is a multipolygon with several members and seems to have the full outer edge, inner edges, is level=0, build_part=yes, is also highway=footway and building=yes
https://www.openstreetmap.org/way/1280137373 - this looks to be the same as the outer way of the multipolygon above. But this is not a child member of the multipolygon, and has sparse other tags.
is the way a duplicate of the relation? Or should these really both exist separately?
A secondary question: we have the railway station as a node: https://www.openstreetmap.org/node/2051794005 - should the relation and way above be named Bahnhof Luzern, considering the node is named Luzern, or should the relation and way just have no name, as train station buildings but not the representative "point of interest" that has a name?
Thank you for your opinions on this and I look forward to learning what is the best practice!
Freundliche gruesse,
Christopher Beddow
Hi Christopher The first multipolygon is wrong. a bilding=yes multipoligon with inner and outer ist used for houses all around an open place. In that case I would not use a multipoligon. An aerea is more clear. This is also the standard tagging for buildings. The open place ist shown as *no building* from the apps. I think the second multipolygon is for fixing that error. But I have not tested it. But I am sure that Josim for example produces a warning "overlapped buildings". *For me it is not clear how on can implement a building inside a building. I think it is not possible, * There is another problem with the Luzern Bahnhof. The "Bahnhofsgeläne Luzern" is a not closed multipoligon. This for sure produces in JOSM an error. BR Thomas I have no better answer.
Am Mi., 22. Apr. 2026 um 23:18 Uhr schrieb Christopher Beddow < christopher.beddow@gmail.com>:
Guten abig,
I was looking at the data for Luzern Bahnhof. I noticed some weird polygons, but maybe it is expected, maybe it is wrong. Can someone share their opinion?
https://www.openstreetmap.org/relation/17482554 - this is a multipolygon with several members and seems to have the full outer edge, inner edges, is level=0, build_part=yes, is also highway=footway and building=yes
https://www.openstreetmap.org/way/1280137373 - this looks to be the same as the outer way of the multipolygon above. But this is not a child member of the multipolygon, and has sparse other tags.
is the way a duplicate of the relation? Or should these really both exist separately?
A secondary question: we have the railway station as a node: https://www.openstreetmap.org/node/2051794005 - should the relation and way above be named Bahnhof Luzern, considering the node is named Luzern, or should the relation and way just have no name, as train station buildings but not the representative "point of interest" that has a name?
Thank you for your opinions on this and I look forward to learning what is the best practice!
Freundliche gruesse,
Christopher Beddow _______________________________________________ talk-ch mailing list -- talk-ch@openstreetmap.ch To unsubscribe send an email to talk-ch-leave@openstreetmap.ch
Thank you! Good insights.
For the multipolygon, it appears to me to be that the holes are where the escalator and stairs are located inside the train building. As if a hole in the floor. I think that is probably not a proper use of the inner ring of a multipolygon because it is still under the roof.
On Thu, Apr 23, 2026, 08:55 Thomas Eugster thomaseugster58@gmail.com wrote:
Hi Christopher The first multipolygon is wrong. a bilding=yes multipoligon with inner and outer ist used for houses all around an open place. In that case I would not use a multipoligon. An aerea is more clear. This is also the standard tagging for buildings. The open place ist shown as *no building* from the apps. I think the second multipolygon is for fixing that error. But I have not tested it. But I am sure that Josim for example produces a warning "overlapped buildings". *For me it is not clear how on can implement a building inside a building. I think it is not possible, * There is another problem with the Luzern Bahnhof. The "Bahnhofsgeläne Luzern" is a not closed multipoligon. This for sure produces in JOSM an error. BR Thomas I have no better answer.
Am Mi., 22. Apr. 2026 um 23:18 Uhr schrieb Christopher Beddow < christopher.beddow@gmail.com>:
Guten abig,
I was looking at the data for Luzern Bahnhof. I noticed some weird polygons, but maybe it is expected, maybe it is wrong. Can someone share their opinion?
https://www.openstreetmap.org/relation/17482554 - this is a multipolygon with several members and seems to have the full outer edge, inner edges, is level=0, build_part=yes, is also highway=footway and building=yes
https://www.openstreetmap.org/way/1280137373 - this looks to be the same as the outer way of the multipolygon above. But this is not a child member of the multipolygon, and has sparse other tags.
is the way a duplicate of the relation? Or should these really both exist separately?
A secondary question: we have the railway station as a node: https://www.openstreetmap.org/node/2051794005 - should the relation and way above be named Bahnhof Luzern, considering the node is named Luzern, or should the relation and way just have no name, as train station buildings but not the representative "point of interest" that has a name?
Thank you for your opinions on this and I look forward to learning what is the best practice!
Freundliche gruesse,
Christopher Beddow _______________________________________________ talk-ch mailing list -- talk-ch@openstreetmap.ch To unsubscribe send an email to talk-ch-leave@openstreetmap.ch
talk-ch mailing list -- talk-ch@openstreetmap.ch To unsubscribe send an email to talk-ch-leave@openstreetmap.ch
I think the holes are the kiosks and shops inside. I used them as a beer SOURCE when i reached luzern after some passes with the race bake. The kiosks saved me.
Christopher Beddow christopher.beddow@gmail.com schrieb am Do., 23. Apr. 2026, 09:25:
Thank you! Good insights.
For the multipolygon, it appears to me to be that the holes are where the escalator and stairs are located inside the train building. As if a hole in the floor. I think that is probably not a proper use of the inner ring of a multipolygon because it is still under the roof.
On Thu, Apr 23, 2026, 08:55 Thomas Eugster thomaseugster58@gmail.com wrote:
Hi Christopher The first multipolygon is wrong. a bilding=yes multipoligon with inner and outer ist used for houses all around an open place. In that case I would not use a multipoligon. An aerea is more clear. This is also the standard tagging for buildings. The open place ist shown as *no building* from the apps. I think the second multipolygon is for fixing that error. But I have not tested it. But I am sure that Josim for example produces a warning "overlapped buildings". *For me it is not clear how on can implement a building inside a building. I think it is not possible, * There is another problem with the Luzern Bahnhof. The "Bahnhofsgeläne Luzern" is a not closed multipoligon. This for sure produces in JOSM an error. BR Thomas I have no better answer.
Am Mi., 22. Apr. 2026 um 23:18 Uhr schrieb Christopher Beddow < christopher.beddow@gmail.com>:
Guten abig,
I was looking at the data for Luzern Bahnhof. I noticed some weird polygons, but maybe it is expected, maybe it is wrong. Can someone share their opinion?
https://www.openstreetmap.org/relation/17482554 - this is a multipolygon with several members and seems to have the full outer edge, inner edges, is level=0, build_part=yes, is also highway=footway and building=yes
https://www.openstreetmap.org/way/1280137373 - this looks to be the same as the outer way of the multipolygon above. But this is not a child member of the multipolygon, and has sparse other tags.
is the way a duplicate of the relation? Or should these really both exist separately?
A secondary question: we have the railway station as a node: https://www.openstreetmap.org/node/2051794005 - should the relation and way above be named Bahnhof Luzern, considering the node is named Luzern, or should the relation and way just have no name, as train station buildings but not the representative "point of interest" that has a name?
Thank you for your opinions on this and I look forward to learning what is the best practice!
Freundliche gruesse,
Christopher Beddow _______________________________________________ talk-ch mailing list -- talk-ch@openstreetmap.ch To unsubscribe send an email to talk-ch-leave@openstreetmap.ch
talk-ch mailing list -- talk-ch@openstreetmap.ch To unsubscribe send an email to talk-ch-leave@openstreetmap.ch
talk-ch mailing list -- talk-ch@openstreetmap.ch To unsubscribe send an email to talk-ch-leave@openstreetmap.ch
Hello,
Looking at the history, the 1st one described the footway area, not the building. Now building+building:part is wrong, building+footway too, building+level too For the 2nd building+buildint:part is also wrong.
For the name : it's look like the name of the area, including building, plateform, rails, ... And not the name of the building. So building shouldn't have a generic name.
Regards, Marc ________________________________ De : Christopher Beddow christopher.beddow@gmail.com Envoyé : mercredi 22 avril 2026 23:17 À : Openstreetmap Schweiz/Suisse/Svizzera/Svizra talk-ch@openstreetmap.ch Objet : [talk-ch] Luzern Bahnhof - duplicate polygons?
Guten abig,
I was looking at the data for Luzern Bahnhof. I noticed some weird polygons, but maybe it is expected, maybe it is wrong. Can someone share their opinion?
https://www.openstreetmap.org/relation/17482554 - this is a multipolygon with several members and seems to have the full outer edge, inner edges, is level=0, build_part=yes, is also highway=footway and building=yes
https://www.openstreetmap.org/way/1280137373 - this looks to be the same as the outer way of the multipolygon above. But this is not a child member of the multipolygon, and has sparse other tags.
is the way a duplicate of the relation? Or should these really both exist separately?
A secondary question: we have the railway station as a node: https://www.openstreetmap.org/node/2051794005 - should the relation and way above be named Bahnhof Luzern, considering the node is named Luzern, or should the relation and way just have no name, as train station buildings but not the representative "point of interest" that has a name?
Thank you for your opinions on this and I look forward to learning what is the best practice!
Freundliche gruesse,
Christopher Beddow