[Talk-GB] FWD: Re: House number ranges that are only odd or even

Mateusz Konieczny matkoniecz at tutanota.com
Thu Dec 10 22:01:47 UTC 2020




Dec 10, 2020, 21:51 by sk53.osm at gmail.com:

> However, I would regard > the Dutch <https://www.openstreetmap.org/#map=19/52.34367/4.77750&layers=N>>  & Polish communities approach of adding individual
> nodes for each address in the building irrespective of the actual address position outline
> as incorrect mapping in the UK. In both cases, and probably > also in Denmark <https://www.openstreetmap.org/#map=19/55.68777/12.58382&layers=N>> , this is most
> likely because addresses have been imported from a national database and this allows
> incremental updates from the same source. The problem with this is that it prevents classic
> OSM iterative refinement, such as accurate mapping for indoor usage, for instance to enable
> guidance for blind people. 
>
At least in Poland separate nodes for addresses are preferred as this:

- more accurate and allows to specify where given address actually is
- for example after mapping entrances, you can be guided to a correct one
- I am confused why it prevents 
"OSM iterative refinement, such as accurate mapping for indoor usage"
(maybe in UK addresses are assigned differently than in Poland)
- maybe it is related to fact that I am unaware of "address position outline"
existing in Poland - address is de facto assigned to building/plot/entrance
and in rare cases to complex objects such as a hospital or group of entrances
- it is common to have on street corner address from two streets in one building
(see https://www.openstreetmap.org/?mlat=50.07413&mlon=19.93361#map=19/50.07413/19.93361
and three nearby buildings), mapping this as an interpolation would not work
(and least I think so)
- and yes, is easier to map and import

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-gb/attachments/20201210/612d1f82/attachment.htm>


More information about the Talk-GB mailing list