[Tagging] Antwort: Re: Aerialway stations

dktue email at daniel-korn.de
Fri Aug 14 07:24:00 UTC 2020



Am 13.08.2020 um 19:06 schrieb Colin Smale:
>
> On 2020-08-13 18:35, Werner.Haag at leitstelle.tirol wrote:
>
>> Hi,
>> in my opinion (i think dktue is right there) it should be easy for a 
>> user to distinguish or extract (overpass query) the upper, mid and 
>> lower stations.
>> That´s not possible at the moment in OSM. Elevation (ele tag) may be 
>> useful, but does not indicate whether a station is the lower or upper 
>> one.
>> On the other hand, a tagging with upper_station or lower_station is 
>> clear and self-explanatory.
> It would also be a denormalisation of the data. In SQL terms you would 
> use a subquery like "WHERE ele = MIN(SELECT ele FROM stations 
> WHERE...)" although I suspect the usual use case will be looking for 
> the whole line, so a simple "ORDER BY ele" would give you all you need 
> to know - bottom station is first row, top station is last row, rows 
> 2..n-1 are mid stations. This would avoid any possibility of 
> conflicting information, e.g. multiple stations tagged as top, or the 
> station with the lowest elevation being tagged as top station.
> All this is based on the assumption that the definition of the top 
> station is the one with the highest altitude.... If any other factors 
> are in play that could mean that this definition does not always hold 
> true, then I am keen to hear.... It's best to check assumptions, even 
> (especially?) if they do appear obvious.
I agree that this is denormalization, but: We still do this in OSM with 
addresses and zipcodes for example and this is a good thing to do: 
Because it allows to find potential errors in the data in an automated 
way. I think it's very hard to find a wrong ele-Tag but it would be easy 
to hint in a quality assurance tool that ele-Tag and station-Tag do 
conflict.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20200814/89861447/attachment.htm>


More information about the Tagging mailing list