[Tagging] [OSM-talk] Cycle lanes & cycle tracks - my findings and a proposal

Martin Koppenhoefer dieterdreist at gmail.com
Tue May 22 15:52:57 BST 2012

2012/5/22 Martin Vonwald <imagic.osm at gmail.com>:
> 2012/5/22 Janko Mihelić <janjko at gmail.com>:
>> 2012/5/22 Martin Koppenhoefer <dieterdreist at gmail.com>
>>> -1, this has been discussed ad infinitum on the German list.
>> I'm sure this was already suggested, but I would like to know: what if the
>> sidewalk had the same name as the street, would that be enough? If the
>> router saw two highways with the same name, it would choose footway or path
>> instead of highway, if possible.
> You still wouldn't know if you could change from the footway/path to
> the street or not. Maybe there's some kind of barrier in between them.
> If you map them separately you simply lose the information what
> (cycleway, sidewalk, whatever) belongs together.
> And before some suggests it: no, I don't think a relation is a
> manageable solution in this case ;-)

Well, the relation offers greater benefit than simply associating
parts of the road: you can also add more parts (dividers/barriers)
which can be mapped explicitly or also only described implicity with
tags (e.g. you can have a dual carriageway and state that between the
two is a 1 metre high concrete wall, without drawing the wall, or you
could draw the wall and put them together with the road into the
relation, or you could have the wall implicitly (tags) in the data but
put the one point where it is interrupted as explicit geometry into
the relation, ...). This allows for mapping of raised and lowered
kerbs, walls, and other barriers between two (or more?) more or less
parallel ways which belong to the same road. You can map these
elements then without having to create "artificially" non-existing
(e.g.) footways to connect a pavement with the road via a lowered kerb
or similar.

See the area-relation-proposal if you are interested.


More information about the Tagging mailing list