[OSM-talk] Cycleways wiki doc enhanced
claus.hindsgaul at gmail.com
Fri Jan 1 22:15:23 GMT 2010
> On Tue, Dec 29, 2009 at 10:32 PM, Pieren <pieren3 at gmail.com <http://lists.openstreetmap.org/listinfo/talk>> wrote:
> >* But it is documented in
> *>* http://wiki.openstreetmap.org/wiki/Key:cycleway since a while and is
> *>* about 100 times in osmdoc. The problem with cycleway=lane is that the
> *>* wiki never says clearly if it is for both sides and both directions or
> *>* if it can apply for one side only.
> I think cycleway=lane clearly implies that there are lanes on both sides,
> hence why replacing that with cycleway:left= is bad. If there is only one
> lane, then fair enough to not use cycleway=lane.
> To me it makes perfect sense to indicate that lanes/tracks are only present
in one side, as several examples on the page illustrates (M2 M4 and M5).
This is also the case for roundabouts btw. I myself have attempted to stay
backwards compatible in such cases with established tags using
"cycleway=lane" combined with "cycleway:left=none" in Copenhagen.
Such tagging will provide renderers with information to represent ways with
visible indication of tracks/lanes in only the correct side and routers with
information wether a road must be crossed.
> >* > 7) You don't make allowance for segregated cycleways
> *>* > ("tracks"/copenhagen-style lanes) that aren't represented as distinct
> *>* ways.
> *>* > Is "highway=residential, cycleway=track" not possible?
> *>* You mean for T1 and T2 but with only one way in OSM, right (the track
> *>* is not traced separately) ? I didn't know it was called the
> *>* "copenhagen-style". I will add it but not as a recommended solution.
> The term "copenhagen style bike lane" is widely used in Australia, seems to
> be used in New York, and maybe some other places - it's hard to tell.
> I am indeed talking about examples like T1 and T2. I don't see why it should
> be necessary to separately trace the bike path if it exactly follows the
> contour of the road. Just like we don't map pavements and lanes and stuff
> like that.
On the talk-dk list we have had a lengthy discussion on how to best
represent such "copenhagen" style bike tracks. Since nearly every major road
in our cities are provided with paved dedicated bicycle tracks aligned
closely with their sides, separated by a curb or a curb and a bit of grass,
it would add a tremendous number of "shadow ways" to represent these as
Making the right connections at junctions is not trivial either; e.g. at
each T-junction, you would need separate micro-mapping ways to describe
access to the cycle track on the opposite side of the road.
In addition, routing software would not be able to suggest a U-turn until
the next junction, as there would be no connections to the opposite cycle
Finally, the concept of being at the same road as the cars would be lost -
and routing software would have to tell you to "drive left at unnamed cycle
track" instead of "drive left at Oxford Road".
The complexity of getting the connections right at junctions and traffic
lights with separate cycle tracks where demonstrated well in the
illustrations at this (Danish language) page by Morten Kjelgaard:
http://wiki.openstreetmap.org/wiki/Da:Cykelstier. Now, imagine the result of
adding representations of sidewalks separately as well in order to be
consistent. Horror! :-)
The outcome of the discussion was by default to represent bicycle
tracks/lanes with "cycleway=track/lane" tags in the accompagning road
instead of separate "cycleway=highway". The following expressed exceptions
were agreed upon and is now pinned out in the Danish guideline (
- distances where the path/contour of the bicycle track differs
significantly from that of the road
- distances where the bicycle track is separated from the road by a
barrier not easily nor legally passed by a bicycle (a curb is regarded as
- distances where the bicycle track has a significant distance to the
road side (~ >5 meter)
- distances where permitted traffic directions of different vehicles can
not otherwise be correctly described (this bullet could well be elliminated
by the present discussion)
- distances where car road and bicycle track properties differs
significantly (e.g. paved road and dirt bicycle track)
Could a similar guideline be useful for other countries?
Civilingeniør ph.d. Claus Hindsgaul
Edvard Thomsens Vej 19, 5. th
DK-2300 KBH S
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the talk