[OSM-talk] Tag:highway=cycleway inconsistency

Michael Collinson mike at ayeltd.biz
Thu Jul 3 16:42:31 BST 2008


At 04:53 PM 7/3/2008, Andy Allan wrote:
>On Thu, Jul 3, 2008 at 3:31 PM, Michael Collinson <mike at ayeltd.biz> wrote:
>
> > I certainly agree that the default for foot access should be one or
> > the other for highway=cycleway.
> >
> > My own preference is for default foot=yes.
>
>I disagree - I think the default should be undefined. After all, it's
>been undefined for a while, and I don't like hidden assumptions that
>will be wrong for entire countries.

Yes, I see your point. A lot of my thinking centres around ease of 
data entry and the fact that when a lot of tags have to be entered, 
folks ... don't.  I take a default as a reasonable but not certain 
assumption. I wonder what our routing fraternity has to say.


>I've been asked by a cycling group how they should be marking shared
>vs segregated, so how about
>
>highway=cycleway, cycleway = shared - both bikes and pedestrians
>sharing the same path, common in the UK. Has the blue sign with the
>bike over the pedestrians.
>highway=cycleway, cycleway = segregated - the path has a line down the
>middle, cyclists on one side, pedestrians on the other
>highway=cycleway, cycleway = cycleonly - bikes only, like the 
>default in Germany

I'd support that.  One practical data entry question.  If  I haven't 
noted whether it is shared or segregated, go back to cycleway=track, 
foot=yes ... or? Tracks often switch from one to the other frequently.


>... but to be honest, I'm not entirely comfortable with it, and I
>still think the lane/lane_opposite doesn't handle things fully either.
>I found a bit in Hyde Park where there was a one-way road with cycle
>lanes on both sides - with all three lanes going in the same direction
>- and I don't know how to model that in OSM either.

Well, I've commented on the rest, so let me take a 
stab.  cycleway=lane.  Does it matter that there happen to be two 
such lanes and their position?

Mike 





More information about the talk mailing list