[OSM-talk] mapping bridges - some ideas

Robin Paulson robin at bumblepuppy.org
Sun Feb 10 22:50:41 UTC 2013


i've been thinking about this for a while and it's bugging me, so i'd 
like to hear what the rest of you think.

the way we map bridges at the moment is by adding tags to the 
road/railway/footpath/etc., something like this:
bridge = yes
layer = 1

which seems a bit clumsy, and doesn't reflect very well what's 
happening.
the problem shows itself best when there is one bridge with several 
ways on it, for example this near where i live:
http://www.openstreetmap.org/?lat=-36.850112&lon=174.793894&zoom=18&layers=M

and this:
http://www.openstreetmap.org/?lat=-36.830896&lon=174.745834&zoom=18&layers=M

all of these roads/railways go over a single bridge, but from the 
mapping each appears to have its own separate bridge, that is one bridge 
per way. i'm sure there are many other examples

what i would like to hear feedback on is this:
perhaps we change the way we map this, and instead map the bridge and 
the way as entirely separate entities. so, the bridge could be 
represented with a rectangle, tagged as follows:

man_made = bridge
layer = 1
name = Auckland Harbour Bridge

the ends of the rectangle would be traced over the ends of the bridge 
structure. then the ways go over, but they no longer have bridge tags or 
layer tags applied to them, as they are applied to the bridge instead. 
it would also allow tagging of the bridge without interfering with 
tagging of the road which goes over it.

the second place this will benefit us is the situation where there are 
bridges with nothing on them. i'm thinking abandoned rail bridges in the 
ex-industrial cities of uk, where the rails have been ripped up, leaving 
nothing to apply the bridge tags too, but there are more than likely 
other examples.

we could maybe apply a similar logic to cuttings, tunnels and similar 
features

so,
comments, suggestions, revisions, problems, etc. please?

cheers,

-- 
robin

http://universitywithoutconditions.ac.nz - Auckland's Free University



More information about the talk mailing list