Bonjour,
Le 11.05.19 à 16:26, Andreas Bürki a écrit :
sollte man auch an visuell Beeinträchtigte denken. Ohne von der Strasse separat gemapptes Trottoir dürften sie einige Probleme haben.
I am part of a thinking tank on osm and disabilities (cartomobilité in french). the separate sidewalks in osm when they are actually connected are much more problematic at the moment than the accuracy gains lower than the accuracy of a gps.
with only one way that represents all lanes (pedestrian bike car), you know exactly where the person is. you can add the width of lanes if you want, the accuracy in the end is higher than any enduser gps.
and much more important you know precisely the difference between : - this street does not have a pedestrian crossing with zebra or traffic light within X meters, despite a large penalty in the routing algorithm, it will be necessary to cross outside a pedestrian crossing you know that it is physically possible to cross since the schema says that the sidewalks are adjacent to the road or - the street and sidewalk do not touch each other, so in addition to crossing the road, there may be grass or hedges or an not-mapped barrier between the street and the sidewalk, so the penalty must be even greater to penalize uncertainty.
there are 2 proposals to divide the ways into lanes while being able to indicate if lanes are touching or not, but for the moment nobody is working on them and I don't know any routing software that uses them.
so please, do not divide a way into way by lanes when these lanes touch, even if it is the sidewalk.
if you want to increase the usability of osm for disabilities, the most useful one at the moment imho : - provide information on pedestrian crossings: 1/2 has no information on whether or not a traffic light is present 90% has no info on the tactile paving 99% have no information on kerb nor wheelchair accessibility - add entrances of public accessible buildings : step_count or highway=steps at the entrance connect them to the public network
Regars, Marc