[Tagging] Proposal for standardization of sidewalk schema (+ import)

Frederik Ramm frederik at remote.org
Tue Aug 2 21:39:12 UTC 2016


On 08/01/2016 11:35 PM, Meg Drouhard wrote:
> Hello,
> Our team is proposing a standardization of sidewalk tagging conventions
> in OSM to simplify pedestrian network annotations and better represent
> the physical reality of sidewalk ways.  This proposal is particularly
> concerned with features of sidewalks that may aid or impede travel for
> people with limited mobility.
> Our schema proposal is available
> here:

As written on the imports list, I think that separate mapping of
sidewalks will not, and should not, be the norm; they should be possible
where it helps modelling a complex layout but you shouldn't rely on them.

In the past, we've had a number of projects aimed at helping older
people or those with impaired vision or movement to navigate cities.
These were often driven by some university or city department with
limited funding and a limited run time, resulting in often shabby
imports of footway-sidewalks ("the community will fix it, surely"), so
that the university/department could quickly show off their prototype
and then forget about everything and go on with their lives.

This is not what brings OSM forward; these people have not kick-started
anything but just hijacked OSM for a short-term goal.

Mapping sidewalks as individual geometries is the most
data-consumer-friendly of all possible approaches; configure your
renderer to use only footways and bam, it works.

However, the result will be as system that works *only* where separate
footways are tagged, and it is highly unlikely that this will happen on
a global scale. A good pedestrian routing schema *must* cope with
not-explicitly-mapped sidewalks.

But mapping sidewalks as individual geometries puts considerable burden
on the mappers who want to work with the data in an editor.

There's also the question of how to link a separately mapped sidewalk to
the street it belongs to, so that proper routing instructions can be
generated (instead of "follow unnamed footway for 500 metres"). And the
question of how rendering should treat these sidewalks.

Reading material for past research and discussions (some of this is
German only or Germany-heavy I'm afraid) e.g.

* https://lists.openstreetmap.org/pipermail/talk/2015-April/072613.html
(Roland Olbricht on routing suggestions)

* Nathanel Lang's bachelor thesis in which he synthesizes a routing
graph with separate edges for left/right sidewalk even when none exist
in OSM https://github.com/Nathanael-L/pedro

* Long discussion on Austrian list

* Micha Meier's work in Graz, Austria - Project "Access2Life"
http://wiki.openstreetmap.org/wiki/User:Species/Access2Life with
detailed tagging schema and even specialized JOSM presets and styles

I'm Cc'ing Roland and Micha explicitly because I don't know if they are
on this list.

There's also people discussing whether to start mapping certain highways
as areas and it is unclear what, if anything, this means for sidewalk

I'm not a friend of separately mapped sidewalks but I can live with
people mapping them if they desire. Importing them, however, is a whole
different ballgame that should only be attempted after careful
evaluation of past problems and previous work.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"

More information about the Tagging mailing list