[talk-ch] Public Transport tagging (was: Didok check map)

Sarah Hoffmann lonvia at denofr.de
Sat Jun 11 12:14:35 CEST 2011


On Thu, Jun 09, 2011 at 10:58:06PM +0200, Beni Buess wrote:
> schrieb Sarah Hoffmann <lonvia at denofr.de>:
> 
> > > there was no discussion on whether we should add the new tagging
> > > schema (1) to the stops, if we touch them. If we do, i think we
> > > should *at least* place the bus stops ON the way with
> > > public_transport=stop_position. if the stop ist not exactly on the
> > > same place for both route directions, 2 nodes should be placed. and
> > > then, there is more about the public transport feature, but read it
> > > yourself.
> > 
> > My personal opinion in this: for simple bus stops, we should stick
> > with the old schema. That is, put a highway=bus_stop node where the
> > pole is and that's it. (For the Didok data, it means indeed to
> > duplicate the imported nodes most of the time because there are two
> > poles per stop.) For the more complicated stops or for micromapping
> > train stations, use the new schema. Just make sure that one node
> > remains that also has the traditional tags, so renderers don't get
> > confused.
> 
> In my opinion it's not a good idea to set the goal on mixing up the two
> schemas. In one way they are complementing each other, but there is
> basically one point where they are different: the place for the bus
> stops. 

I never quite understood the need for a stop_position. The only person
who would need this information is the bus/tram/train driver. And if
he/she needs OSM to find the stop position then this country is in
serious trouble. Routing should always start and end at the platform.

There is another reason to prefer the platform node over stop_position:
if rendered on a map a human can immediately determine which direction
the stop serves when the platform node is rendered. That is valuable
information you get for free. Why throw it away?

> We have to decide on whether we want to map the stop_position or the
> platform if we want to keep the bus_stops as simple as possible.
> 
> I propose to tag the stop_position of the new schema, as this is IMHO
> the node that should be in the route relation at least. 

Why? Why not put the platform in it? If you need a point on the way
you can always determine the nearest point on the way. PostGIS has 
functions for that.

> That would mean
> the tagging for basic bus stops would look like this:
> 
> highway=bus_stop //optional, you may leave it there to support the
> current renderers
> public_transport=stop_position
> bus=yes
> uic_name= //if there is no stop_area relation
> uic_ref= //if there is no stop_area relation

I simply don't see what is the point in replacing one simple
and understandable tag (highway=bus_stop) with two rather
mystic tags (public_transport=stop_position/platform, bus=yes)
That is exactly why I think that the new proposal tries to 
make life of mappers as miserable as possible.

Especially as your little statistic here proves that it is
very, very simple to determine, whether that highway=bus_stop
is meant to be a stop_position or platform.

But that is all just my 2cents. Let each mapper decide. In any
case, you need to support the two schemata because even if we
decided right now to follow one or the other, it would take
a while to consolidate.

Sarah



More information about the talk-ch mailing list