[OSM-talk] Bus Stops
peter.miller at itoworld.com
Sat Aug 11 07:21:04 BST 2007
The conventional way of handling Bus Stops in the public transport industry
is to have a node for each individual point at which one can get on a
vehicle, so if there are two bus stops on opposite sides of the road then
they are represented as two nodes. If there are three bays in a row on one
side of the road then they are represented a 3 nodes in a row. Every Bus
Stop in the UK has a unique code, and this is sometimes printed on the bus
In the EU standards they are called 'Stop Points' (rather than Bus Stops) so
they can cover buses, tram, rail, ferry planes etc.
In railway stations there is a Stop Point for each Platform (and each bay in
a bus station, each Gate for an Airport and each quay in a Ferry terminal).
Groups of local Stop Points (as they are called) are then arranged into Stop
Areas where they are very close to each other.
These Stop Points are not within the road layer because Stop Points are a
distinct dataset managed separately; they are then associated with a street,
sometimes using the Street Name and sometimes based on proximity.
I recommend that we use 'Bus Stop' and 'Stop Point' for this low-level
purpose and construct entities as we need them.
The database of all these points in the UK is called 'NaPTAN' (standing for
'National Public Transport Access Nodes'), there are about 350,000 of them,
and keen people can find additional information here:
A new CEN standard is in the process of being ratified, called IFOPT which
can be used for describe much more complex transport interchanges, such as
major airports and railways stations, detailing every corridor, lift,
check-in desk escalator etc. CEN standards are used throughout the EU and
There is also a modelling standard for public transport in general published
by CEN called transmodel which covers the modelling in general and is used
behind most professional transport products used in Europe.
Of course, I am not proposing that we 'implement' all of the above, but
where we choose modelling approaches and terms for entities it would be
sensible to choose the same names.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the talk