[Talk-transit] Ideas for a simplified public transportation scheme

Stephen Sprunk stephen at sprunk.org
Tue Apr 30 17:34:49 UTC 2019

On 2019-04-30 05:50, Dave F via Talk-transit wrote:
> On 29/04/2019 19:39, Markus wrote:
>> On Mon, 29 Apr 2019 at 17:18, Stephen Sprunk <stephen at sprunk.org> 
>> wrote:
>>> Part of what seems to have started the PTv2 mess is that bus stops 
>>> were
>>> sometimes mapped on the way and sometimes beside the way, and both 
>>> cases
>>> were tagged the same.  PTv2 tried to separate those into "platform" 
>>> and
>>> "stop_position", to bring uniformity across modes.
>> It would have been a lot easier to just recommend placing stops beside
>> the road. :)
> If there is a problem on the OSM database I believe sorting that
> problem is beneficial rather than 'papering over the cracks' by adding
> extra tags. It may seem quite laborious, but just as quick as adding
> those tags.

I agree.

>>> We need platforms beside the way so routers can get people to/from 
>>> the
>>> stop on foot.  This is a big deal because trains are long and can
>>> usually be boarded along their entire length, unlike buses where a 
>>> node
>>> often suffices.
>>> OTOH, we need stop positions so routers can get people from stop to 
>>> stop
>>> on the buses/trains.
>> Routers just need the platforms (the places beside the road) because
>> the journey begins and ends there.
> Please clarify what you mean by 'platforms'? Many UK bus stops are
> merely signs clamped to telegraph poles. In rural areas there may not
> even be a pavement, let alone a raise platform. Please remember that
> we should be mapping the physical world. PT schema should fit in with
> what's actually there.

A platform is where people wait to board; if they stand at a pole 
(typical for buses), then the pole is logically the platform.  That's 
easily distinguished from large platforms because it's a node rather 
than a way/area.

>> Stop positions (on the road) are
>> irrelevant for routing. If someone, for whatever reasons, needs the
>> stop positions, they can be calculated (projection of the stop node or
>> centroid of the platform to the highway or railway way).
> Wouldn't a stop position be easier to locate if it's a node on the
> highway, rather than an imaginary, offset 'platform'?
> Please show me a router which uses platforms as I'm struggling to see
> the benefits atm.

I think the idea was that nobody _could_ build routers with the data we 
had, which was inconsistently tagged between areas and sometimes even 
between mappers in the same area.

If you're trying to construct a route that involves walking to a bus 
stop, riding the bus to another stop, and then walking some more, then 
you need a linkage connecting the bus route (using stop positions) with 
the walkways (using platforms).  I'm not saying that's the only way to 
do it, but it's the only way that was proposed.


Stephen Sprunk      "Those people who think they know everything
CCIE #3723         are a great annoyance to those of us who do."
K5SSS                                             --Isaac Asimov

More information about the Talk-transit mailing list