Public transport journey planning has a public purpose - get people efficiently to their destination.
This planning function *might* not have financial compensation - there might be no ticket sale involved.
The likes of Trainline then have very little incentive as I see it to provide excellent journey planning. But if they can’t, and the likes of DB won’t, internationally… then who can?
@jon One angle to this problem is: what makes it work for road navigation & flights, and why is that not applicable to rail? Listing some thoughts.
Economically (for the routing services):
- Road: routing apps are provided essentially for free, and funded indirectly. For example, having a mapping application with directions is table stakes for a smartphone manufacturer, in order to sell their hardware. But the B2C demand of long-distance rail is lower than road so that's not table stakes.
- Flights: B2C price comparators don't operate on a huge margin (I guess?), but take a commission on sales and there's a bunch of competing comparators (pushing quality up?). Also there's money to make on B2B integrations with large companies.
As long as the B2C & B2B demand is much higher on road/flights than rail, the supply of good rail planning (from private companies) won't follow?
@jon Economically (for the operators):
- Trains: European state railway incumbents have incentives to slow down the opening of their data, to make their new competitors undiscoverable. Perhaps something a competition authority might address (if politics don't block it like MDMS)?
- Road: the few toll roads are monopolies, so no need to hide themselves. And the price isn't shown upfront when asking for directions, which is even better for the operators.
- Flights: competition is widespread already on most routes, so being the one hiding your data makes yourself less discoverable, benefiting your competitors.
@jon Technically (thinking routing algorithms):
- Road: the network is very complex with many streets & roads. But importantly, (1) there is no schedule element, i.e. roads are available at any time (except rare sections closed at night) and (2) incorrect data (e.g. unplanned road closure) is not catastrophic, an unplanned detour in a city only costs a few minutes. Planning rail with many connections requires excellent data quality.
- Flights: similar to trains, but the hub model (with a huge fanout of directions at each airport) means that one can go from anywhere to anywhere in a few hops only. And the worldwide airline alliances mean that these few-hops trips can be booked in a single ticket. Unlike rail, the routing doesn't become increasingly more complex as distance increases (and more railway operators get involved).
To take a concrete example, Google Maps has a transit option modeled over road navigation (showing duration and route), which works well at a city or regional scale. But it's not much useful on long-distance trains where pricing and availability (+ having a way to book the tickets) are critical.
Jon Worth
Guillaume Endignoux