A year ago someone split up the tokyo station node into three.
One for “local” trains and two for highspeed trains. I guess because they have different operators. Is this the right approach? I think there should be only one station node with all operators merged.
see this changeset: Changeset Map
In fact, it’s 4 objects. There’s one for Metro.
It could be justified as they are separately gated paid areas. They are similar to being different stations. A 2nd level would be structurally different station boxes constructed at various times linked up together, as in Otemachi, Nihonbashi, etc stations, which also have funky combined shapes. The 3rd level is entirely separated, as in Akihabara Station for JR, Tsukuba Express, and Metro.
Check London. There are various distinguishing, although many can argued to have a different name= etc. This is not to say it must be correct, but it can be understood.
Some attributes can be different. This can’t be matched easily. How can we know which ref= belongs to which with semicolons easily and reliably? Viz ref=TYO only applies to JR East. (Marunouchi LIne’s ref=M17 is better limited to =stop , being numbered for stops on each line, not stations as a whole)
The 3rd level is entirely separated, as in Akihabara Station for JR, Tsukuba Express, and Metro.
Check London. There are various distinguishing, although many can argued to have a differentname=etc. This is not to say it must be correct, but it can be understood.
Some attributes can be different. This can’t be matched easily. How can we know whichref=belongs to which with semicolons easily and reliably?
several refs indicate several objects, having them comma separated on one object, just as with names, hints at a mapping problem.
In Berlin there are also several stations at the same place:
https://www.openstreetmap.org/node/3856100103
https://www.openstreetmap.org/node/2459919675