Check: Railway end node near other railway

[newsScott@…]

There allready exists a "Way end node near other highway" check. The same would be very helpful to improve the quality of railway data.

Some false positives may be avoidable if the code would also check for the tag "railway:buffer_stop" in the end node.

[anonymous]

i realized that there is allready some support of this. But i think the current implementation needs some changes:

in org/openstreetmap/josm/data/validation/tests/UnconnectedWays.java:

The default distance (mindist) of 10 meters is not useful for checking railways. I my experiments, something like 1-2 meters is best. So should there be a mindist_highway and mindist_railway?

The code finds some false positives: many platforms are still tagged "railway=platform" instead of "public_transport=platform". In member function "visit", ways tagged with railway=platform should not be taken into account:

something like

if (w.hasKey("highway")

w.hasKey("railway") && !w.get("railway").equals("platform")) {

set = endnodes_highway;

}

[skyper]

The test was reworked. Does it work better, now?

[skyper]

Anything left, besides #21038.

[GerdP]

I think #18272 fixes this. Please open a new ticket if further issues are found

[Don-vip]

Marked as duplicate of #18272.

[skyper]

Closed as duplicate of #21038.
See r18272.