Validator should flag ferry routes which are not connected to a highway=* or man_made=pier

[GerdP]

Sometimes ways with ​osmwiki:Tag:route=ferry are not properly connected to a highway. This can cause routing problems.
I know these typical problem cases:
1) the route is split into parts, so that an end node is only connected with other route=ferry ways
2) the connected highway is mapped with a low class, e.g. as hw=service instead of e.g. hw=secondary

I've attached a sample file created from current OSM data in Canada which contains both problems.
Nodes 660826272 and 323308217 are samples for 1), Node 1682319513 is a sample for 2)
If this check requires Java code please let me know and I'll try to code it.

I am not sure how to handle ferry routes which allow vehicles (e.g. bicycle or mofa) if they end at
a man_made=pier (and no other highway).

[naoliv]

Is man_made=pier routable or it needs explicity access=*?
And here I see cases like the examples in ferry-example.osm​:

• a ferry not connected to a highway or
• a ferry that ends in a terminal (and this terminal is not connected to a highway)

[GerdP]

My understanding of the wiki is that the access=* tag is about mooring boats.

[Klumbumbus]

Regarding 1) Routing should work even with splitted ways. It works for highways too. However it is written in the wiki that "The ferry route must not branch in the water", so I think we can add the test.

Regarding 2) I don't see a problem if the last way to the ferry terminal is mapped as highway=service. That shouldn't raise any routing problems.

[GerdP]

Reg. 1) In my example data the way is not only split. Multiple ferry lines end at the same node in the sea. Routing is not as on highways since it is not possible to change the ferry at such a node, and also this point in the sea is probably not really
fixed. So I think the wiki is right.

Reg. 2) OK, this might be a special problem that exists in Garmin GPS units, so I agree that this test is not useful in JOSM.