Not sure how that node got there.

I may have somehow mislabelled it inadvertently, but there seems to have a number of odd railway changes worldwide picked up by http://osm-subway.maps.me earlier this week though, in some cases despite no changes recorded in the history of the affected object,

I did correct the name for Poplar a couple of days ago *node/7147562182/history), but looking again now you may have mislabelled this as Langdon Park a month ago,

With the changeset you mentioned, that was a temporary change and the bus/train route relations involved have changed quite a bit since.

Hi Jon, I had a look, and in changeset/81327871 I've 'fixed' the 3 crossings that I could detect in this area.

Regards,
Mac

Hi. As you mention, the problem with the wiki is that it only considers the situation where the junction node has the traffic_signal and doesn't deal with situations where the traffic_signals are mapped separately on each of the junction roads.

In the past, after reading the wiki, I also used to do the way you seem to be doing it, so I can understand the confusion. It seemed illogical that cyclists were stopped for a traffic_signals that was behind them, yet that was how asl's were being generally mapped.

However I later realised what people were were doing is mapping a traffic_signals node where the stop line for motorists is (the point where the traffic_signals control takes effect), and that this node was not being used for where the traffic_signals are physically located.

Although cyclists have a waiting area in advance of this, it is the cyclists' stop line (i.e. the front line of the box) that is the cycleway=asl.

Basically, cyclists would be subject to the traffic_signal node 'behind' the asl but can pass in advance of this traffic_signal node to stop at the asl.

It's very possible that a definition of an 'Advanced Stop Line' could be the motorists' stop line rather than the cyclists' stop line, but I can't see why a "cycleway=*" tag in OSM would be used for the purpose of mapping the rear stop line (instead of a "motor_vehicle=asl").

I don't think the asl tagging is satisfactory, but I'm also not sure how much it's being used in routing etc.

Thanks Toni. I found an example of one of these 'segment' routes at relation/2125451 just to see a specific example of the tagging used. The tagging seems to be in use mostly in Germany for the members of a few route=road parent relations (https://taginfo.openstreetmap.org/keys/segment#map).

The equivalent for the Folkestone - Calais segment would be to make a collection of the railway tracks, but I'd expect that any route via this segment (along a collection of railway tracks) would still need to be a route=train.

For now, I have applied "segment=yes" to all the nested route relations. Marco, hopefully that tag could in some way be made use of by you.

A couple of potential issues I though of with the use of nested relations:

(i) Would all bus stops be better placed at the top of the parent relation, i.e. before the nested route segments?

(ii) Maybe there might be a need for separate train relations for each bus route It might transpire that the train segment will need to have the same ref/network tag as the parent relation.

An alternative simpler (dream!) solution would be if any route tagged with e.g. "shuttle=bicycle;vehicle" or "cargo=bicycle;vehicle" would be 'valid' in PTv2 for use by bicycles + vehicles, so all ways could be used in a single route relation.

Regards,
Mac

Hi Toni,

I've been in contact with Marco at changeset/81031404. I understand he's previously been trying to find a solution to these routes showing as invalid on OSM Inspector.

I'm still working on some experimental testing of how my changes pass with the OSM Inspector validator.

The initial changes I made did make these invalid routes 'valid', so I'm now testing if reverting one route to the preferable "route=bus" keeps it valid. If it does, I should at least be able to solve/revert the route=bus;train" tagging to "route=bus" later tonight.

Regards,
Mac

Hi Marco,

I agree it would be nice to be able to have a continuous route. I'd previously hoped that tagging the Eurotunnel train route relation with 'service=car_shuttle + cargo:bus=yes' might solve the problem but it didn't. I also tried adding the Eurotunnel route relation (instead of the railway tracks segments) to the bus route without any success.

With regard to an IT solution to the use of nested relations, I wonder if https://hiking.waymarkedtrails.org/help/rendering/hierarchies *might* be a pointer to finding a solution (site is run by osm.wiki/User:Lonvia).

In the background, she has somehow managed to combine nested relations for individual walking/cycle routes to display them on a map as a single route, e.g. London's Capital Ring walking route (https://hiking.waymarkedtrails.org/#route?id=10952) is made up of 15 subrelations. Another example from the cycle routes is https://cycling.waymarkedtrails.org/#route?id=1207220&map=5!56.2246!14.5649.

I should mention, I did replicate the changes from this changeset in the other coach routes using the Eurotunnel train in advance of seeing your initial comment here.

On last night's OSM Inspector update, all these routes did show as valid, but I'm waiting to test out a couple of further things I've done like reverting an individual route to the preferable 'route=bus' to check if it remains valid when OSM Inspector updates again tonight.

On a different route I'm testing out whether the 'virtual ramp' to the train track can be included in a bus route.

Regards,
Mac

Hi Marco,
Unfortunately it isn't valid for a route=bus to run along a railway=rail.

Although the buses themselves do go onto the train, passengers then generally get off the bus and this section of the public_transport journey is a route=train.

PTv2 bus routes that run along train tracks are displayed as invalid on OSM Inspector (http://tools.geofabrik.de/osmi/?view=pubtrans_routes&lon=1.34715&lat=51.04895&zoom=10&overlays=ptv2_routes_invalid). This also makes it difficult to identify any additional errors that might occur along this long route.

Regards,
Mac

Thanks for the update.

Southwark also had a tmo for contraflow cycling on way/506015142 ~5 months ago... that hasn't yet been enacted. So this C35 work might not be done for a while yet.

http://public.londonworks.gov.uk/roadworks shows that there is 'Q9' work planned in April around way/7998498, which might well be the start of the planned work here for the C35 route.

I've now added these tags.

I'd be a little surprised if work on the proposed cycle tracks has already been done, but I agree the Commercial Way junction might well be worth the survey.

If you get the chance, there is a planned pathway realignment/straightening (way/567718774) that could be worth keeping an eye for, although I don't expect this has been worked on yet.

The railway=station is an interchange station with 6 platforms but, as far as I understand it, platforms 3 + 4 are not for access to public transport.

Platforms 1 + 2 (suburban service) are part of a stop_area that has e.g. a different "naptan:AtcoCode" to the stop_area for platforms 5 + 6 (national services). Previously these stop_areas were mapped as 2 separate 'Wembley Central' railway=station, but there is a common entrance to the interchange station here.

If the 2 'public_transport=station' stop_areas were meticulously mapped as areas, then neither area would include platforms 5 + 6... but these platforms would still be part of an accurately mapped railway=station area.

Thanks. I've now removed this station=subway tag. The station node is also now a member of the 2 stop_areas.

Regards, Mac

Hi,
The point I was trying to make is that the London Overground should be in the same WMBYDC stop_area as the London Overground - they both use platforms 1 + 2.

https://en.wikipedia.org/wiki/Wembley_Central_station actually states "Wembley Central is an interchange station" rather than 2 distinct stations.

Also National Rail uses platforms 5 + 6 of this interchange station, rather than e.g. platforms 1 + 2 of a separate station.

For this reason I've removed the duplicate railway=station tag in changeset/80385615 which has attracted attention at http://osm-subway.maps.me/uk.html

You might cast an eye over the tagging of the railway=station 'interchange node' at node/7175566127 in case it's not fully/properly tagged.

As an aside, I don't see any subway entrance mapped here.

Regards,
Mac

The node on the tracks is a stop_position, and is not part of the Overground as the network tagging of this track node had implied. The previous 2 stop_areas had already been consolidated into one.

If separate stop_areas need reinstating with 2 public_transport=station tags, would it not be best to have 1 railway=station tag for the Wembley Central interchange?

Hi,
The turn restriction (restriction=only_u_turn) that you added here looks very wrong.

If you explain what you were trying to achieve with the turn restriction, I can correct it to what it should be.

Regards,
Mac

Hi,

Red is the official colour on NCN route and the background colour for route numbers on signs.

Use of colour for routes is mentioned in osm.wiki/Relation:route#Tags.

This colour tag is used by e.g. https://cycling.waymarkedtrails.org/#?map=13!51.4031!-0.2934 to set the colour used to underline route numbers.

In a similar vein, the osmc:symbol tag is used with hiking route, e.g. https://hiking.waymarkedtrails.org/#route?id=10058046

Regarding new NCN 1 relations: there are alternative NCN 1 route sections that hadn't been mapped as part of the NCN, e.g. sections that would have to be used to follow NCN 1 when way/605748677 is closed at night.

Also, as a relation:route should be one continuous line of travel, I've mapped branching sections of NCN 1 in individual relations where necessary. In practice, this make identifying gaps or similar errors in any type of relation:route easier.

Regards,

Mac

Hi,
I'd assume the cycleway construction here should be completed by now as someone tagged it in July 2018.

Sustrans had this mapped as a proposed section in the past. I can see they now have it mapped at https://osmaps.ordnancesurvey.co.uk/ncn/51.61174,-1.24531,18 as being part of NCN 5 though.

I've therefore added this section to the NCN5 relation. Rather than presuming the highway=construction tag is definitely outdated for the cycleways, I've just added a FIXME tag instead.

Mac

David, http://static.walthamforest.gov.uk/sp/Documents/chapter4-calibri.pdf#page=17 does show the Olympic route (in purple) finishing at the entrance to Epping Forest on Snaresbrook Road.

relation/8751401 is definitely signed as a LCN from way/380369498, but it does actually appear to be a ONEWAY link route, i.e. basically a shortcut for westbound cyclists on New Cross Road to head southwards to join onto the Crystal Palace route.

I've now added role=forward to its ways in order to tag it as a oneway route. Also I've added the relation for this short route as a subrelation (with 'role=link') of the route it links into.

Errors with routes tagged with "public_transport:version=2" get highlighted at
http://tools.geofabrik.de/osmi/?view=pubtrans_routes&lon=0.57186&lat=51.71202&zoom=12&overlays=ptv2_routes_invalid,ptv2_error_ways. 'Errors' would include gaps and misordering of route members, or use of role=forward/backward/etc.

It gets updated daily at about 11pm UK time.

I only rectified PTv2 bus routes, so any other bus routes here might need looking at.