There is only 1 set of ticket validators to access the Eurostar.

Can you please make sure that the platform multipolygon relations aren't broken when you are fixing the platforms like this relation/8108183
?

Sorry, they represented two steps, one over the other.

I have fixed it and added level tags now.

Sorry what did you do to the steps? The map is wrong now and I have to fix it.

Why do you keep removing the platform_edge all around the country? I surveyed these platforms and these platform edges are needed to correctly map the extent the whole length of the train, as part of the platform is fenced and trains do not stop.

I have seen that you have added some naptan:AtcoCode tags into the stop position on the tramway. This is wrong, and it should be placed on the platform instead.

In addition, the code themselves are wrong as well. The TfL timetable data shows that the platforms for the two directions have their own codes with 1 and 2 as their suffix.

I fixed the problem yesterday.

Please make sure that the platforms at different layers aren't joined together.

You have to add access=no for emergency paths, otherwise the router will attempt to use it to get to the platform.

I also did the same to some other stations, such as Tottenham Court Road, a while ago.

Those I left separate are those where there is a clear boundary on the ground between areas of responsibility, where the parts can be operated separately, like Liverpool Street.

It was intended. West Ham (the rail station) is now clearly one station (the entrance and concourse are shared so no part of the station can work independently without using that entrance and concourse) so there should be only one stop area for that station, and the name is just "West Ham".

The meaning of the ref= is the official route code. The codes I have put in are used officially in the TfL Go application and website to represent line status.

This is also the same in Hong Kong MTR network as well, where the 3-letter code is put in the ref.

Each of the 11 Underground lines is represented by one route master, therefore for consistency each of the 6 Overground lines need to be represented by one route master as well.

This changeset has two problems:
1. you have blocked off platform 3
2. access=no is not right here, since this bridge is accessible when trains do call at platforms 1&2. This has resulted in my journey planner not able to produce the correct route when a train calls at platform 2.

The bare ref= should contain the primary reference value as used by the operator itself. The CRS code should only be listed in ref= if it is actually a National Rail station, as London Underground or any other networks have their reference codes.

These CRS codes are all valid and in use, which should not be removed.

I have no idea, it was just the result of merging duplicate objects.

https://tfl.gov.uk/cdn/static/cms/documents/rfliccostpr2026v1.pdf

The timetable planning rules clearly lists the operational CRS codes for all the Elizabeth line CCOS stations - CRS codes exist for stations and also other places like depots, junctions, as seen in Network Rail data (which is open data), so the use of ref:crs is appropriate.

ORR derives the stations usage statistics from ticketing data, which do not separate PAD/PDX, STP/SPL, ZFD/FDX, LST/LSX, etc., because they are considered the same station in terms of rail ticketing, but operationally they are clearly two different stations (apart from Farringdon which is disputable) which can be easily verified as ground truth.

LSX is the code used in the timetable data for the Elizabeth line station, this code can be found in the Network Rail data under the column crs_reference_code but it is not publicised.

Similarly for PDX (Paddington Crossrail), FDX (Farringdon Crossrail), SPL (St Pancras Low Level), etc. The existence of separate codes, along with separate entrances / concourses / gatelines or even operators for ground truth both confirm that they are two physical stations for operational purposes.