Montreal REM: Structuring and naming of routes

[tfitzg]

Hi. I’m looking for expertise with this approach to mapping the Réseau express métropolitain in Montreal:

@tfitzg

There was a discussion on the Agora Montreal forum about how to tag the A1, A2, A3 and A4 directional indicators used on the REM. The consensus there was that they should be four one-way routes using the sigle and name (e.g. “A1 - Brossard”), terminating at the station in question, i.e.:

• a three-route master leading from the other ends to Brossard for A1
• an A2 from Brossard to YUL
• an A3 route from Brossard to Anse-à-l’Orme, and
• an A4 route from Brossard to Deux-Montagnes

I mention it here (a) because I’m new to this forum and don’t want my first action to be to start a new thread, (b) to raise awareness and foster discussion if this is not the right approach, and (c) because I think there’s already other changes being made over these.

Since posting this, the structure of the relations have been set up to the following:

• Relation: Réseau express métropolitain (6169927): type=route_master
  • Relation: REM A4 - Deux-Montagnes<>Brossard (19792160): type=route_master
    • Relation REM A4 : Brossard → Deux-Montagnes (19669299): type=route
    • Relation REM A4 : Deux-Montagnes → Brossard (19668643): type=route
  • Relation: REM A2 -YUL-Aéroport-Montréal-Trudeau (19672326): type=route
    • Relation: REM A3: Anse-à-l’Orme <> Brossard (19792138): type=route_master
    • Relation: REM A3: Anse-à-l’Orme → Brossard (19668926): type=route
  • Relation: REM A3: Brossard → Anse-à-l’Orme (19672327): type=route
  • Relation: REM A2: YUL-Aéroport-Montréal-Trudeau → Brossard (19668927): type=route
• Relation: REM A1 - Brossard (16141765): type=route_master
  • Relation: REM A1 : Montréal → Brossard (16139923): type=route
  • Relation: REM A1 : Brossard → Montréal (6170078): type=route

This does not represent reality, because there will never be an A2, A3 or A4 train to Brossard, and the A1 line shown only represents a small initial segment separate from the network. There are only REM trains : all trains heading to Brossard (regardless of what branch they start from) are A1, A4 trains will always go toward Deux-Montagnes, A2 trains toward A2 Airport and A3 trains toward Anse-à-l’Orme.

I propose we reorganize and rename these relations to reflect reality:

• Relation: Réseau express métropolitain (6169927): type=route_master
  • Relation: REM A1 - Brossard (16141765): type=route_master
    • Relation: REM A2: A1 - YUL-Aéroport-Montréal-Trudeau → Brossard (19668927): type=route
    • Relation: REM A3 A1: Anse-à-l’Orme → Brossard (19668926): type=route
    • Relation REM A4 A1 : Deux-Montagnes → Brossard (19668643): type=route
  • Relation: REM A2 - YUL-Aéroport-Montréal-Trudeau (19672326): type=route
  • Relation: REM A3 - : Brossard → Anse-à-l’Orme (19672327): type=route
  • Relation REM A4 - : Brossard → Deux-Montagnes (19669299): type=route

In this setup, I believe that these relations become redundant:

• Relation: REM A1 : Montréal → Brossard (16139923): type=route*
• Relation: REM A1 : Brossard → Montréal (6170078): type=route*
• Relation: REM A4 - Deux-Montagnes<>Brossard (19792160): type=route_master
• Relation: REM A3: Anse-à-l’Orme <> Brossard (19792138): type=route_master

*These two relations represented the introductory service of the REM from July 2023 to November 2025, but operationally the direction toward Montréal (Gare Centrale) was never labeled A1.

Does this relationship of routes and master routes make sense? Does it break something that I’m not aware of? I welcome all feedback.

Tagging @Claudius_Henrichs , @bouzinac , @SomeRailfan as you all experienced mappers who have been involved in this part of the map in recent months.

[tfitzg]

To illustrate the situation, here’s an image from the “Upcoming depatures” board in a REM station:

(Image source)

Only A1 and A4 appear as A2 and A3 are not yet in service (something else we should reflect in these relations’ tags)

[bouzinac]

Bonjour, les missions vers les quatre antennes du REM sont elles toutes omnibus (s’arrêtent à toutes les gares)?

[tfitzg]

Exact. Aucun trajet “express” comme font les lignes du Subway de New York. Tout train arrête à chaque station entre Brossard et Bois-Franc, et ensuite chaque station sur l’antenne qu’il emprunte.

[Claudius_Henrichs]

Thanks for taking the initiative to untangle this REM confusion and attempting to map it to the OSM data model. I find it utterly confusing that they decided to deviate from the international standard of naming one line service in both directions with the same line reference number. This deviation makes it tricky to map it to the OSM public transport model.

In this established OSM public transport model a route_master relation links all variants of a line service. E.g. in a normal system "northbound and southbound subway routes of line 1”. Most of the time these are two routes, but some times there are line variants for shorter services or branches.

In the REM scenario it seems there are these three “line services” between terminus stations:

1. Brossard ↔ YUL-Aéroport-Montréal-Trudeau
2. Brossard ↔ Anse-à-l’Orme
3. Brossard ↔ Deux-Montagnes

Each of these I would suggest to represent in OSM with a route_master linking two route relations for each direction. (2/3 of this is what’s already mapped in OSM right now). The only problem with existing OSM metro validators now would be that each direction has a different reference number, but that’s a warning we have to ignore.

Let me know if this makes sense or you have more questions. I’m happy to help update OSM data accordingly if we agree on an approach.

[SomeRailfan (Nicolas Landucci)]

Technically, the whole REM system constitutes line A, with the termini designated A1 (to the south), A2, A3 and A4 (to the north). In practice, the operator (almost) treats A1-, A2-, A3- and A4-bound services as wholly different lines in some public communications.

[Claudius_Henrichs]

Does that mean you support my proposal to treat the A2, A3, and A4 bound services as three separate OSM routes then?

[SomeRailfan (Nicolas Landucci)]

Indeed, I think it would be best to tag A1-bound services together, and A2-, A3- and A4-bound services separately.

[tfitzg]

So is this the proposal?

• Relation: REM A1 - Brossard (16141765): type=route_master

  • Relation: REM A2: A1 - YUL-Aéroport-Montréal-Trudeau → Brossard (19668927): type=route

  • Relation: REM A3 A1: Anse-à-l’Orme → Brossard (19668926): type=route

  • Relation REM A4 A1 : Deux-Montagnes → Brossard (19668643): type=route

• Relation: REM A2 - YUL-Aéroport-Montréal-Trudeau (19672326): type=route

• Relation: REM A3 - : Brossard → Anse-à-l’Orme (19672327): type=route

• Relation REM A4 - : Brossard → Deux-Montagnes (19669299): type=route

Relation: Réseau express métropolitain (6169927): type=route_master

[tfitzg]

I’ve submitted changes in this changeset:

Changeset: 176451224

I struggled with the removing the old A1 Brossard-Montreal relations, so I tagged them with a “Fix Me” tag.

[Claudius_Henrichs]

Thank you. I’ve further updated the REM route relations so they reflect their active or under construction state.

Changeset: 176745526

[tfitzg]

I have a proposal. Because it seems the GTFS data for the REM has the same challenge adapting these A1, A2, A3, A4 directions as we do in OSM.

They’ve given a route representing the equivalent of relation 16141765 a route_short_name of A4-A1. (There’s a second route currently, but it’s temporary while they bring the branch leading to A3 online.)

Would you be okay if I update relation 16141765 with a ref=A4-A1?

And I’ll remove the wikidata tag as it’s not a 1:1 match anymore, but I can add a network:wikidata=Q19582331