Google's unofficial route_type 8 to 12#174
Conversation
|
Can you please reference where you retrieved this information (8-12) from? Since it is not on https://developers.google.com/transit/gtfs/reference/#routestxt Are you representing Google in this pull request? |
|
I'm not aware of any public documentation of those route types that I could refer to. They come directly from conversations with Google. @aababilov: Could you maybe share your perspective? |
|
+1 for 11 Trolleybus, and I don't think 3 needs to be changed, because where they use trolleybuses people call them like "trolley" vs "bus" anyway. But 9 is a bit of a problem, because if we'd add it then the definition of 3 probably should drop the "long-distance", but that's a non-backward compatible change. |
|
BTW what about Rickshaw? https://en.wikipedia.org/wiki/Rickshaw |
Can you present a timetabled Rickshaw? |
|
About richshaws, I assume @flocsy may be referring to the "Auto-rickshaw", like the Tuk-tuk in Nairobi (https://en.wikipedia.org/wiki/Auto_rickshaw), which has been described using GTFS by the University of Nairobi as part of "Digital Matatus" (http://www.digitalmatatus.com/intro_lite.html, the GTFS is in the "Map" tab). It is one of the project by DigitalTransport4Africa, which also described the informal transit network in Accra (Ghanna). But maybe we should keep the informal transit conversation for another thread, and try to first see which among those Google unofficial route type we should merge with the official spec. |
|
+1 from Google. Trolleybus (11) and monorail (12) are used by private providers in 5 countries. Monorails are also perceived as a special vehicle type, so they need their own route_type. |
|
@mbta has trolleybuses, but our riders don't think of them as separate from regular buses: the routes w/ trolleybuses also have trips served by regular buses. |
|
I understand that Google is a big player, but wouldn't it be more logical to have Trolleybus and monorail as 8 and 9? For google to support it we could talk about having an indicator of the GTFS version or supported feature like discussed in #171 |
|
We at Google don't think that it is important to have sequential numbers for route types. However, if it is the goal, I can suggest more vehicle types to fill the gap:
All those route types are substantially different from the existing standard route types. E.g., it is fine to give commuter rail as 2 (rail) but there is no way to represent a people mover using types 0-7. By the way, is there a conflict with existing large providers or consumers that may use 8-12 in some other senses? If so, the standard GTFS may declare 8-12 as "reserved, do not use". We at Goolge would like to avoid defining versioning of the spec. While this can make lives easier for data providers, data consumers may have to take the burden of supporting official versions of the spec. |
|
+1 for rack-and-pinion railway |
|
I see three options being discussed here:
We have to keep in mind that option 3 may prolongate the conversation, and that we'll also then need producer of those values. So: @gcamp: Would you be against adding 11 and 12 for now and discuss then the potential needs for 8 to 10? |
|
What would be the end goal with the gap? Close it with new values that should match what Google has or fill it with any value that we find useful? How important is the backward compatibility with Google's extension? |
|
Values 8 to 10 aren't used by Google (even if they are defined), so they can be reallocate without having any impact. So the gap would likely be filled with other needs with time (Alexej listed a few possible options), even if I guess nobody could ensure that it will be. |
|
Good, then I don't mind really. I would prefer 8 to 10 but won't block otherwise. |
|
Oh ok. Then I can open the vote with the current 11 and 12, and then we'll add the other when we'll need it. We just need a producer. @aababilov Could you open a dataset with 11 and 12? |
@LeoFrachet in our opinion, intercity bus should likely be added, but I think we would need producer input on this since many intercity buses operate in ways that are indistinguishable from intracity. It's not clear what specific characteristics would be used to separate types For example, one of our partners, NJ Transit, has bus routes that go between New York City and various cities in New Jersey [1] [2]. Today, they're represented simply as buses ( [1] http://transitfeeds.com/p/nj-transit/409/latest/route/3 |
|
@mattjoseph why would you want to differentiate between them? A car is a
car whether you just go to next street or next country. Can't the same be
said about buses? Is there a difference between the two types of buses
you'd like to represent with different route_type?
…On Mon, Aug 12, 2019 at 10:24 PM Matt Joseph ***@***.***> wrote:
- route_type=9 intercity buses aren't really described in GTFS, they
are usually represented with route_type=3 bus, which is defined by
"Bus. Used for short- and long-distance bus routes.". => *Should it be
also added?*
@LeoFrachet <https://github.com/LeoFrachet> in our opinion, intracity bus
should likely be added, but I think we would need producer input on this
since many intercity buses operate in ways that are indistinguishable from
intracity. It's not clear what specific characteristics would be used to
separate types 9 and 3.
For example, one of our partners, NJ Transit, has bus routes that go
between New York City and various cities in New Jersey [1] [2]. Today,
they're represented simply as buses (3), and it's not clear which, if
any, would become "intercity buses".
[1] http://transitfeeds.com/p/nj-transit/409/latest/route/3
[2] http://transitfeeds.com/p/nj-transit/409/latest/route/58
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#174?email_source=notifications&email_token=AAHI4RE5S4HOH6LE3QVSNODQEG2IJA5CNFSM4IF33GH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4DSDRQ#issuecomment-520561094>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAHI4REII66GIOKZKSU3RTLQEG2IJANCNFSM4IF33GHQ>
.
--
Gavriel Fleischer
|
|
@flocsy in essence, it would provide for better granularity regarding the function of the vehicle (level of service provided), rather than the physical characteristics of the vehicle. This would seem to be in line with the other For instance, |
|
So, in my first message I was just stating the current internal Google definition and asking "Should it be also added?". I personally think that we should not add intercity bus as a new route type, otherwise we just won't know if a route with @flocsy: We could define a distinction between both. As @devadvance said, this is what is done for rail service. And for bus, they are a big physical distinction between city buses and coaches. Passenger are expected to just stand between the seats in the first one, when it's too crowded, when for coaches the maximum number of person allowed is the number of seats. But for backward compatibility reason, I would just avoid making that distinction. |
|
In Israel it's usually as you wrote (about standing on buses vs sitting in
coaches), but sometimes you can see "coaches" in the city buses (on the
same route). There can be random reasons. (rush hour, broken down bus...)
Also many inter-city buses act as local bus on the way out from the 1st
city and on the way into the 2nd city. I mean: cityA --> cityB bus will
load passengers on numerous stops in cityA, and unload passengers in a few
stops in cityB, wile on the "long leg" it won't stop, or only in the main
stations of a few bigger towns on the way.
Anyway, I'm also against, because (even without considering the backward
compatibility problem) it's unclear both as a producer and as a consumer
what those values mean.
…On Tue, Aug 13, 2019 at 10:19 PM Leo Frachet ***@***.***> wrote:
So, in my first message I was just stating the current internal Google
definition and asking "Should it be also added?".
I personally think that we should not add intercity bus as a new route
type, otherwise we just won't know if a route with route_type=3 is meant
to be an intra-city bus, or if the producer is using the old definition and
still define intercity bus like that.
@flocsy <https://github.com/flocsy>: We could define a distinction
between both. As @devadvance <https://github.com/devadvance> said, this
is what is done for rail service. And for bus, they are a big physical
distinction between city buses and coaches. Passenger are expected to just
stand between the seats in the first one, when it's too crowded, when for
coaches the maximum number of person allowed is the number of seats. But
for backward compatibility reason, I would just avoid making that
distinction.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#174?email_source=notifications&email_token=AAHI4RB3JX7FFXDXSUJPCB3QEMCKRA5CNFSM4IF33GH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4GWH5Q#issuecomment-520971254>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAHI4RG2YH6LWLHMSQ7H6ULQEMCKRANCNFSM4IF33GHQ>
.
--
Gavriel Fleischer
|
|
Here is the open GTFS data kindly provided by Easyway. 11 - Trolleybus - a feed for Lviv (Ukraine) 12 - Monorail - an open subset of the Moscow feed (Russia) |
|
So the full request has been open for more than 7 days, and we have a producer (Easyway) and a consumer (Google). Therefore I'm opening the vote on it! We are Wednesday 14th, so the vote will be open until Wednesday 21st 23:59:59 UTC. @aababilov, @flocsy, @devadvance, @gcamp, @skinkie, @paulswartz you all contributed to the conversation. You can cast your vote now. |
|
Is this necessary, given the existence of the extended route types proposal? Would it be more advantageous to officially adopt extended route types, and not add a small number of redundant codes that only cover certain modes? |
|
+1 for adding Trolleybus and Monorail (with any id the others decide) |
|
+1 from Google for 11 (trolleybus) and 12 (monorail) I find extended route types dreadful. Many of them are unclearly defined (say, what is the difference between metro and underground?!) and many types are simply missing there (no way to tell between tram and light rail etc). I am supporting the GTFS-Modes proposal which gives the full flexibility to data providers and consumers (http://bit.ly/gtfs-modes-and-networks). I would like to keep the list of route_types short and dense. |
|
Re: @aababilov following your comment I have a question: What's the purpose of the route_type field vs the purpose of GTFS-Modes proposal? To me it looks like the latter is more end-user facing (a bit like l11n or translation) Locals are used to see Subway vs Path. But the route_type maybe shouldn't aim at end-users. For example in Moovit we use route_type in the fare calculations in Sydney. |
|
@flocsy what is the reason you do not select the fare information in Sydney by only route_id? The end-user wants to see or select a mode of transport when exploring transit options. I fail to see why we wouldn't want that for the end-user. |
|
@LeoFrachet can you fix the conflict and reopen voting? |
|
Thanks for the update, Tim! Please could you reopen voting? +1 from Google. |
|
Little wrap-up on the PR discussion A consensus was reached on adding 2 new route_type values ( As the PR clarifying the route_types Consumer: Google The vote is open until Tuesday, February 4th at 23:59:59 UTC. @flocsy, @aababilov, @mgilligan, @devadvance, @abyrd: as you voted on this PR back in August, don’t hesitate to vote again if you want to. |
|
+1 from Transit |
|
+1 Moovit |
|
+1 from Google |
|
+1 from Kisio |
|
+1 from Ito World |
|
+1 from IBI Group |
|
+1 from Trillium |
|
The vote has ended and the new route_types Producer: EasyWay 7 votes for: No abstentions, no votes against. |
|
I am so glad to see trolleybus and monorail added. Thanks everyone! |
sccmcca
added
proposal
GTFS Schedule
16 participants
Google uses the following unofficial
route_type:8: Horse Carriage (used in fact as "easter eggs")9: Intercity buses and coaches10: Commuter Rail11: Trolleybus (As defined by https://en.wikipedia.org/wiki/Trolleybus)12: MonorailThey'd like to officialize some of them, namely:
11: Trolleybus (As defined by https://en.wikipedia.org/wiki/Trolleybus)12: MonorailRegarding the other ones:
route_type=8could be defined as an "easter egg" route type (Catapult in CityMapper, Hogwarts Express in Transit, Dragon service, loch ness monster...) => Could be worth also addingroute_type=10commuter rail is already broadly represented usingroute_type=2, which is defined as "Rail. Used for intercity or long-distance traveled". => Should maybe not be addedroute_type=9intercity buses aren't really described in GTFS, they are usually represented withroute_type=3bus, which is defined by "Bus. Used for short- and long-distance bus routes.". => Should it be also added?This conversation can be linked to the extended route types, which are not part of the specification but which are officially used by Google.
(Ping: @aababilov )