Talk:Proposed features/Public Transport v2 Vehicle Type "coach"

From OpenStreetMap Wiki
Jump to: navigation, search

Dividing line is not clear

It is not clear if this proposal is about the vehicle type (so we might see a highway access-tagged as "coach=yes"), or the usage to which it is being put, or the routes for scheduled coach services.

The vehicle type (its construction) is basically no different to a bus in a legal sense, except that if it complies with certain extra regulations, it may be allowed a higher speed limit.

The usage is split into two types: the long-distance scheduled services, and private-hire arrangements. The scheduled services, and their routes, are not intrinsically different from bus routes, and could be distinguishable by means of operator and network tagging. Private hire to take groups of people from A to B is still called coach travel (in the UK) but it is a very different animal to the scheduled services. For all sorts of operational reasons, a bus vehicle may operate a coach route, or vice versa. The concept of "psv" may help here; a scheduled service may be a "public service" thus making the vehicles "psv", but once again, is it about the vehicle or the usage?

I am not sure if the scheduled services are allowed to use bus lanes. Private hire is almost certainly not, but that may depend on individual jurisdictions.

There are other grey areas as well... Is a vehicle returning to the depot after service still a "coach"? Is a vehicle operating a "rail replacement service" a bus or a scheduled coach or private hire (or is it still a train)?

At the moment I am not convinced that there is sufficient distinction between "bus" and "coach" to justify a new tag.

--Csmale (talk) 06:54, 4 October 2017 (UTC)


As the proposal is written now, I don't think it is clear enough about what distinguishes a "coach" route from a "bus" route. I can see that buses in a typical urban network with stops every block or two would be "bus", and that a route connecting two cities 300km apart would be "coach". But what about all the services that do not fall into either category? For examples buses connecting a city to surrounding towns and villages. These might have quite a few stops in the city itself, but fewer than city buses, and lots of stops in the surrounding region. Looking at the departures board for the main bus station in my city, the majority of destinations served fall into this kind of "provincial" or "regional" category, and I wouldn't know from this proposal how to tag them.

If the distinguishing factor is distance, isn't this already available from the route relation? If the important factor is that it connects at least two large cities, maybe "intercity" would be a clearer tag than coach, and make clear that the distinction is about the type of route rather than type of vehicle?

--Alan gr (talk) 12:46, 4 October 2017 (UTC)


I think the vague distinction between bus route and coach route can be identified in informations like route name and stops. Therefore a new type is redundant.

--Austin Zhu (talk) 18:36, 4 October 2017 (UTC)

@Austin please let's not put "(coach)" in the name or anything like that. The difference, if we ever agree on what that is, should end up in the tagging in some structured way.

The main thing needed for are feature like this is a list of attributes which discriminates the two kinds of buses/coach.

Possible distinction based on route vs stops

The longer the route, the less stops per km. Very long routes probably don't actually define the route, other than as a list of stops. The driver may have discretion about the actual route, possibly depending on traffic etc. Shorter city routes would tend to be the other way around - the actual route is sacred, and there may be a bit of flexibility in the stop locations, like jumping off at the traffic lights.

That would imply that there is not much point in making the ways members of the route relation in these cases. The stops, in order, are the limit of what is actually defined beforehand. --Csmale (talk) 19:07, 4 October 2017 (UTC)

That may be right in many countries where there are a lot of different roads available, but if there is only one highway between different cities then the service will always use the same route. So I think we can and should (at least in this cases) add the ways to the route relation following the Public Transport Scheme. --Ialokim (talk) 22:25, 4 October 2017 (UTC)

Other possible tagging scheme

Okay, I agree with @Csmale and @Alan_gr, the proposal is a bit confusing whether it is about the vehicle or the service type. Reading your opinions, I see that it is very difficult to distinguish the two types in all cases, because there are simply more than two. I'm sorry about this proposal without having been very clear myself of the concept :/

I agree to mark this proposal as canceled or obsoleted, but I'm convinced we should think of another type of distinction between different bus service types. I've been searching a bit more and found this discussion on the Buses-Talkpage. It dates from 2009 with the latest comment made in 2012 referring to the (than) newly approved Public Transport Scheme v2. But this scheme, as already stated in my proposal, lacks the differentiation between different bus route service types that also can be very important for rendering engines (long distance routes should be rendered at lower zoom level whereas urban routes could be omitted). The original Oxomoa-Scheme proposal indeed suggested a better method (IMHO) that seems to have been omitted in the approved proposal. Are there any arguments against a classification following that proposal? --Ialokim (talk) 21:32, 4 October 2017 (UTC)

I second this motion. I was wondering why this proposal wasn't just adding bus=<value>, didn't know that this already existed in the Oxomoa schema.
--Adavidson (talk) 23:12, 4 October 2017 (UTC)


Using another tag, rather than changing the meaning of route=, seems like a more promising approach and less dependent on the assumption that there are only two important categories. A minor point: I wonder if the specific bus= tag would cause confusion with stop_position, where it is a yes/no tag. Mappers might think that bus= on the route should match bus= on the stops.
Looking at the specific examples given in that schema, I feel that maybe separate concepts are being thrown in together. E.g. whether a service is express or stopping is independent of whether it is long-distance or not. Many cities have express airport buses as part of the urban network, or express variants of normal city routes. And many long-distance routes have both express and stopping services. Which value gets priority for this tag? Maybe something like stops=non-stop/limited/many should be a separate tag?
Is there an analogy between what you are trying to achieve and the use of network= on hiking and cycling routes? This allows routes to be classified into a hierarachy (international/national/regional/local). From what I have seen, mappers in specific countries tend to interpret this in a sensible way according to how trails are organised in their country. I think this may be used for rendering at various levels at waymarkedtrails.org (but I am not certain, it might just depend on length). It would have to be called something else for buses as network= is already used for the name of the network, but maybe something along the same lines? Of course hiking/cycling trails are easier in the sense that stops do not have to be considered.
--Alan gr (talk) 09:17, 5 October 2017 (UTC)


the key "bus" is already used nearly a million times: https://taginfo.openstreetmap.org/keys/bus#values
it is used as an access tag and also to say which kind of vehicle stops at a public_transport platform / stop.
Common values are:
"yes" 98%
"no" 0,5%
"designated" 0,5%
"urban" 0,1%
"unofficial" 0,1%
I agree the values you propose can be interesting, especially long_distance, train_replacement (although this is generally a temporary thing for hyperactive high density mapping areas), school, urban, "shuttle"
I wouldn't suburban, but would see them included in urban
intercity is also somehow unclear, as cities might be very close together (i.e. it's kind of "urban") or very far away. "long_distance" seems less ambiguous (althoug it is also relative, maybe your list lacks some kind of "regional" which can be used both, in metropolitan and in rural areas for "medium" distance routes).
On the other hand, while the distinction by function (school, train_replacement, shuttle, express) is not (clearly) inferrable from other properties, the distance travelled and the admin entities served can easily be gotten from looking at the actually route.
--Martin Koppenhoefer via mailing list


Thanks for your comments @Alan_gr and Martin Koppenhoefer, based on them I thought of separating the different ideas that "are thrown together" with different tags, I like your idea of using a "stops" key! And yes, it's a similar idea to the way it works on hiking and cycling routes.
What do you think about the following keys and values (but I'm still not very happy with the key names):
stops=non_stop shuttle service which only stops at start and end points
stops=limited express service which only stops at few (big) stations
stops=many regular service that stops at almost all the busstops between start and end point
bus=local or bus=urban route only inside towns / cities
bus=regional "suburban", route in towns / cities and rural area (villages, suburbs etc.), route length does not pass region size
bus=national or bus=long_distance route between towns and cities that are separated by a long distance
service=train_replacement temporary bus service to replace train/tram/etc. service
service=post post bus (still exists in some countries (e.g. Africa))
service=share_taxi share taxi (currently suggested as route=share_taxi on the Wikipage Relation:route#Bus routes (including trolley bus), but not part of the approved proposal!)
service=on_demand on demand bus service
service=school school bus service, only on peak times and not during vacations
time=night service only at night (e.g. nightliner etc.)
time=peak service only at peak times
The key stops=* already exists with the number of stops on a route, instead of bus=* I was thinking of extend=* or radius=*, the tag service=* already exists for highway/railway/waterway services (what about specialty=*) and I think for time=*, we should search another key too.
What about using the prefix "bus:" and therefore bus:stops=*, bus:radius=*, bus:specialty=* (or bus:type=*) and bus:time=*?
Also, the two latest keys should not be obligatory, but only applied in the case there is anything special.
add tourism to service/bus/bus_type/bus:speciality or how you want --AgusQui (talk) 18:21, 13 October 2017 (UTC)
OSM definitely needs a way to distinguish between metropolitan bus lines and coach / long distance lines. This proposal with bus=long_distance/local etc. tag seems very reasonable, so maybe you should start a new proposed features page? Also note that an important difference between "bus" and "coach" services is the ticket system. At least here most people using the metropolitan buses have a monthly pass to the network, whereas coach tickets are typically sold separately per trip. Keimo (talk) 08:43, 14 October 2017 (UTC)

tag network scope <> name

for bike, network tag is used to tag local <> regional scope maybe it could be useful to try to reduce differences in meaning with buses by having a unified way to map the type of network scope and network name

A welcome initiative, but ...

This is a much needed initiative, but I think the word coach is probably not the best choice (see Colin Smale's remarks above).

Bus/coach services have a large number of different types. Probably the best analogy is the varieties of trains from Inter-city, Regional, Suburban etc. As it stands route=coach is directed at limited-stop long distance routes, but there are several other types which I am not sure are adequate represented. These include (ones I've used):

  • Regional bus services which will stop at any stop but tend to have a fare structure which precludes local use. Typical example the Trans-Peak Relation 167839 (XML, iD, JOSM, Potlatch2, history, analyze, manage, gpx) from Derby to Stockport, Krakow-Zakopane, Bellinzona-Thusis or the Glasgow-Campbeltown service.
  • Express services between cities. Nottingham-Derby has an express bus Relation 366750 (XML, iD, JOSM, Potlatch2, history, analyze, manage, gpx) with one intermediate stop operated by well-equipped coaches.
  • Express commuter services (e.g., Oxford-London Relation 164517 (XML, iD, JOSM, Potlatch2, history, analyze, manage, gpx))
  • Shuttle services (e.g, airport buses, park-and-ride services) with limited stops
  • Long-distance coach services such as those provided by National Express in Britain, Riga-Tallinn, Tallinn-Turku, Puerto Madryn-Bariloche, Geneva-Val d'Isere, Madrid-Caceres. (this is the class mentioned in the proposal). It is in places like the Baltic States, Spain, Argentina etc where long-distance coach travel is the main overland transport network that this information is most important. The ability to render such routes in a similar way to (often) less-frequent rail services would be a major useful enhancement.

I'm sure I've forgotten some categories, but I think the way forward is to ensure that we can categorise bus routes more precisely. Note that none of the above were operated as part of a local bus network. Long-distance bus travel in Poland may not fit either. Many routes are operated by small buses (20-30 seats), stop in most towns and villages and are a pretty unpleasant way of travelling any distance. SK53 (talk) 18:37, 12 October 2017 (UTC)

New proposal published

See Proposed features/Differentiation for routes of public transport for the new proposal. --Ialokim (talk) 16:34, 20 October 2017 (UTC)