2017-12-10 13:11 GMT+01:00 Ilya Zverev <i...@zverev.info>: > subway entrances and elevators: you suggest to add both tags to the same >> object, but subway entrances are defined as nodes only while elevators can >> be mapped as well with a way. > > > > Well, you could add a subway_entrance node on a contour in that case. In > practice, I've yet to see an subway entrance elevator mapped not as a node, > and I've seen all the metro systems in Germany, Austria, Helsinki and some > other cities that allow entering a station via elevators.
I don't know if there are situations where this currently is a problem, but the proposed suggestion is creating a conflict with the elevator documentation, and there really is not need to require it for all cases to have the elevator tags and the entrance tags on the same object. It doesn't improve your proposal to require this (IMHO), you could rephrase it like: "the subway entrance nodes should be put at all places where you can enter a subway station (btw.: with underground railway stations there is no difference, although we don't provide a semantically nice tag for this currently), this might be an elevator. (e.g. in case of an elevator mapped as line you could add the subway entrance tag to the "streetlevel" (usually) node of the elevator). > > > linking stop positions and platforms via relation with a station: if the >> station is mapped as an area, this shouldn’t be needed, as there’s already >> a spatial link >> > > Oh come on, we're talking about underground stations. A spatial link will > get you all the bus platforms overground. > you would have to take the level / location / layer tags into account > > Also, imposing using a spatial link for a distributed number of objects > instead of a small relation is not thinking about data consumers. Which PT > mappers have been doing for 9 years now, but should it stop somewhere? yes, admittedly it requires an additional preprocessing step. I'm unsure on this, because generally we put easy mapping and data stability above ease of data consuming, but we also help applications a lot (in other fields as well) where in theory they could also figure things out themselves by analyzing the data. > why are tracks optional for route relations? Isn’t that in contradiction >> with the generic route relation definition? Shouldn’t this become a >> different type of relation if it isn’t about an explicit route anymore? >> > > In my opinion, tracks should be optional for all public transport route > relations, but let's start with rapid transit routes, when most of the time > you don't see tracks, because they are underground. Mapping what you don't > see and cannot check on the ground is questionable. Requiring to do that > mapping from people who just want to mark a route with stops is plain bad. > You can check underground tracks on the ground, at least in some cases, e.g. by finding ventilation shafts. Or by finding historic imagery from the time of construction. Or by importing publicly available data. It might even be possible to do measure them with accelerometers in the train. > Regarding the first responses to the proposal, I was even angry for a few > minutes. I say "please discuss at the wiki" — there are no comments yet. I > spend few more hours to polish a proposal — nobody cares. And had I not > been a senseless drone, I'd be angry still, writing disappointed posts in > every media about the community. Because starting your replies on a > positive tone should be simple and would frame responses much more polite > and caring. For now it looks like everyone here wants to maintain a status > quo, because... I don't know why, 50 of 60 rapid transit networks in Europe > already adheres to the proposed changes. Interesting you speak about being angry and mention 50-60 rapid transit networks in Europe in the same paragraph as already adhering to the proposed changes, because I was recently "angry" as well, and it might fit into this context: I spotted a new mapper with a particular interest in subway and railway mapping, maybe you have also heard about them, the user name is IraSergeeva, and although they have only a total of 18 edits, all of them seem to be about railway mapping in metropolitan areas in Europe. Those edits are unfortunately terribly inconsistent and also bare of any facts: they managed to mark all main (some of them international / highspeed) railway stations in Rome as "light_rail" stations, as well as (that's one of the reasons why I wrote "inconsistent") those smaller railway stations on the railways to Naples, Florence and Pisa. I have spent some hours researching and fixing the less obvious cases, but it really isn't good style. Interested by the Rome incident, I started to look around and found similar remote mapping activities in other areas, e.g. in Madrid there was a changeset with the meaningful description "stop:areas" but actually changing a lot of stations to light_rail: E.g. this https://www.openstreetmap.org/node/2006607646 which by some basic internet research isn't a light rail station, as isn't this one from the same changeset: https://www.openstreetmap.org/node/286225469 (just query the Spanish railroad provider a route from Madrid to Salamanca and you'll see that trains stop at these places). Judging from the context and given the obvious errors, these are automated edits by armchair mappers (in these areas they are not local and probably not familiar with the actual situation). (I didn't check all of this edit, the mentioned stations above are 2 out of 3 I checked, so there are likely many more errors). Also in Athens, Greece, I found suspicious action from the same time period (and yet another user, with only 12 total edits and a concentrated in interested in European metro railway systems): https://www.openstreetmap.org/changeset/53163011 E.g. here https://www.openstreetmap.org/node/251083102/history what is apparently representing both, a Metro Station and a Train Station, becomes effectively a metro station by adding station=subway without the mapper creating the now required train station. While I corrected the Roman stations, I didn't mess with the Spanish or Greek ones because I lack the required local knowledge, but I suggest the Madrid and Athens community should check the station tags on railway stations. Cheers, Martin
_______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
