OpenStreetMap
Аксиомы классификации дорог по важности
Posted by Zkir on 13 August 2012 in Russian (Русский). Last updated on 22 August 2012.Эти аксиомы призваны дополнить существующие принципы классификации , и прояснить то, что было оставлено за кадром.
- Тупиковый участок от конца дороги до ближайшего перекрестка может иметь только самый низший статус( уровень важности) - residential или unclassified.
- Статус может меняться только на перекрестках, или на границах населенного пункта, если в нем отсутствуют перекрестки.
- Граф дорог, в который входят дороги некоторого статуса и выше, должен быть связным.
Кто-нибудь предложит четвертую аксиому? ;)
(Это, разумеется, затравка для предварительного обсуждения)
Discussion
Comment from xushaoaa123 on 13 August 2012 at 10:08
http://www.burberryzsale.com/ Burberry celebration plc is frequently a British luxury style house, producing clothing, fragrance, and style accessories.
Comment from Sanderd17 on 13 August 2012 at 10:29
К сожалению, я не говорю на русском, это было переведено с Google Translate.
Несмотря на то, как правило, верно, что дорога классификации лишь изменения в переходах, я не думаю, что это всегда так.
Иногда дорога классификация также изменения в границы (государственных границ э даже границах сообщества). Как здесьhttps://maps.google.com/?ll=50.934538,3.079777&spn=0.009155,0.026157&t=h&z=16&layer=c&cbll=50.934601,3.079926&panoid=0b8wFrYgiY_Lp4JquVjlxQ&cbp=12,237.07,,0,18.09 , На границе мое Oostnieuwkerke деревни и города Roeselare, дорога изменений с двух полос (классифицируется как третичный) в одну полосу (или несекретные). Или вот, на франко-бельгийской границы: http://www.openstreetmap.org/?lat=50.780064&lon=2.739279&zoom=18&layers=M дороге не изменится в виду, но Бельгии и Франции использовать различные схемы классификации.
Original text:
Although it is generally true that road classification only changes at crossings, I don’t think it’s always that way.
Sometimes, road classification also changes at borders (state borders er even community borders). Like here https://maps.google.com/?ll=50.934538,3.079777&spn=0.009155,0.026157&t=h&z=16&layer=c&cbll=50.934601,3.079926&panoid=0b8wFrYgiY_Lp4JquVjlxQ&cbp=12,237.07,,0,18.09 , at the border of my village Oostnieuwkerke and the city Roeselare, the road changes from two lanes (classified as tertiary) to one lane (or unclassified). Or here, at the French-Belgian boundary: http://www.openstreetmap.org/?lat=50.780064&lon=2.739279&zoom=18&layers=M the road doesn’t change in view, but Belgium and France use different classification schemes.
Comment from osmisto on 13 August 2012 at 10:43
Статус на границе НП менятся не может? Например повышаться: secondary ведёт в town, но в самом городе она primary, или даже транзитная в другой город.
Comment from osmisto on 13 August 2012 at 10:44
Oh :) I forgot to reload before posting my comment
Comment from Zverik on 13 August 2012 at 11:19
Sanderd17, I can’t but comment that in Russia, physical attributes of a road do not affect its highway tag value. The latter is defined as an importance of the road in the routing graph. That is, the more drivers choose the road for their trip, the higher value is set for highway tag. Neither official classification (E105), nor surface quality and number of lanes do not matter.
Comment from Zverik on 13 August 2012 at 11:20
(sorry, do not matter -> matter)
Comment from Diomas on 13 August 2012 at 13:28
А если дорога ведет к 2-3 населенным пунктам которые находятся на этой дороге друг за другом и в этих населенных пунктах нет перекрестков?
Comment from Zkir on 13 August 2012 at 13:42
Про границы НП - очень дельное замечание. Это надо будет учесть. Обычно так бывает, когда unclassified меняется на residential. Но это один и тот же уровень значимости. Может ли tertiary меняться на границе на residential, надо подумать :)
//Например повышаться: secondary ведёт в town, но в самом городе она primary.
Так быть категорически не должно - это нарушение третьей аксиомы. Главная улица в городе не может иметь статус выше, чем самая статусная дорога к нему. Если я выбрал дороги primary и выше, у меня не должен получится висяк в виде главной улицы уездного города N.
Comment from Zkir on 13 August 2012 at 15:29
Sanderd17, I agree that status can change (in special cases) on the city boundary. Obviously status can also change on the borders of the countries too.
But note, that in Russia classification by relative importance in the road graph is used. So those axioms are about this kind of road classification.
Also, it’s a BIG problem for me, that in Europe not all contries use this scheme, because it is not possible to create routable overview map of Europe. It is not possible to fetch main roads by highway tag, because they will to make connected graph.
Comment from Zkir on 13 August 2012 at 15:30
… because they will NOT make connected graph.
Comment from Sanderd17 on 14 August 2012 at 06:52
And how do you calculate the “relative importance in the road graph”? We have no data to determine that.
We only see the data on the ground. That is the official classification and the design of the road. We try to determine the relative importance from that data:
Exxx and Axx -> motorway
Nx and Nxx with special “autoweg” sign -> trunk
Nx and Nxx -> primary
Nxxx -> secondary
two lanes -> tertiary
one lane -> residential or unclassified
This was done in the hope that our state makes those decisions based on the importance of the road. You may know that Belgium has a very bad spacial ordering. We don’t have the normal residential centres but lint-construction. There used to be centres in the middle ages, those were connected with roads, and from than on, people build their houses along those roads. So you get huge lints of buildings.
The example I showed you is also this case: http://www.openstreetmap.org/?lat=50.934473&lon=3.080182&zoom=18&layers=M
You see that the teriary road starts together with the lint of houses. All those people living there have to be frequently in the city. So it’s difficult to determine the importance, as it accumulates while you go via a lint.
Btw, our state doesn’t give us the best data. The recently turned a primary road (with the ref N37) in a no-exit road. Off coarse, in cases like that, we deviate from our scheme to follow the numbering.
And because it’s so difficult to calculate the “relative importance”, Belgium indeed has some different definitions than France. Btw, you will see that the routing graph is still mostly connected (although the best way often leads over a few residential roads or tertiary).
Comment from Zkir on 22 August 2012 at 15:44
We have guidelines.
If in short - traffic generators should be assigned with low importance (residential, unclassified) and traffic conductors should be assigned with high importance (tertiary-secondary-primary-trunk)
Traffic conductors are also classified by kind of the traffic. Very-very roughly: traffic between Moscow and region capitals is considered trunk, traffic between regions - primary, traffic between towns of the same region - primary, traffic between town and villages inside region - tertiary.
Comment from Zkir on 22 August 2012 at 15:45
This is possible because we still have very distinct population centers (cities, towns and villages) in Russia.
Comment from Sanderd17 on 22 August 2012 at 16:24
This is indeed very clear with distinct population centres. But that wouldn’t solve the road classification inside a city as Moscow.
Some consider Flanders to be one big city with some holes in it (there’s not a single place where you can’t see a building around you). So I’m particularly interested in the classification inside those big cities.
See Moscow and the central part of Flanders on the same zoom level:
http://www.openstreetmap.org/?lat=51.068&lon=4.27&zoom=10&layers=M http://www.openstreetmap.org/?lat=55.732&lon=37.715&zoom=10&layers=M
But because Flanders didn’t evolve from a single city, we don’t have a motorway around it, but only motorways through Flanders. But still around the old cities.
I think, in Moscow, the situation is more clear than in Flanders. In Moscow, you can assume a star-shaped road network to start from, and with the rules such as “trunk may only end in primary roads”, you have a decent classification. While we have something quite random to start with.
So I guess we’re quite stuck with our classification (because our government never saw the benefit of spacial ordering, and nobody in Flanders wants to live in a city).