[Tagging] Dispute prevention: meaning of lanes tag
imagic.osm at gmail.com
Mon Apr 23 14:55:25 BST 2012
In my opinion, parked cars should not be considered. Keep it simple!
The tag lanes should give you the number of lanes. The tag width gives
you the width. And if cars usually park on the road, we should use a
tag - maybe parking:lanes or similar - for that one.
What we not should do is put a completely different meaning into a tag
-> "Actually it is lanes=3 but I only tag lanes=1 because cars are
2012/4/23 Kytömaa Lauri <lauri.kytomaa at aalto.fi>:
>>Am 21.04.2012 um 13:34 schrieb "Ilpo Järvinen" <ilpo.jarvinen at helsinki.fi>:
>>> ...What I don't really care if it is called lanes=1.5 or
>>> lanes=1/2+some_other_agreed_tag_which_is_not_an_estimated_width=x, but
>>> simply saying that use lanes=1/2 alone instead I oppose.
>>I would recommend lanes=2 and width=xxx. Maybe give some examples for the widths of some common, "narrow" roads? Can someone provide photos and widths?
> There are a few things to this, that haven't yet been mentioned. I've
> been writing this as the discussion has progressed, so this got a bit
> long, but I tried to rearrange it to a comprehensible presentation of
> the issues.
> All this discussion is taking place, because there are roads where
> lanes=2 would be wrong part of the time, or for some motorized road
> users, and lanes=1 would be wrong on that same road at some other
> times. IMO we should not omit the lanes tag altogether on these
> roads, when the "between 1 and 2" tells us something significant of
> the attainable speeds and the layout of that road. Even if the
> actual width is measured and entered.
> First: I give you two clear examples why we must not limit the
> lanes tag to roads with a painted line:
> Urban road 9 meters wide (measured from aerial images), parked cars
> on one side (always), no markings. Definitively lanes=2 - that's 3,5
> meter wide lanes, enough for the bus that runs here.
> Rural road less than 5.7 meters(*), probably 5.5m. No markings. Low
> traffic, passenger car and a hgv can pass even if most drivers slow
> down a bit because they have to drive so close to the edge. Has
> passing places for the easier driving in the rare case of oncoming
> hgv's, even if they can fit side by side with only few cm margin.
> Passing places are also in place for winter time, when the snowplowed
> road edge has too much snow for driving safely in a straight line,
> when a bus and a passenger car need to fit side by side. In winter
> two passenger cars would most of the time disregard the passing
> I'd still say lanes=2:
> *) Here the road marking rules state that "generally" no center line
> markings are used on roads less than 5.7 meters wide; even that is
> 2.8 meters per lane, i.e. enough for the widest road legal hgv's to
> pass. I'd believe the point being, that a center line on a road
> narrower than that would make it impossible for the hgv to stay
> within the lane it is supposed be driving in.
> About car widths: typical European car widths are 1.80 meters, give
> or take 5 cm. That does not include mirrors. That is why a row of
> parked cars takes up 2 meters from the road width. (Also, not
> everybody can park every time less than 5 cm from the road edge.)
> The widths have grown some 20 cm between the 1970's and present day.
> Likewise, the 2.55/2.60 meters for trucks and buses does not include
> Second point: often we would, I believe, assume that a road tagged
> as having two lanes "generally always" allows unimpeded traffic flow
> in both directions. Where it's 4.2-4.4 meters wide, that holds for
> passenger cars. A case I often see, especially in urban areas built
> between 1920 and 1960, are residential roads that are 5.5 to 6
> meters wide without a center line, but allow parking on one side of
> the road carriageway - and they're generally often full of parked
> cars. It's not a "parking lane", but part of the road reserved for
> traffic; when there are no parked cars, even hgv's could pass each
> other with care. On 5.5 meters wide roads, many passenger car
> drivers will wait at a random free space between the parked cars,
> but on a 6 m wide road people will just slow to a crawl (or halt)
> when passing oncoming cars. The wider one could be lanes=2, but the
> narrower one hardly. See below for streetview examples, the last two
> example links.
> Third point: were we to estimate widths visually (not all areas
> have good enough aerial imagery), there would be lots of "relative
> errors" between nearby roads: IMO it's bad if such measures are
> recorded that claim that a road is narrower than the next, when it's
> the other way round - especially when the claimed-to-be narrower
> road might have two clear lanes, whereas the lane count isn't that
> unambigous on the other road. With lanes=1.5 we have a rough scale,
> but one that's correct relative to nearby roads with definitively
> 1 or 2 lanes.
> Examples, both clear cases and ones that are to date without a
> lanes tag, or with lanes=1.5. None of the examples below are oneway
> roads. Only now did I measure the width from Bing aerials.
> One can take the phrase "Generally always" below to mean that if you
> were to go there on 10 different days, on 8 or 9 days the situation
> would be as depicted.
> Clear cases of lanes=2:
> Rural road 7.5 meters wide (+shoulders), marked lanes, lots of margin
> within the lanes for the widest of vehicles.
> Rural road about 6.5 to 7 meters, marked lanes, even a hgv still has
> lots of margin within their own lane.
> Not so clear cases:
> 8 meters wide residential road, parking allowed on both sides
> (roughly always half of the allowed space is in use), no markings:
> Where cars are parked on both sides, only 4 meters is very tight but
> possibly manageable for passenger cars. I have long ago tagged this
> as lanes=2
> 7 meters wide, parked cars generally always on both sides. That's 3
> meters of free road. Can that be lanes=1? What if there were most of
> the time much fewer parked cars?
> 6.7 meters wide, parked cars always on one side. Should be 4.7
> meters free for traffic, ample room for two passenger cars. Here the
> cars parked on the left side are illegally half on the curb so that
> they leave the required 3 meters of road free for traffic - there's
> is no formal parking restriction on either side of the road.
> Urban unpaved residential, width between the fences is about 5.2 to
> 5.7 meters - hard to measure from the aerial images because of the
> trees - and the carriageway itself a bit (up to 1.5 meters) narrower.
> As you can see from the parked car, a hgv (say, a garbage truck)
> could just barely get past it, and even passenger cars need to use
> the very edge of the road in case of oncoming vehicles. This one I
> have long ago tagged as lanes=1.5; it's not one lane road, but you
> couldn't call it two lanes either.
> 5.5 meters, parking allowed actually on both sides but everybody
> parks on one side of the road, less than half of the allowed space
> normally used by parking, no markings. I can't say if this is
> lanes=1 or lanes=2 - if parking were forbidden totally forbidden,
> it would be 2. Quite often you have to wait for a few seconds for
> the lane to clear.
> 5 meters wide, otherwise just as above.
> * some of these roads can't have a lanes tag at all,
> * or they're misleadingly tagged as lanes=1
> * or misleadingly as lanes=2
> * or tagged with lanes=1.5. And not all of those are 4 to 4.2 meters
> wide, but even wider.
> I'll still have to dig to find a case where such a border case narrow
> road has dedicated (painted) parking spaces, which excludes them from
> the width usable for traffic.
> Tagging mailing list
> Tagging at openstreetmap.org
More information about the Tagging