[Tagging] Dispute prevention: meaning of lanes tag

Martin Vonwald imagic.osm at gmail.com
Fri Apr 20 08:09:26 BST 2012


2012/4/19 Alan Mintz <Alan_Mintz+OSM at earthlink.net>:
>> * PSV lanes SHOULD be included (also [2]). Example: lanes=3 and
>> lanes:psv=1 means we have three lanes and one OF THEM is for PSV only.
> Same goes for HOV (high-occupancy-vehicles) lanes, unless they are
> separately mapped (which is a better solution for routing, given their
> controlled access).

I will think about a phrase, that will cover all those lanes. For
english and russian: suggestions from native speakers are welcome!


>> * Turn lanes SHOULD be included (see [2] and [5]).
>> * The lane count should change, as soon as a) new lane has reached its
>> full width or b) a lane starts to disappear (usually a merge with
>> another lane) (also [5]).
> Technically, yes, but it doesn't seem practical in developed areas in the
> US, which typically change lane configurations at every major intersection
> and then change back again.
> ....

Yes - and no. That's called micromapping. I fully agree with you, that
under normal circumstances it should not be necessary. But for example
on motorways I actually tag this way, especially since turning lanes
can be properly mapped. This way routers could precisely determine
e.g. the start and end of lanes exiting the motorway and give very
accurate instructions.
As there are no obvious reasons to not include turning lanes, we
should not exclude them. But I think about adding a statement, that
usually only on major roads or very complex junctions those lanes are
actually mapped. Can we agree on this?


>>  - Two-way roads with a specified lane count, but without a specified
>> lanes:forward OR lanes:backward and a lane count, that is divisible by
>> two, are assumed to have half of the lanes in each direction, e.g.
>> lanes=4 means two lanes in each direction if not specified otherwise.
>> I will add a recommendation for this situation, to add explicit
>> values.
> If an odd number, assume a center turn lane (e.g. lanes=5 means 2 forward, 2
> backward, 1 center).

This is simply not working that way. If we would use that assumption,
we would assume a lot of center turn lanes in Austria. I don't know 1
(in words: one) of them. Completely omitting those default assumptions
might also not be a good idea, because in my opinion it should not be
necessary to tag the lanes count on e.g. normal residential roads.
How about a table for the most common types of roads? Example:
residential is assumed to have one lane if one-way, two otherwise. For
motorways and trunks I would not add any assumptions, because they
simply differ too much.
Can we agree on that?

Martin



More information about the Tagging mailing list