[OSM-talk] Contraflow bus lane

[Andy Robinson (blackadder-lists)]

David Earl wrote:
>Sent: 23 October 2008 9:25 AM
>To: Richard Fairhurst
>Cc: osm
>Subject: Re: [OSM-talk] Contraflow bus lane
>
>On 23/10/2008 09:00, Richard Fairhurst wrote:
>> Andy Robinson (blackadder-lists) wrote:
>>
>>> I also have one in the centre of Sutton Coldfield. Think when I
>>> mapped it
>>> (ages and ages ago) I made it two separate ways which obviously
>>> isn't right.
>>> So interested to hear what people think.
>>
>> +1. There's one in Worcester. Except it's labelled (in the car-flow
>> direction) by "BUS & CYCLE" painted on the road, where cycle is the
>> usual pictogram.
>>
>> Consequently you could, and I almost have, cycle along there in the
>> belief that it's a with-flow cycle lane... only to be flattened by an
>> oncoming bus. Not funny (well, maybe a bit).
>
>Was that a +1 for separate ways or a +1 for "which obviously isn't
>right"? The latter I hope.
>
>Given the existing use of psv=opposite_lane I'll go with that for now.
>(I don't think it does clash with psv=other things BTW). psv rather than
>bus is less than obvious for non-English speakers. Changing to
>bus=opposite_lane is merely linguistic tinkering, though I know how much
>that excites people on this list.
>
>However, it occurs to me that there is a more general problem here. Many
>roads are divided into lanes with different uses (obviously with some
>basic defaults), including different types of vehicle - bus, with or
>without taxi, bicycle, motorbike, hgv; cycles with other traffic allowed
>too or not; multiple vehicle occupancy; ... - and different types of
>permitted manoeuvre - right turn only; left turn only; ...
>
>The x=lane notation is going to bend at the knees doing this. Maybe we
>need a general mechanism that describes each lane when the road is more
>than a simple footway + n any-vehicle traffic lanes arrangement. Perhaps
>include the ones on the verge too. Hills Road in Cambridge
>(http://www.openstreetmap.org/export/embed.html?bbox=0.13402,52.17604,0.149
>51,52.18583&layer=mapnik&marker=52.18100,0.14119")
>would need a support description like this:
>
>footway (on verge)
>cycle track (on verge)
>cycle lane (mandatory) (*)
>any vehicle
>any vehicle opposite
>cycle lane opposite
>footway


I see exactly this type of approach in the longer term but was not sure if
it should all be an OSM thing or whether I should think about it in terms of
a gazetteer approach, ie all the attributes and information about a street
stored in a different database but using OSM data as the starting point.
Obviously though I can see the benefits of including everything within OSM
itself if it can be made to work.

Because streets are liner features between logical node points its
relatively easy to define the sequence as you cross from one side of the
highway (Right-of-way) envelope to the other, adding widths and descriptions
for each characteristic from grass to tarmac, kerb details the lot (if you
are so inclined). The only issue is way direction for 2-way roads in that
the information becomes erroneous if the way direction is swapped.

Cheers

Andy