[OSM-talk] A new highway tagging scheme - thinking about tagging highways for cycling and walking as well
peter.miller at itoworld.com
Mon Aug 27 09:34:48 BST 2007
I suspect that we are going to tease out national differences in road
classification and that our model should be able to accommodate this
differences. We also need to agree a way of classifying Ways for different
sorts of users.
There seems to be agreement that the 'highway' tag should really be an
indication of it's drivability, ie how useful it is for the driver of a
vehicle, rather than its classification (A road, B road etc). btw, I have
noticed that Google Maps don't rigidly stick to 'Red for A road' , 'brown
for B road' etc, it seems to be much more about relative levels of use. For
example, in the example linked to below there are two B roads coming off the
A1152 as it heads east out of Woodbridge, the first, the B1083 heads south
before the golf course, and the second, the B1084 (Orford Road) heads East
after the golf course. One is rendered by Google as a very minor road, the
other as a 'tertiary' road and it is certainly true that Orford Road is the
busier road. Also, follow the A1152 further east and notice that it turns
into the B1069 the other side of Rendlesham (according to my mapping) but
Google keep it red and fail to show the new road class and code.
In the past I tagged roads strictly according to their classification
(trunk, A, B, etc), but I now tag them so that the drivability of the place
is revealed. Personally I agree that administration should be a completely
Note however that as a cyclist or as a walker the classification is
completely different. How about keeping the 'highway' tag for classification
of drivable routes for motor vehicles, and invent something else for
cyclists and other groups of road users?
For cyclists a national cycle route would probably be given the highest
cycle classification, a regional cycle route the next one, and a local cycle
route the 3rd level. Other Ways recommended for cycling could also be tagged
with a cycle-classification and Ways can also be tagged with 'no' if they
should not be used.
To keep it open for later development, would something along these lines be
class_cycle=[1-9]; where 1 =highest (national cycle route), 2 next (regional
cycle route, 3 (local cycle route), 5=mountain bike only, 9 = non-cycle-able
class_walk=[1-9]; where 1=national walking route (Pennine Way etc), 2
regional walking route , 9=non-walkable
class_ride[1-9]; where 1=idea soft wide legal surface for galloping and so
I guess the current highway tags could alias onto the class_drive tag and
the highway tag 'residential' would alias onto
class_drive=unclassified;abutters=residential. So we would also have
Class_drive[1-9]; where highway='motorway' == 1, 'trunk' == 2 etc
The class-drive tag would no longer be used for ways that are not accessible
at all by motorised traffic, the tag would be 'class_drive=no' and the
Highway tags 'footway', 'cycleway' and 'bridleway' would be aliased onto the
appropriate class_cycle/ class_walk/ class_ride. Are these actually
extensions of the foot, horse, bicycle tags?
This change has a number of benefits:
It is extensible to new modes
It is easy to render for any mode
It teases out issues to do with surface and abutters from the highway tag
It also allows different classifications to be used for the same Way for
different user groups
It is orthogonal and does not have the traditional bias towards the car, the
car is just one of a number of different modes
All this is very relevant to the new cycle mapping from gravitystorm, a fine
OSM project that is tackling the issue of rendering cycle mapping as we
> Message: 6
> Date: Mon, 27 Aug 2007 04:00:20 +0200
> From: Frederik Ramm <frederik at remote.org>
> Subject: Re: [OSM-talk] A new highway tagging scheme
> To: talk at openstreetmap.org
> Message-ID: <20070827020019.GA28425 at lochewe.mathy.remote.org>
> Content-Type: text/plain; charset=iso-8859-1
> > Often the administrative classification of a road is unclear / not
> > Again, as far as I know, there's no obvious way where I am to determine
> > the legal classification of a road as trunk/primary/secondary/etc from
> > any signage on the road itself.
> Any attempt at distinguishing between physical/administrative
> situation should allow both *or* only one of them to be used, with
> software expected to use appropriate logic to make assumptions about
> the missing information.
> I may be on the ground and thus able to authoritatively tag the
> physical side of things; or I may be using an out-of-copyright (or
> otherwise legal) map which will give me the administrative status of
> something but I'll have no idea what it is like on the ground.
> On another note, as far as I understand Andy Robinson is working on a
> general (not just limited to highways) concept for making the
> distinction between physical and administrative aspects. Search for
> "STAGS" on the Wiki or the lists (there was a presentation at SOTM). I
> believe the concept has evolved a bit since it was presented, and has
> also been renamed (to "Cosmic" or "Cosmos" if I am not mistaken).
> Frederik Ramm ## eMail frederik at remote.org ## N49?00.09' E008?23.33'
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the talk