[OSM-talk] Advanced highway tagging

[Karl Newman]

On Nov 20, 2007 3:24 PM, Andrew MacKinnon <andrewpmk at gmail.com> wrote:
> > Do you have any examples of other maps where road drawing is not based
> > solely on objective features? I'd be interested to see how they work.
> > Or are you breaking new (cartographical) ground with trying to combine
> > autocartography with subjectivity? I can imagine specific cartography
> > (a map of a limited set of roads, drawn individually by someone)
> > incorporating specific subjective features, but I'd be interested in
> > altases or similar doing so).
>
> All commercial online maps decide which roads are major and which are
> minor on a subjective basis. Obviously, the fact that there is a road
> at point x is completely objective, but deciding how important a road
> is is at least somewhat subjective. Look at any of the commercial
> online map services - nearly all of them decide whether a road should
> be coloured as major or minor in a somewhat arbitrary fashion, not
> directly based on the road designation, the width of the road, the
> number of lanes, or any other objectively measurable factor or
> mathematically derived combination of objective factors.
>
I would bet that they actually are derived from some objective
factors; those factors just may not be known or obvious. Some
objective factors which could influence the categorization: official
designation, lanes, traffic (can be quantified as a frequency:
cars/day), surface type, vehicle weight rating, vertical clearance,
single/dual carriageway, traffic controls, ramp-only access, speed
limit, etc. Obviously at some point a map maker has to subjectively
decide which combination of those factors are important, but that
decision should be left to the data consumers, not stuck in the data
itself. It's the old Model-View-Controller idea. The map vector data
is the model; you get to decide what view of that data you want to
present.

Karl