I think in parts of this discussion we are confusing grouping and categorisation of facts with subjectivity and information loss.
For example, ski runs are categorised into Green/Blue/Black runs. A run may be classified as black if it exceeds a certain narrowness, or a certain roughness, or a certain gradient, or a certain length/duration. That doesn't make the classification scheme subjective. It can be a largely objective classification, based on specific facts that are verifiable, leading to a higher level classification. Information is lost in the categorisation, but this doesn't make it unverifiable. Cartography is necessarily a simplification and categorisation of what is on the ground, otherwise in the extreme case we end up with just a 3D image of the road. Good cartography is preserving the right information. At the highest level we've chosen to classify roads as primary/secondary, etc. We could perhaps have instead use number of lanes, average traffic speed, average traffic capacity, etc and left the classification to an algorithm based on those facts. The answer here seems to be that we need to have a classification scheme based on verifiable criteria. I think the classifications being proposed largely meet this. We also need to have the corresponding tags to identify (at least) those input criteria so we can capture the extra information when possible. I see this as a particular issue with 4wd tracks, where one trip through with a grader or rain can make a huge difference to road condition. Ian.
_______________________________________________ Talk-au mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk-au
