On Fri, May 30, 2008 at 11:25 AM, Andy Allan <[EMAIL PROTECTED]> wrote: > Other words *could* be mapped into the same numbers. But since we can > see quite clearly that there are more than 10 types of administrative > boundaries in the world, and different people have different opinions > as to which are equivalent, what advantage is there in trying to > shoe-horn them into such a narrow set? And can anybody, in advance, > name every boundary type in the entire world and get the inter-nation > equivalence correct and uncontroversial? I think not.
Because 99% of applications don't care the slightest how the internal subdivisions of some random country in the world compare to those in england. All renderers care about is "is boundary A more or less important than boundary B". Numbers work perfect for this and that's why they're used. There is absolutely no implication that admin_level=4 is the same everywhere. We know it isn't but we do know it's within a country, because countries are admin_level=2. Just like highway=tertiary is not the same everywhere in the world, but customised to the local situation. > I think we should store the actual boundary types, and if a user of > the data (e.g. a renderer) considers that English counties are > equivalent to US states then he can process them into both being the > same numerical value. If he considers English counties and US counties > to be equivalent, he can do so too. So the numerical equivalence table > should be on the rendering end of things, and the database should > store the actual factual data. If you are willing to maintain a mapping table for the 160+ counties in the world and maintain the megabytes of rendering rules that would be required to make it render sensebly then we can talk. Until then let us use admin_level and be done with it. Have a nice day, -- Martijn van Oosterhout <[EMAIL PROTECTED]> http://svana.org/kleptog/ _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
