OpenStreetMap
Let me get this straight for American roads:
Motorway: Interstates and other Freeway/Expressway grade roads
Trunk: Divided highways with graded and level intersections (ex: parkways in NY, NY 12 north to S of Boonville, NY 840 in NY Mills, the old NY 17, maybe 5 between Amsterdam city line to NY 103??)
Primary: US Highways, and I believe State highways with high levels of importance (ex: US 9, US 20, possibly NY 5, NY 7??)
Secondary: Most State highways
Tertiary: County highways
Residential: Most town-maintained roads
Correct?
讨论
42429 于 2010年04月11日 20:13 的评论
Right! At least in New York!
However, each state has its own system. In some states, like North Carolina and Virginia, State highways also include the lowest level of rural road network, which are designated as tertiary, whereas other states, like Minnesota and Wisconsin, have County highways of high importance, which are designated as secondary.
Baloo Uriza 于 2010年04月11日 20:49 的评论
In general, yes, that's a good starting point. Depending on local conditions, a road may rank higher or lower.
asciipip 于 2010年04月12日 11:46 的评论
That's a reasonable set of criteria. I tend to go with the US Highway Functional Classification System ( osm.wiki/Highway_Functional_Classification_System ) for tagging primary/secondary/tertiary/unclassified. So the designation of who maintains the road is less important than how the road functions: non-motorway/trunk roads that are central to moving traffic between regions are primary, roads that feed traffic into the primaries are secondary, and roads that feed traffic into the secondaries are tertiary. Anything for local traffic (i.e. not used as a through street) is residential if it's used for accessing residences in an urban/suburban setting, and unclassified otherwise (rural, or used for accessing business/industry in an urban setting).
Unfortunately, there's not really a "correct" set of rules. A lot of the classification is fuzzy, so take a look at what other people have done in your area, if possible (don't blindly trust the classifications from the TIGER import; they're often reasonable but not always best).