"strategic" may not be the right word (my english is rusty :) ) The thing is, I want to avoid those dotted highway=trunk like this : https://www.openstreetmap.org/#map=13/48.5884/-1.4035 (trunk then primary in the town then trunk again), I'd prefer trunk - still trunk in the town - trunk, like in England : https://www.openstreetmap.org/#map=14/51.1057/-2.1245
2017-08-19 14:00 GMT+02:00 Greg Troxel <g...@lexort.com>: > > Colin Smale <colin.sm...@xs4all.nl> writes: > > > Interesting approach, which might work for Europe, but at the moment I > > am not entirely convinced. What is strategic at a European level might > > not be so strategic locally, and vice versa. The European numbers are > > also not signposted everywhere, so there may be a challenge of > > verifiability. I believe the European definition basically defines the > > endpoints and a few waypoints, and it is left to national authorities to > > join the dots as they wish. So it may or may not achieve your goal of > > having a harmonised definition between countries. > > I agree with caution in trying to change anything. > > "Strategic" is an imprecise term. In the US, different people have > different opinions about which roads are important, depending on where > they live and where they want to drive. > > The tagging scheme is very much the UK system, and has adapations in > other countries. In the US, motorway/interstate is easy, and we more or > less have primary for US highways, secondary for state highways, and > tertiary for the next level of importance (reaching adjacent population > centers). > > In the US, "trunk" is fairly well defined. It's a road that is > substantially more than a regular highway in that it has some aspects of > a motorway. To be "motorway" (interstate class), the road needs to be > divided, multiple lanes, no stoplights, no at-grade intersections, with > controlled access. To be trunk, the road has to be part way to that > standard. So that means that almost all trunks are divided with > multiple lanes in each direction, but they typically have some > intersections (every few miles to maybe every mile), and may have some > but not a lot of non-ramp access. They often have narrower lanes than > Interstate specifications allow. > > Around me, the poster child for trunk is Route 2. It's the second most > important east-west road in Massachusetts,, and in many places has two > lanes each way, is divided, and has occasional farmstands and roads on > the edge, but fairly few. Lights (and one rotary) are at least a mile > apart, and sometimes 5ish miles apart. But, way out west, it is no > longer trunk - it's just ~Main street, undivided, one lane each way, > lights, houses. There it's highway=primary, because it's still a key > route, as important as a US highway. In some places it meets motorway > specs and is tagged as such. > > Calling a regular highway trunk, when it should be secondary or primary > would defeat the purpose of trunk, which is to identify roads that feel > intermediate between regular US highway and Interstate. We don't have > any formal designation between US highway and Interstate. > > In addition, there's history of people wanting to retag highways to > match their own view of these tags, and many others being unhappy about > this. This is hugely important in OSM, which is about a group of people > cooperating to improve the map. > > > Are you actually sure there is a problem to be solved? Do you have > > examples of inappropriate or inconsistent use of highway=trunk? > > That's a very good question. Certainly I find cases where roads are > over or under tagged, and as long as it's occasional, I just fix them to > match the norms of the larger surrounding area. This is very different > from trying to make large-scale changes in what the tags mean. > > Perhaps the OPs could rephrase the discussion in terms of what seems > wrong with actual tagging and how that impacts users of the database > (not any particular render :-). > > All that said, a problem in OSM is the blurring of road classification > and road characteristics. But that partially reflects a larger societal > blurring in terms of how people identify and use roads. It partially > reflects a choice made by renderers to emphasize classification. > Imagine a map where colors depend on whether the road is divided, how > many lanes it has, and intersection frequency, and aren't affected by > government labels! > > _______________________________________________ > talk mailing list > talk@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk > >
_______________________________________________ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk