Re: [Talk-us] Trunk VS primary,
Eric H. Christensen wrote: > The routing engine should be able to take into account > the road surface It can and often does. Your problem there is that only 2% of highway= ways in the US are explicitly tagged with surface; probably only 30% are implicitly tagged; and sometimes the implicit stuff gets broken, like when people start retagging gravel roads as secondary without adding a surface tag. (Numbers are estimates but I think not far off.) > Any idea why trunk was established in the first place? It's a word from the UK road classification system, because OSM was invented in the UK. But the letters in the word aren't really important. OSM has five broad-brush motor-road tags (trunk, primary, secondary, tertiary, unclassified), plus special-case ones at either end of the hierarchy (motorway for limited-access high-speed roads, residential for roads with the main purpose of providing access to houses on that road). If you don't think you need five, you don't need to use all five. If you need more than five, you are free to use additional tags to supply extra nuance, as the Germans do with motorroad=yes. I would say that 15 years is probably more than enough time to decide what roads you're putting in what category, but hey, this is OSM. Richard -- Sent from: http://gis.19327.n8.nabble.com/USA-f5284732.html ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
20 Dec 2019, 01:25 by ba...@ursamundi.org: > So, for example, in the US, instead of motorway, trunk, primary, secondary, > tertiary, perhaps something more like freeway, expressway, > major/minor_principal (just having this would fix a *lot* of problems with > Texas and Missouri and their extensive secondary systems), > major/minor_collector...the US just has a way more complex view of how > highways work. > > Or at least some more serious consideration given to the proposal at > > https://wiki.openstreetmap.org/wiki/User:UltimateRiff/HFCS> (but perhaps > with "other principal arterials" as primary and a new "highway=quartinary". > Fitting thing like road classification into UK system is irritating at times. But idea of each country with separate tags for roads is simply a bad idea. This info is probably worth recording, but legal status should go into a separate tag. ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
On Thu, Dec 19, 2019 at 1:19 PM Martijn van Exel wrote: > I actually like your suggestion that highway=trunk does not add much value > to the U.S. map, Eric. > We love to add detail / granularity to OSM so much, it can become hard to > envisage taking some away. > Not saying we should abolish trunk right here and now, but something I'd > consider as one outcome. > I'd like to see a lot more left up to the data consumer and more regional values to be widely acceptable. For example, instead of trying to smash the entire planet into the UK's prescribed values and trying to come up with equivalences, use the terminology each country uses. So, for example, in the US, instead of motorway, trunk, primary, secondary, tertiary, perhaps something more like freeway, expressway, major/minor_principal (just having this would fix a *lot* of problems with Texas and Missouri and their extensive secondary systems), major/minor_collector...the US just has a way more complex view of how highways work. Or at least some more serious consideration given to the proposal at https://wiki.openstreetmap.org/wiki/User:UltimateRiff/HFCS (but perhaps with "other principal arterials" as primary and a new "highway=quartinary". Much like moving route refs to highway relations (freeing the ref=* tag on highways for situations where the road and the route have different refs), leaving the mental gymnastics up to an algorithm and leaving less confusion to the mapper is getting to be long overdue. ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
On Thu, Dec 19, 2019 at 5:13 AM Mike N wrote: > On 12/17/2019 10:19 PM, Evin Fairchild wrote: > > some US routes are more important than others and lumping them all as > > primary doesn???t make any sense; > > The arguments here about relative importance of parallel routes makes > sense. > >Some massive changes such as in > https://www.openstreetmap.org/changeset/78620805 are raising roads which > have no other major choices, but are apparently just because they are > the most important. > This smashing everything to the highest possible value I would generally consider to be an undiscussed and problematic mechanical edit. Going with the lowest level that fits feels a bit more correct (think "minimum effective dose" from medicine, for example), does give routers more information where there's lots of routes available, and humans more of an idea what kind of road they're going to encounter at a glance. ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary
I now reiterate the fundamental struggle in this discussion (which can be summed up as "both"): highway=trunk is another level of granularity (above primary) to describe "high performance OR high importance roads" (emphasis mine). Additionally, (from the US-specific definition from our wiki): highway=trunk is a "surface expressway: a relatively high-speed divided road (at least 40 MPH with a barrier or median separating each direction of traffic), with a limited amount of intersections and driveways; or a major intercity highway." Similar to "descriptive vs. prescriptive," this semantic "struggle" might be described as these two definitions being "relative vs. absolute." Some people say a gravel road (or even a dirt road, if that dominates, say, in a developing country) is important enough to be tagged "trunk," for example in Alaska. That is using "trunk" in its relative sense: relative as to what is also meant by primary, secondary, etc. IN THAT LOCAL/REGIONAL CONTEXT. Some people say "trunk must be divided with a barrier/median and medium-to-higher speed..." (or fill in some hand-waving additions). That is using "trunk" in its absolute sense. "Both." Yes, both. I'll say it again, OSM, I doubt it very much, will ever, EVER get away from how we now define "trunk" as "both." Look at our wiki and see how there are differing definitions for differing countries and see how we define it relatively. Look at our wiki's text and see how we define it absolutely. Both. Can we (as humans) and routers (as software) learn to live with this apparent dichotomy? I can. I believe the rest of us (humans and routers alike) can, too. SteveA California ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 ‐‐‐ Original Message ‐‐‐ On Thursday, December 19, 2019 2:19 PM, Martijn van Exel wrote: > I actually like your suggestion that highway=trunk does not add much value to > the U.S. map, Eric. > We love to add detail / granularity to OSM so much, it can become hard to > envisage taking some away. > Not saying we should abolish trunk right here and now, but something I'd > consider as one outcome. It does seem like there is a lot of arbitrary conditions separating some of these road types. It would be nice to have solid reasoning for tagging a roadway a certain way instead of how "important" it is. The routing engine should be able to take into account the road surface, number of stop lights, and other factors into consideration if there are multiple routes of similar highways that are "important", IMO. Any idea why trunk was established in the first place? --Sparks -BEGIN PGP SIGNATURE- Version: ProtonMail wsFcBAEBCAAGBQJd+93IAAoJEIB2q94CS7PRz34QALpd+bUfV2YniWX9qVtb nzFxEsdPp8dzt/s4HewS93wDtSrVKM7rtuPLTnDetFrI4Yl5bZ3JfkrDRfDR KkxYbgU1/Fjrjz/MCJ+aoDh0SwKVwSUDMpX9nJtNxNHfaqjxu6Z86eKLcBX1 ceAUnTMoUQfDnWSH3TUPadQOkwij1r0Ja2EreKuSJwMk+geGs8qlVU0sNf0x 3KooDARDgIvTbul1344kt0QceK42py0P4RjqRsFJnjg9cCGa07BD1pb6R+qk 2xWL9Q0cCnbF/mhCQ1LoQvIzdgQl/4jvccQ7vyBwgkqbg1f63X0dDbQMPT6F mRtZ6almPUg0XP0jO33s7JKruMuO4OaJhMYY7h4N1EscjiDcm1zXzz5OCsXP 2QEKnSDkP6h0DAGL58A3rmI0c8z9qAUT+kVrLU1nJDb4WhUsNcpch87Rswi3 w9VFKBEAYVuD9EwJN37e8xG21SO289alvb1hVGzI914aovSbOElluDrQ3F8p yyD8d7m2wEebEjNzaKRHXVTC1bsiBzruZiUZt9K69Cxv5m3YJHW7giPagKEW V/o6kwBJOrXnwtF9U/AZCA+EhPF7Myb2tfMdPsWzr0kq4Yy0vfb9V7Y7Ldnw MlTc8QtIRUfjejWGV2mXDPxghGQ50YGnipG4NjVuSOhhs/yO3ZDOB0pDorxF VtwA =uMfV -END PGP SIGNATURE- ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
I actually like your suggestion that highway=trunk does not add much value to the U.S. map, Eric. We love to add detail / granularity to OSM so much, it can become hard to envisage taking some away. Not saying we should abolish trunk right here and now, but something I'd consider as one outcome. Martijn On Thu, Dec 19, 2019 at 7:27 AM Eric Ladner wrote: > I personally dislike "trunk". Its definition is vague and leaves a lot to > interpretation (and argument). It doesn't really add anything to the > information on the map, IMO. A US Highway is a US Highway regardless of > how much traffic it carries or how many stoplights it has. > > Maybe if the definition of "trunk" was solidified to something more > specific, it would have a more valuable use case. > > On Thu, Dec 19, 2019 at 5:15 AM Mike N wrote: > >> On 12/17/2019 10:19 PM, Evin Fairchild wrote: >> > some US routes are more important than others and lumping them all as >> > primary doesn???t make any sense; >> >> The arguments here about relative importance of parallel routes makes >> sense. >> >>Some massive changes such as in >> https://www.openstreetmap.org/changeset/78620805 are raising roads which >> have no other major choices, but are apparently just because they are >> the most important. >> >> ___ >> Talk-us mailing list >> Talk-us@openstreetmap.org >> https://lists.openstreetmap.org/listinfo/talk-us >> > > > -- > Eric Ladner > ___ > Talk-us mailing list > Talk-us@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk-us > ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
I personally dislike "trunk". Its definition is vague and leaves a lot to interpretation (and argument). It doesn't really add anything to the information on the map, IMO. A US Highway is a US Highway regardless of how much traffic it carries or how many stoplights it has. Maybe if the definition of "trunk" was solidified to something more specific, it would have a more valuable use case. On Thu, Dec 19, 2019 at 5:15 AM Mike N wrote: > On 12/17/2019 10:19 PM, Evin Fairchild wrote: > > some US routes are more important than others and lumping them all as > > primary doesn???t make any sense; > > The arguments here about relative importance of parallel routes makes > sense. > >Some massive changes such as in > https://www.openstreetmap.org/changeset/78620805 are raising roads which > have no other major choices, but are apparently just because they are > the most important. > > ___ > Talk-us mailing list > Talk-us@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk-us > -- Eric Ladner ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
On 12/17/2019 10:19 PM, Evin Fairchild wrote: some US routes are more important than others and lumping them all as primary doesn???t make any sense; The arguments here about relative importance of parallel routes makes sense. Some massive changes such as in https://www.openstreetmap.org/changeset/78620805 are raising roads which have no other major choices, but are apparently just because they are the most important. ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
Okay, this is going to be a long message, but I’d strongly suggest you read ALL of it before responding. I’d first like to address the assumption that some people seem to have that those who support using trunk on roads other than divided highways are “tagging for the renderer” because we allegedly just want to see these roads appear at lower zoom levels. For me, that is NOT the case at all. In reality, I support having trunk for major intercity highways because there needs to be more levels of indicating importance for US highways than just primary. For example, in Nevada, US 50 and US 6 are very lightly-traveled roads that only connect a few small towns in Nevada, and are known for having very little traffic. Tagging them as primary is perfectly fine IMO. However, US 95 connects the two biggest cities in Nevada—Las Vegas and Reno—but it’s also tagged as primary. Surely US 95 between Vegas and Reno is more important than US 6 and US 50, right? In my home state of Washington, there are two US routes that cross the Cascade Mountains, US 12 and US 2, both currently tagged as primary. Both of these roads are kept open in winter thru the Cascades. However, there is another road that crosses the Cascades, WA 20, that is also currently tagged as primary (which makes sense given that it is a very important cross-state highway), but it is NOT kept open in the winter. It doesn’t make any sense that a road that is not open in the winter is tagged at the same importance level as other roads that are kept open in the winter! Secondly, I think some things from the wiki need to be pointed out here. On the wiki page for Key:highway, [1] the definition of highway=trunk is “The most important roads in a country's system that aren't motorways. (Need not necessarily be a divided highway.)” Those who say “Trunk roads should ONLY be divided highways, no ifs, ands, or buts” are going against what is explicitly stated on the wiki page for key:highway. Also, at the bottom of the aforementioned wiki page, there is a section entitled "Assumptions,” which states in the first paragraph: “Only highway=motorway/motorway_link implies anything about quality. Other road types, from highway=trunk through highway=tertiary to highway=residential=residential/service or highway=path/footway/cycleway/track do not imply anything about road quality.” These words speak for themselves. Now, if we want to indicate road quality in some way (e.g. whether a road is divided or not), we ought to use the expressway=* tag like others have suggested rather than using the highway=trunk tag just for that. Even if you don’t use the expressway tag, you can still tell if a trunk road is divided or not because the default render shows divided roads as having a thicker line than undivided roads. A good example can be seen by looking at western Canada, where the most important intercity roads are tagged as trunk regardless of whether they’re divided or not. [3] You can clearly tell if a road is divided or not even if undivided roads are tagged as trunk because the divided roads have a thicker line than the undivided. Therefore, it is not necessary to reserve highway=trunk for divided roads only. TL;DR I support tagging undivided roads as trunk because 1) some US routes are more important than others and lumping them all as primary doesn’t make any sense; 2) the wiki says that only the motorway designation implies anything about quality of the road; and 3) the renderer shows divided roads with a thicker line than undivided roads. [1] https://wiki.openstreetmap.org/wiki/Key:highway#Roads [2] https://wiki.openstreetmap.org/wiki/Key:highway#Assumptions [3] https://www.openstreetmap.org/#map=7/51.618/-112.972 From: Greg Troxel Sent: Tuesday, December 17, 2019 1:16 PM To: Paul Johnson Cc: Mike N; talk-us@openstreetmap.org Subject: Re: [Talk-us] Trunk VS primary, Paul Johnson writes: > On Tue, Dec 17, 2019 at 7:24 AM Mike N wrote: > >> >>I think many of the trunk VS motorway VS primary conflicts come from >> 2 points of view: on the one hand, people like to zoom out and see a >> coherent network of interconnected roads. > > In which case, rendering based on network on the route relations would be > more appropriate. This is the crux of the matter. Calling things trunk so they render is tagging for the renderer in a bad way. ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
Paul Johnson writes: > On Tue, Dec 17, 2019 at 7:24 AM Mike N wrote: > >> >>I think many of the trunk VS motorway VS primary conflicts come from >> 2 points of view: on the one hand, people like to zoom out and see a >> coherent network of interconnected roads. > > In which case, rendering based on network on the route relations would be > more appropriate. This is the crux of the matter. Calling things trunk so they render is tagging for the renderer in a bad way. ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
Re: [Talk-us] Trunk VS primary,
On Tue, Dec 17, 2019 at 7:24 AM Mike N wrote: > >I think many of the trunk VS motorway VS primary conflicts come from > 2 points of view: on the one hand, people like to zoom out and see a > coherent network of interconnected roads. In which case, rendering based on network on the route relations would be more appropriate. >In the end, this would suffer from the same connectivity issue: > should the US highway remain a trunk as it reduces to 2 lanes and drops > to 30mph passing through a tourist area? Would that tend to draw GPS > navigation routes from nearby faster, parallel streets? No. What usually causes this is a regional speed limit where the local speed is not yet known to OSM and/or priority signage hasn't been mapped yet that obviate staying on the highway as the best route to the renderer based on ground truth. > Or would it > look like an ugly gap in the trunk road if it switched to primary in > that tourist area? > Depends on if you're rendering based on class or based on network. >As an aside, I sense that the tendency to upgrade results in all OSM > streets being promoted by one level, resulting in a compression at the > top end and less class distinction at those levels. > This tends to be the case. Seems like based on the AK2 conversation, this is a prolific problem in northern Canada, where roads are uncommon in general. ___ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
