*Downgraded Griffith Park Dr. to unclassified, got cut off

I posted my comment in the other changeset coming from a place of assuming good faith, in the hopes of convincing you of my (and a broad consensus of mappers') opinion. However, calling those who legitimately disagree with you on a tagging matter vandals and trolls is no way to go about a respectful argument, and it goes against OSM's community principles, indeed against the principles of any community effort.

To bring in an example from another part of the state that I think has some parallels, I was initially uncomfortable with proposing CA 1 and US 101 through the city of San Francisco as trunk. I lived in San Francisco for 18 years, and had taken both these routes myself many times. Their built characteristics match other primary roads in the area, as does their use by many SF residents: as the best routes between neighborhoods and from neighborhoods to the freeways, good primary uses in urban areas. But these really are the best routes to go from the Peninsula and points south to Marin and points north, that is, they provide regional connectivity, and thus meet the trunk definition. By not indicating this, the OSM map contains less useful information than it could if these routes were trunk, but it's something that was easy to miss since I was so familiar with the local area. I'll also add for the road at hand specifically, though I know Adamant has consented to CA 299 being trunk through town at this point so I don't want to belabor the point, that by the "more objective" regional designations (NHS, Caltrans FCs, the Strategic Interregional Corridors), CA 299 through Redding to CA 44 meets the trunk route-equivalent standard. Again, to reiterate I don't think any of these are authoritative, but they do qualify as some evidence in support of the road's regional importance.

I had refrained from commenting much on this, as I feel like the acrimony was about more than a road classification. But since the merits of the whole enterprise have now come up, I do want to describe why I think this trunk classification project is useful and an improvement of OSM data.

I want to start by saying to Adamant that I do think local perspective is valuable and important, and I respect your knowledge of the area. For instance, after reading your descriptions of local circulation patterns (and judging for myself), I agree that Market St/CA 273 is best classified as primary, not trunk, and I don't think many disagree. But I also think it's important to come up with a consistent, nationwide way of tagging roads, and to do this we need to incorporate the perspectives of those outside our local regions. Otherwise, our data is useless to consumers, be they routing engines or just people looking at the map. So I don't think anyone has "authority" over the map, and of course no one ever can: the map is editable by anyone. All we can do is work to establish uniform practices across the country/world and build consensus for them.

As to the merits of this approach specifically, I think there's some validity to the assertion that there exists some vagueness in the proposal's wording, making local verification difficult. I think a lot of people participating in the project (certainly me) would prefer a single, authoritative source to base our road class tagging on. Unfortunately, no single criterion exists in the US which matches OSM's road class definitions, which are based on official UK classifications. A number have been tried (traffic counts, number of lanes, NHS classification, Caltrans functional classes, membership in Caltrans' "Strategic Interregional Corridors" to name a few), but none were found to perfectly reflect OSM's definitions (which are a little vague themselves!). One particular difficulty, as Adamant has alluded to, is that roads in different regions are built differently. So road classes have always had a subjective component, which of course leads to disagreements.

In the UK, trunk roads are classed by their importance to the regional road network, i.e. how they link cities of regional importance. So the approach here, matching the approach in dozens of other states, is to decide which cities are of "regional importance", and decide what the best routes between them are. Certainly, there is room for disagreement here, and indeed there has been. But one principle that has achieved broad consensus across the country is that roads do not lose their regional importance when they enter urban areas, that the best regional route is still the best, even if it's heavily trafficked by locals as well. Another related idea is that network connectivity is important, that there should be no dead ends in the trunk network except in rare cases of terminal geography. So this is why even if a road may have similar characteristics to other primary routes on the ground, its regional importance makes trunk a classification that contains more information for map users in my opinion. This also matches how many other US maps, paper and online, map regional routes through cities.

It sounds like things are already getting ironed out in a constructive way, but as one of the other authors of the California guidance I just wanted to reach out. Thank you for providing feedback Adamant and stevea, I definitely understand being wary of major edits in areas you've worked hard to curate and improve. I think in general, how to connect the ends of routes to and through the urbanized areas they connect is one of the thornier issues with this highway classification project, as it can lead to up-classifications of roads due to their regional importance, even if they don't seem much more locally important than surrounding primaries. I hope we can begin/continue collaborating to make sensible classifications that allow us to match both the spirit of a consistent, connected state-wide trunk network and the reality of on-the-ground experience. I'll let others speak to the specifics of this tagging, as I don't have much to add that hasn't been addressed (I've been up this way a few times, but it's not an area of personal expertise)

I've only hiked it, and I remember it being fairly narrow, and so I think changing the highway= tag from cycleway to footway is a good choice. What I was saying is that I think you also changed the surface= tag on two of the segments from "concrete" and "asphalt" to "unpaved", and was wondering if that was based on a more recent survey or not.

Hey MrDan, I noticed you edited the LMU trail. Is the whole thing really unpaved? I hiked this a few months ago and seem to remember most of it being asphalt, with some unpaved segments near the ends, which is how I tagged it. Note that there are two trails, an upper one and a lower one, and I believe the lower one is dirt. I have no argument for or against its suitability for bikes.

Hi! I noticed you split PCH into a dual carriageway in Santa Monica. From my memory and from mapillary street-level imagery, it looks to me like the two directions are separated only by a center turn lane in this area, with no physical separation (like the posts in other parts of the road). This means the road should probably be mapped as a single way: osm.wiki/Dual_carriageway. This is important because if it's mapped as a dual carriageway, it implies that you physically cannot turn across the road to get to houses and parking lots, but I think you can. Is there now physical separation that I'm not aware of?

Best,
Will

Hi Jason, thanks for editing OSM. I noticed you're modeling intersections. Typically in OSM, intersections involving dual-carriageway roads are modeled as keeping the dual-carriageway through the intersection, rather than squeezing it to one segment at each intersection, even if the physical separation briefly disappears. This site has some helpful illustrations of typical practices https://labs.mapbox.com/mapping/mapping-for-navigation/modeling-intersections-for-map-navigation/.

Best,
Will

Hello Florian, welcome to OSM! I see you're adding detail to some golf courses, which is welcome. However, I am not sure you are following typical OSM practices for naming. I believe holes are typically mapped as ways along the shortest playing path, with a ref=1 for the first hole, for example. Names are usually omitted from the actual green and fairway areas. I encourage you to take a look at the documentation for golf holes at golf=hole, or at nearby Rancho Park Golf Course, which I believe is mapped more similarly to what's laid out in the wiki.
Happy mapping!
Will

Hello,
I noticed you recently made a series of changes similar to this one in Culver City. I am not sure whether naming these nodes is appropriate tagging. What do these numbers represent? Addresses may be better suited for stand-alone nodes, or attached to their corresponding buildings, as addr:housenumber.

Hey AV Safe Streets,
I noticed there was some back and forth on this highway designation. Looking online, I don't see any sources that CA 122 is an official state route, and I wasn't able to find any evidence on street-level imagery that it's signed on the ground either. Are you sure this is a real state highway? I'd appreciate it if you pointed me toward the evidence you had for adding it.
Best,
Will

What is your justification for tagging this as meadow? Most of this park is covered in scrub, and it certainly is not used for grazing. To me it appears you're tagging for the renderer. Your changesets also appear to have broken the tagging of the state park.

Hi, please take care when making these edits, as you appear to have broken the Ackerman union building with this changeset. Also, your source map link does not work, is this source compliant with OSM's license?

Hi, and thanks for improving OpenStreetMap! These are apartments rather than houses. One tagging recommendation: for an apartment building, even if it's part of a larger complex, you want to tag it building=apartments. If the entire complex has a name, you can draw a separate area around the complex and tag that landuse=residential and residential=apartments. It's a little confusing, but we want to keep buildings tagged as buildings. I'm going to change the buildings to building=apartments, but feel free to let me know if you have any questions!

What is your justification for upgrading this segment? These are normal city roads, with 35 mph speed limits, and La Cienega distinctly becomes an expressway (higher speed limit, separated, limited-access) south of Obama.

I think that'd be great!

I agree that leisure=park was not correct for Griffith Park, as it is not manicured and so boundary=protected_area is better. However, I don't think tagging the whole park as landuse=forest is the best approach, as it implies the whole park is covered by trees and perhaps used for timber (per osm.wiki/Forest), whereas it's mostly chaparral/scrubs and man-made features like golf courses. After some discussion on the OSMUS slack #tagging channel, I think leisure=nature_reserve is more appropriate, given the definition in osm.wiki/United_States/Public_lands#General_Tagging_Guidelines, so I'm going to change the overall park to that, and try to tag more specific landcovers (natural=scrub, natural=wood, etc) separately at some point. But happy to discuss further if you disagree!

I don't think this change is correct. Based on the surrounding streets, all WWI battles, this street is almost certainly "Marne" as it was before, not "Marine".

Hi! I've been doing some mapping in this area. Could you explain your rationale for changing 842458772 from amenity=school to boundary=school? I'm not sure the boundary tag is appropriate, its only two other uses in CA are both for school districts, not school property.