I see your points, too, Minh. And BTW, “Local 11” is far from the only numbered local route in Santa Clara County. I’ve wiki’d dozens of them, although most don’t have signs with the route numbers, many more than 11 actually do. (Some of them, like in Sunnyvale, in a three-digit numbering protocol, as our wiki mentions).
Yes, Route 11 is the only signposted numbered route in San José. A few of the other cities post numbered routes too, but most of the numbered routes in Santa Clara County are from a long-gone planning document that turned out to be aspirational. I’ve been meaning to delete the numbers from the relations.
Its (San José’s / Santa Clara County’s) being-updated bicycle network numbering protocol is actually “being redesigned for the 2020s,” is how I’d put it.
About a year or two ago, there were open meetings soliciting input from the public. I had (have on an old spreadsheet) some sketched ideas that were a re-imagined superset of the now-mapped (and agreed, old, obsolete) numberings, but I didn’t submit it (my newer numbering protocol ideas) to the meetings.
Think more about how people navigated like two decades ago, without GPS. In those times you rather drove on state routes. Since it was much easier to follow that sign than checking in the map when to switch from a county road to the next. Nowadays, it doesn’t matter and you just follow your GPS. On a bike that’s still different. Those *cn are something similar to differentiate between a state route and a county route.
The overall issue is, that as a cyclist you want to cycle a nice route, but every cyclist has a different understanding of nice. 
Of course, routers try to find a nice route, and those routes are considered by a lot cyclist as nicer than other roads. I agree, for routers a differentiation between routes is not necessary.
Right, the state and national routes are for long-distance travel planning – and just as much for economic development purposes. @Richard has given us useful feedback about the challenges he faces trying to avoid sending cyclists into surprising situations. To me, this is a call to supplement route relations with something else like functional classification, even if for the most part it aligns to the route networks for now.
We’ve been discussing the USBRS a lot in this thread, but the ACA’s privately designated routes are more emblematic of these challenges. I think it’s fair to say that those routes serve a different audience. Unfortunately, we’ve forced data consumers to know the difference just based on the cycle_network=* value.
I think we are pretty close on the lcn and rcn but obviously there are different understandings about ncn and icn between Europe (40+ countries) and North America (3 countries)
Don’t get me wrong, not saying that the International Selkirk Loop not deserves a icn, just saying from European perspective, a route from LA to Seattle should be at least on a similar “level”.
That isn’t the fault of the hierarchy or our notation(s) in OSM, since “hierarchical notation” is exactly what *cn are. Any resulting friction / misunderstanding / uncomfortableness / this-is-different-from-what-I’m-used-to might be in the perspective of the beholder. Such cross-cultural differences happen, especially in an international context like a global map that purports to well-establish international, national, et cetera hierarchical relationships.
There are a number of examples of the vice versa (things that happen in Europe from the perspective of someone from the USA) which I find odd as I travel to Europe. I take it in stride: when in Rome, do as Romans do.
Really, this seems more and more of “hm, it would seem that certain ‘adjustments’ are (often) required when parsing bicycle routes between (North) American and European perspectives.” About ten years ago, I had @voschix over to my dining room table in Santa Cruz set up with a big desktop JOSM session. He does a lot of OSM bicycle work in Italy and Germany. We both learned a LOT from each other that day about how different our perspectives could be (and were and ARE!) about bicycle routing in OSM. It is fascinating how the differences are there, observable, explainable and ultimately, a version of “we do it like this, you do it like that.” A bit different, not terribly too difficult to understand once you DO understand.
I would rather suggest to separate signposted routes and “roads you want to send a cyclist on” as a router. A route takes a lot more into consideration than a router can do. It’s rather you plan a trip from LA to Seattle and you want to bypass,… Where a router tries just to find the “best” way from LA to Seattle. A router will never suggest you, hey, make a detour to see Mt Hood on the way 
To enhance routing, you need to add physical facts, like lanes, shoulders,… roads containing a route are just a good sign, some human made once the assumption, that road is better to cycle on than the nearby alternative.
Sharing a little DOT insight, these routes often form a “spine” network where bike/ped improvements can be prioritized. Still, many bike routes are nothing more than a sign (if that). A designated route could nonetheless be required for an improvement project to be funded.
And, we really should not confuse a designated route with the characteristics of its infrastructure. These are two different things, which can be analyzed independently. Richard knows this, it is part of the secret of success of cycle.travel. A route only means “a” route, not necessarily “your” (ideal) route.
1 Like
Why is that unfortunate, why is that insufficient?
Edit: I’ll add that USBRs being added to the System (and mirrored in OSM) have, can and do quite exactly follow ACA’s routes. Not always, but when they do, they often do so so substantially, they are nearly identical (USBR 66 in Kansas, both the intentions and actuality of the pending USBR 95 in California…).
A given cycle_network=US:US may or may not be signposted and may or may not come with infrastructure improvements, depending on what the current funding situation allows. But a given cycle_network=US:ACA is guaranteed to go unsignposted until it is replaced by a USBR. You’re also much less likely to run into a local along the route who knows about the route. To my knowledge, the ACA doesn’t go out blazing the routes as a hiking association would.
A reliance on published guidebooks and websites is nothing new when it comes to privately designated recreational routes, but I could definitely see how a data consumer, steeped in OSM’s traditional emphasis on on-the-ground verifiability, would find that surprising. Given all the times the word “quasi” has come up in conversations around these routes, wouldn’t some sort of tag specifically about that “quasi” characteristic be appropriate? At least unsigned=yes?
So what? (To the first two sentences). A USBR makes no claims whatsoever that it will “come with infrastructure improvements.” Why would anybody expect that? (I don’t). If I’m prepared to follow a (private, proprietary, copyrighted) route like one which might be tagged cycle_network=US:ACA (and I wouldn’t expect to find in OSM to begin with, as these violate our ODbL, even though some again-vestigial reasonings from deep in OSM’s past have allowed small segments of these private route data to slip into OSM and we strive to minimize these…) I would have ponied up the cash to ACA to purchase one of their pretty route guides for that particular ACA route. I don’t mean to strike a hostile tone as I ask “so what?”, but why would I expect to find private route data in OSM? And, there is no guarantee that any particular ACA route will (or won’t) turn into an USBR, although some rather substantially and pretty exactly have. No, ACA doesn’t blaze routes. They charge money for publishing them privately, so it makes no sense for them to blaze them. ACA’s routes are distinctly private routes, not quasi-private routes. They are quite different than routes which make their way into OSM, for reasons explained in our wiki.
I wouldn’t find this surprising in the least. I would find somebody getting surprised that ACA routes aren’t in OSM as uneducated about the difference between private, copyrighted bicycle route data and public (or publicly-accessible, essentially the same thing) bicycle route data. Or, perhaps she or he didn’t read our wiki about how these are fundamentally different things and putting copyrighted data into OSM isn’t allowed. This is an odd disconnect between us, Minh…I don’t understand it.
The word quasi- in front of “public” or “private” is defined in our wiki. Quasi-anything does not apply to ACA routes. OSM(-US) uses “quasi-national” to describe national-scope routes which have names or acronyms on the signs, distinct from actually-national bicycle routes (which are numbered USBRs). This has been clear to a lot of people for over a decade.
If it makes sense to you that segments of bicycle routes be tagged witih unsigned=yes I don’t have a problem with that (where true). But I find that an odd disconnect between us and strange direction for the conversation to have taken; OK, sure. The “quasi- characteristic” about routes can have all its questions answered by the wiki. But if any specific ones remain unanswered in your mind, I urge you to ask them directly here and I’ll do my best to answer it / them.
If / as / once the state of Oregon adds its “OC” (Oregon Coast) bicycle route to the collections of network=ncn, ref=95 components in our USBR 95 super relation, Relation: 95 (super) (7644263) | OpenStreetMap , USBR 95 will stretch from California’s coastal border with México (at Imperial Beach) all the way (continuously, only Oregon is now excluded) to Delta Junction, Alaska. (And you could continue northwesterly on USBR 87 as far north as Fairbanks). Yes, some of that (between Bellingham-Fairhaven, Washington and various ports-of-call in Alaska are via bicycle-ferry, that’s still (or will be, with Oregon’s inclusion) one of the longest bike routes on Earth. (I don’t know that, but I can look at a map and see it’s true). OK, the “Pendleton Gap” in San Diego County, California is a minor omission, alternatives exist.
Should we call this “international” because of such length? No: it is in the USA’s national numbered system of routes, connecting three (maybe four) US states.
The “level” that our routes are “in” (here, in USA / North America) has to do with whether they cross an international boundary (and these are / this is):
• international route(s) tagged network=icn,
• national routes (USBRs) and quasi-nationals (national in scope) tagged network=ncn,
• regional, often state-wide / administered at state-level routes tagged network=rcn, or
• local routes tagged network=lcn.
Key:cycle_network - OpenStreetMap Wiki 's United States section clarifies this as well.
This is not about copyright. This is about expectations, right or wrong, that people have when they encounter features with certain tags. I don’t see why cycle_network needs to be the beginning and end of characterizing a route. My hypothesis is that the nuances you’re describing are not well understood among data consumers. After all, rendering and routing software does not normally consult the wiki so that they could learn these nuances. A dedicated tag would help with discoverability.
As a mapper, I would like to clarify the situation through tagging. For example, this route relation across Kentucky calls itself the “Ramblin’ River Bike Tour”, supposedly part of the Kentucky state bike route network, sourced to a defunct page on KYTC’s site about the Underground Railroad Bike Tour, with a note that this route is not even a proposed route yet. A renderer would reasonably mark it with a green oval inscribed with “RRT”, based on the cycle_network and ref tags.
I remember when this relation first appeared in OpenCycleMap. As someone who used to live nearby and followed local trail development closely, I understood that this was merely a tour published somewhere and entered into OSM for convenience, not because there’s anything tangible about it. Eight years later, at least KYTC still pencils in the route as a potential corridor for a future USBR 35.
I get that these things take time, but in the meantime, shouldn’t data consumers have some way to tell that this route is more tentative than others in this network or similar cycle_networks? Shouldn’t a cyclist be informed of a caveat emptor that they shouldn’t expect to encounter any wayfinding aids or locals who have heard of the route along the way? Or do we hope that, by forcing routers to present this route as a faît accompli, the route will be that much more likely to become reality?
Sure, Minh: data consumers can, should (and do) expect a lot. Let’s give them what they want and need. Kentucky’s RRT is a good example, but how can we improve it? The wiki really IS a good start, but I hear you loud and clear that our data should reflect “something different” about things like “corridors” (which are not routes, but are “could be a miles-wide area where a future route might be developed”). I’ll say one of the smartest things we ever did (circa 2011-12) regarding bicycle routing and how the USBRS was newly re-emerging at that time was eliminate any sense of “corridor” entering OSM at all.
As a cyclist, and one who seeks routes, I’d really like to know whether I’m viewing, for example, a corridor or an actual route. An actual route has had real infrastructure vetted and certified to be “suitable for cyclists” perhaps with a lean-in-a-particular direction denoted about it, like “better for experienced, adult, long-distance-oriented riders.”
It sounds like you might be itching for a (newer) scheme that describes these attributes of actual or potential routes, better than the conventions which have emerged over the last couple of decades in our project. Once again, I’m all ears and listening.
In the meantime, we build data in our map database and we wiki our way to both document what is (being descriptive) and sometimes (when explicitly stated that this is what we’re doing), we sprinkle a little bit of prescriptive into our wiki. (Here’s how we SHOULD be tagging).
There’s room for plenty of such clarification in OSM. Sometimes in our data, sometimes in our wiki, sometimes amongst ourselves in good discussion, sometimes all of the above.
If I understand correctly, many of these named routes started out as actual tours that someone once published online. Maybe people even used to follow these routes at some point. Some of them follow common-sense routes that may well become USBRs without modification at some point in the future. But I don’t think OSM should be in the business of speculation about the future.
This “Ramblin’ River” tour follows a road in Northern Kentucky that used to ramble right into the river below after every major storm, forcing KYTC to more or less abandon the road for a few years. When I closed off the road and detoured Kentucky Route 8 around the area, I didn’t quite know what to do with this bike route. (Currently, the road is currently in a somewhat better state of repair as Kentucky Route 6335, the highest number in the state – in other words, the least important route in the state.)
As far as I can tell, no relevant agency or organization promotes this tour as a specific route anymore, not even the local tourism boards. It only has any notability as a corridor for a future USBR. Should we delete the route relation to avoid further confusion? And does this situation inform our next steps regarding the original topic, the Pacific Coast route?
I’d certainly consider deprecating this Kentucky “route.” If it is something that KYTC truly insists (at a statewide, DOT-level) “this is a bike route, suitable for cyclists to traverse its entirety, from one end to the other, by bicycle” then I’d say let’s keep it in OSM. If it is (as some might argue) a vestigial planning corridor, influenced by ACA’s routes (UGRR, others) in the area, well, there are good arguments for deletion / deprecation.
I don’t like seeing wholesale deletions, so at least squirrel away a .osm or .gpx version of the route data if you DO delete it, but it is good discussion for us to determine, “y’know, this doesn’t really belong in OSM.” Maybe it does, maybe it doesn’t. If it goes, squirrel away a copy (or mention I might like to before it is deleted…I sorta “collect” such things), update its mention in our USBRS wiki and wave goodbye.
We’re on the right track here: better harmony amongst routes (and how they fit together into their “actual-ness” in the real world, and whether they belong here or there in a hierarchy…) as well as wider understanding of what our data actually mean. We’re OK!
What does this mean for PCB in California as USBR 95 Phase II might be Approved? I think something along the lines (as I originally specified) that it is vestigial, and so its superseding data (USBR 95) might cause those overlapping elements of PCB(R) to be removed from OSM. I imagine heads continue to nod about this, though I could be wrong.
At most we would delete the route relation, not redact it, so this Overpass query will always return the current route geometry and tags.
Cool. Helpful. Good one. Thanks.
Now, to delete or not to delete? And similar, the original question: do we delete vestigial PCBR elements after 95 “supersedes” them? In the case of 95/PCBR, I think we do.