Vestigial "Pacific Coast Bicycle Route"

I’m being careful here to thread many needles simultaneously, as the plot is historical, thick and somewhat complex.

California, through its Department of Transportation Caltrans, has recently applied to AASHTO to essentially complete (Phase II after Phase I already approved from Oregon to Daly City, south of San Francisco) United States Bicycle Route 95, extending from Daly City / Pacifica to its border with México (at Imperial Beach).

There is at least one long-existent-in-OSM “Pacific Coast Bicycle Routes” (PCBRs) which are either vestigial or proprietary, which USBR 95 in California is deliberately intended (as stated by its architect) to replace (USBR 95 replaces the proprietary ACA version, AND the “devolved-from-Caltrans-to-local-jurisdictions” version). One of these is the Adventure Cycling Association so-named route (private, deliberately not entered into OSM as it is proprietary and copyrighted, which would be in violation of OSM’s ODbL). The other is a route which used to be administered by Caltrans, but has devolved into being administered by counties and cities which were included along the route. These numerous belts and spurs along the route might remain in OSM as (purple) regional route segments, rather than the (red) “national trunk” of USBR 95. I do not say that as “tagging for the renderer,” this is beyond that.

In effect, this devolved route (not the ACA proprietary route, as USBR 95 intends to replace that) is little more than vestigial, having been essentially replaced by USBR 95 (California’s Phase II version now submitted to AASHTO). In OSM, there are northbound (https://osm.org/relation/7063452) and southbound (https://www.osm.org/relation/53722) relations as well as a super-relation (Relation: ‪Pacific Coast Bike Route‬ (‪7063453‬) | OpenStreetMap) tying these together.

What I propose is to simply delete the members of the now-vestigial relation members now overlapping with USBR 95 (after it is Approved by AASHTO, of course). This seems “cleanest,” though I welcome perspectives there may be merit in doing something of an OpenHistoricalMap (OHM) interpretation method of tagging PCBR as it fades away (is replaced by USBR 95 in California from Pacifica to México).

Really, I’m simply opening a discussion. I think what will end up happening is that “bits of purple belt and spurs called PCB(R) along USBR 95 in California will sometimes show up as locally-devolved-and-administrated components of a vestigial route glommed onto the newer USBR 95, which replaced this.”

I am in listening mode. I’m not terribly familiar with OHM principles and tagging. It seems both easier and correct to simply “duck out” those elements in the purple route (network=rcn) which are being “replaced” (superseded) by the USBR 95 red route (network=ncn). Thanks for any feedback offered.

Edit: An additional step which may make thing easier is this. Vestigial ACA routes are now (since 2012) tagged network=rcn. We might “further demote” the vestigial PCB segments to lcn (as “low as they go”), as they (the “devolved” segments) are (to the extent they ACTUALLY are!) administered by counties or cities…“lower” in the administrative hierarchy. That is, compared to international, national and regional routes. It all makes sense, really; demoting from purple (mid) to blue (low) while red (high, national) replaces (supersedes) allows heads to further nod. Feedback welcome.

Would it be fair to say that the Pacific Coast route is just a common naming practice at this point? It would be analogous to how plenty of streets along the historic El Camino Real retain that name, without necessarily implying that a coherent route remains. If so, then perhaps OSM shouldn’t maintain a route relation for the Pacific Coast route, as relations aren’t categories and routes aren’t merely collections of similar things.

As far as I know, there isn’t a formal process by which a numbered route “replaces” a named route. The two can coexist with substantial concurrencies. If the name falls out of use due to the establishment of the numbered route, then the named route’s relation can be repurposed or, more correctly, deleted. But as El Camino Real demonstrates, a name can persist independently of a named route.

Consider what happens when a building is stormed down and replaced by another. Sometimes a mapper will reuse the preexisting way to represent the new building, adding some vertices and deleting others to achieve the desired shape, but this reuse is for the mapper’s convenience only. They aren’t necessarily implying that the old building has been meticulously transformed into the new one through adaptive reuse.

I think there’s a wide recognition that the three-letter acronym values of network=* are a flawed tagging scheme rooted in an aversion to typing. The international community worked around the ambiguity by backronyming lcn to mean “local cycle network” rather than “London Cycle Network”, and the U.S. community worked around the oversimplicity with cycle_network=*. With a bit more foresight, maybe we would’ve chosen a workaround that can accommodate walking trails as well.

Ideally, we would’ve recognized much earlier that a given trail segment’s importance doesn’t necessarily correspond to the kind of administrative area it traverses, and that a given route’s importance doesn’t necessarily correspond to level of government that designates it. network=*cn makes both poor assumptions simultaneously.

We should’ve decoupled trail functional classification from route designation a long time ago, just as we did for highway routes. It may not be too late, if we think of network=*cn as a moribund tagging scheme along with crossing=* on crossings or ref=* on roadways. But any change would be aspirational unless we can convince the established cycling renderers to go along with it.

Which brings us to OpenHistoricalMap. Nothing is stopping us from mapping the former state-coordinated Pacific Coast route along trails and roadways, other than having to map those trails and roadways first. I think you’d be in a better position than most to map these features, because of your substantial knowledge of their historical development. As you map the ways, you can begin to cobble together the associated route relations. This discussion about evolving roadways may help you settle on a workable modeling strategy:

Since we know network=*cn is so problematic in OSM, we might as well avoid it in OHM in favor of some other approach, such as a hierarchical network=* format that resembles cycle_network=*. Nothing renders recreational route relations as in OSM, so we have plenty of time to figure out a solution for functional classification that looks very different than OSM’s status quo.

Thank you for your as-usual-very-thoughtful reply, Minh.

I now ponder much and may have more to say here in a day or two.

I would suggest to create a new relation for USBR95, without involving the existing PCB-relations, since the USBR95 isn’t liken a new version of the PCB.

Not sure whether it’s useful, but I’m thinking it might be better to transfer the PCB into a network=US:ACA or something similar. If they remain in the *cn schema, they supposed to be rather ncn, since they are on national scale, rather than regional or local. At least based on European usage of that tag. If this data is illegal or “outdated”, we should remove these relations from OSM.

If you need support, adding the additional parts of the USBR95, let me know.

1 Like

Henning, the new relations for USBR 95 are denoted in United States Bicycle Route System - OpenStreetMap Wiki , the Proposed section (95 is near the end of the table).

As I said, there are TWO routes named PCB(R): the “vestigial” route which is what is purported to be in OSM. This is the “devolved from Caltrans to counties and cities” version. There is ALSO a version of PCB(R) which is ACA’s (copyrighted, proprietary) route, which is exactly what USBR 95 is specifically designed to “replace” (according to numerous discussions I have had with the System’s decades-long architect). Not only would it be inappropriate to specifically enter the ACA route (OSM does not have authorization to do this and their entry would be a violation of our ODbL), but, again, USBR 95 is designed to be a 1-for-1 replacement of ACA’s PCBR.

While there are some historical segments of other ACA routes (Underground Railroad and Transamerica Trail), it was worked out between OSM and ACA in the early twenty-teens that it was OK for these (older, possibly not even correct to today’s routings) to remain in OSM as REGIONAL (network=rcn) routes, not national (ncn) routes, entered state-at-a-time to further emphasize they are not meant to be “at national scale” (OSM in the USA has several “quasi-national” bike routes for these; see our wiki’s Talk pages regarding some of this history). These ACA routes have been (deliberately) diminishing in OSM as more and more of these segments become directly incorporated into OSM as USBRs, for example the entirety of USBR 76 in Kansas was originally a component of ACA’s Transamerica Trail data which virtually 100% was re-entered as USBR 76 when AASHTO Approved this route in Kansas.

As Minh makes some excellent points above, and I still ponder what best effects might be enjoyed by OSM data consumers as eventual solutions, it still remains not-yet-clear to me what the best course of action forward might be. The three levels of national, regional and local have been taken about as far as they can (or beyond, to the level of abuse and/or confusion), with quite a great deal of pulling and stretching their semantics over the years, such that they are “less meaningful” than they might once have been. Hence, the US (and other parts of the world) have started to alleviate this with cycle_network=US:* tagging.

It is also why I suggested in my last-minute edit to my original post that the vestigial components of PCB(R) which remain as minor belts-and-spurs hanging off of USBR 95 (as it is Approved) be changed from their present network=rcn tagging (which seems to have been assigned as a misunderstanding that this route represented the ACA route, which it does not) to network=lcn tagging, to better emphasize that these “vestigial route remnants” are “locally” administered (by counties and cities).

I’ll keep pondering, this topic can continue with suggestions (please) and eventually, we’ll settle on a solution good for everybody.