Currently, the maritime area surrounding Dry Tortugas National Park (way/309644644) isn’t included in CONUS (relation/9331155). It is part of Florida, and Florida clearly belongs to CONUS, so shouldn’t way/309644644 be part of CONUS, as well?
To be honest, I would question whether the CONUS relation even belongs in OSM. The fuzziness of the definition is highlighted by this question!
My answer would be no. CONUS should only include the lower 48 states and Dry Tortugas is offshore of Florida.
1 Like
I figured that might have been the reasoning for excluding it. But San Nicolas Island, off the coast of California, lies farther offshore, yet in OSM it is considered as part of CONUS.
I’ve always viewed CONUS as more of an administrative concept (the boundaries of the 48 lower states) rather than geographic (contiguous landmass), but haven’t seen anything authoritative.
Ah, yes, I see now that Dry Tortugas is in fact considered part of Florida administratively. So sure, I guess that can be part of the CONUS relation.
As a side note, all these extra relations makes editing a massive pain in the butt when doing any work on the US international boundary. Consider a randomly selected border segment on the Maine/Canada border - it’s a member of 15 different relations!
Ha, that’s a lightweight! The real joy begins with route relations. This street belongs to a whopping 247.
Most editors (software) should ease the pain by automatically updating relations whenever a member is split (e.g. separate parking rules for part of a street), but not all of them do, so it’s easy for gaps to develop.
However, any alternative to relations would require keeping multiple copies of the same coordinates in sync, which to me seems far worse.
I’m pretty sure it’s just a concept that exists in the public consciousness. Nothing administrative about it. There are all sorts of informally named regions like this that don’t really make sense to include in OSM. For example:
- The Midwestern United States
- The Southern United States
- Appalachia
- Any of the various Tri-state_areas
1 Like
That mirrors Brian’s assessment above. Maybe best approach is to remove CONUS altogether? For now, I’ve added Dry Tortugas to CONUS, so at least it is consistent with “the admin territory of the US States ex Alaska/Hawaii.” (Dry Tortugas is part of Florida)
However, any alternative to relations would require keeping multiple copies of the same coordinates in sync, which to me seems far worse.
we could have relations that represent route sections, they would have the highways as members just like the routes do now, and the routes would not refer the highways anymore, so the roads would become easier to refine and overlook because instead of all the routes there would be only a single section (or very few).
The routes wouldn’t get new versions just because the road is split for a changing property on the sidewalk, and for changes where the ways change, often you would only update one route section, and all the routes will update “automatically” because they consist of these sections and not of individual highway ways (both in public transport and with hiking routes, there are usually “hubs” (like stations in transport or parkings, mountain huts and peaks for hiking) around which several routes use the same ways, leading in the extreme to cases such as https://www.openstreetmap.org/way/30892492 )
You can add New England to that list as a region with no administration, though it does have a definitive boundary unlike those others. Personally I have found it convenient to have in OSM so that I can query for “X in New England”. Also, without it, searching for New England would send you to New England, South Dakota. And I suppose it could get used for anyone looking to generate a blank outline map of the region, among other uses I haven’t considered.
That is to say, I believe both New England and CONUS belong in OSM and “no administration” is not a great reason to exclude a region.
In United States admin level - OpenStreetMap Wiki we say there are “regions” in the USA (CONUS actually seems to qualify, though we don’t say that there), like “the Midwest,” “the South”… and that these “regions” are largely geographical, possibly (I’d say broadly) cultural regions. There are geographical subsets of even these “regions” which include multiple states, some of which are what some government agency finds it convenient to group together, some of which are quite informal.
Clearly, these (subset of the whole country) “regions” are not especially well-defined, and if they were, it is not certain that whatever entity is defining them is authoritative to do so. I believe CONUS is like this, although I believe the US Defense Department (DOD, our military) uses the term, though it may or may not be especially well-defined. And if it were, would the DOD be the ultimate authority to do so? What if another US government entity defined Dry Tortugas (as being part of CONUS, or not) differently? We already have this in, for example, the US Virgin Islands, where the Commerce Department (Census Bureau) says there are three administrative divisions of the islands, and the Interior Department (I think) says there are two (grouping two of the three islands into a single administrative grouping, or what might more accurately be called a statistical grouping, even if the islands themselves don’t do this).
The whole point of pointing this out in our wiki is to admonish OSM users who want to so classify such sub-groupings to be very careful — usually NOT to do this. OSM certainly DOES have truly administrative boundaries (and we’ve been careful to get these right in the USA…it has taken a decade or more of sometimes rancorous discussion) which should be in our map. But for “statistical boundaries,” either “not so much,” or outright “no.” Sometimes it makes sense to “keep a relation to hold all the pieces together for convenience” (as in the case of the statistical grouping US Minor Outlying Islands), but it can be instructive to drive the point home (early and often) that this is “a statistical grouping, done for convenience, and not anything authoritatively administrative.” I think CONUS might be another example of this. We might keep “for convenience” a relation describing it, but as to its authority or “officialdom,” that almost immediately gets either questionable or contentious rather quickly.
Indeed the New England object is also questionable to have in OSM. New England is a cultural region and as such doesn’t really have a definitive boundary. For example, I would exclude Fairfield County, CT as an honorary piece of New York, while others might disagree 
. Meanwhile us northern Vermonters are culturally more connected to Montreal than Boston.
It certainly is convenient having regions like this in OSM, so why not add more? Just in Vermont I could add region polygons for Northern Vermont, Central Vermont, Southern Vermont, the Northeast Kingdom, the Champlain Valley, Northern Green Mountains, Southern Green Mountains, the Taconic Region (crossing into NY), Northern Piedmont, Southern Piedmont, and the Upper Valley (crossing into NH). Many cities could have a Greater {City Name} Area region polygon around them. Etc, etc. Would this improve OSM? Perhaps. Would it be an annoying source of disagreement between mappers about where exactly these regions begin and end? Yes, definitely. Such disagreements about CONUS and New England are avoided since the polygons happen to coincide with administrative boundaries, but this seems like a poor reason to include these regions while excluding others. We should either map region polygons or not, regardless of how they are defined.
2 Likes