Broad undiscussed New England place name reclassification

Hello New England mappers. There has been another widespread place name reclassification across the region that appears to to significantly misinterpret the meaning of various settlement place tags. I’ve reviewed the changes in Vermont so far and have noticed the following issues:

  • Upgraded appropriately tagged place=village nodes to place=town
  • Downgraded appropriately tagged place=village nodes to place=neighbourhood
  • Downgraded appropriately tagged place=town nodes to place=village
  • Downgraded appropriately tagged place=city nodes to place=town
  • Added unnecessary duplicate place nodes, often with name of municiple boundary rather than settlement name
  • Moved place nodes from appropriate location to a less appropriate location

There proabably are some classification changes needed in Vermont, but these don’t appear to align with previous discussions we’ve had or the generally accepted meanings of these tags. I plan to revert these three changesets affecting Vermont and will invite the mapper (@edops) to discuss in this thread.

Similar changes have been made across New Hampshire, Massachussetts, Rhode Island, and Maine.
I assume these follow the same pattern and probably also qualify for reverts, but I’ll wait for others more familiar with these states to look them over.

Previous topic on New England place name inflation for reference.


Thanks Zeke, I looked at the Rhode Island batch and summarily reverted it.

Some changes were debatable, like whether the minor cities around Providence should be downgraded below =city. Certainly if Cambridge, MA is a city, there isn’t an obvious good reason why significant places around Providence couldn’t get the same treatment. (Not really opening the debate – just pointing out that we still have yet to come to a consensus on the suburb question and thus unilateral changes aren’t appropriate).

At least one change I viewed as flat-out wrong based on our discussions in the other thread (the downgrade of Newport to place=town). I mean, it has triple the population of Montpelier and far superior weather to boot… :smiley:

I think it’s important to get a real consensus on moving forward and not just have people randomly reclassify things back and forth.

1 Like

Representative from Maine acknowledges that place classifications could be improved in Maine but there would need to be at least a conversation about methodology before it happens. I’d like to revert all of edop’s changes, but I don’t feel confident enough to do it myself.

I told @blackboxlogic I’d do the revert in Maine, but after looking through all the changesets editing place classifications I’ve decided against doing a wholesale revert. A lot of the changes are downgrades from town to village or hamlet that seem entirely appropriate to me given the settlement sizes. There are also a lot of inappropriate changes like downgrades of city and town nodes for much larger, more important looking settlements and many neighbourhood nodes that should be hamlet instead (since they are isolated rural settlements, not parts of larger ones). Despite the problematic changes, I’d hate to revert the appropriate ones. I think a more targeted approach dealing with just the inappropriate changes is probably best here.

I reviewed Changeset: 149318265 | OpenStreetMap containing the New Hampshire changes.

The downgrades from to city to town for Concord, Nashua, Portsmouth, Keene, Lebanon, and others should be reverted in my view. I’m on the fence regarding the downgrade of Franklin. But that’s a good candidate to revisit after reaching broader consensus on classification.

I have been following here. Would you ask any DWG warning? Has there been any communication with the user?

I’ve left comments on several changesets and sent a direct OSM message with requests to discuss in this thread. I’m hoping DWG intervention won’t be needed. @edops is a new mapper who is clearly motivated to improve place classification across the region, but has just interpreted the the tags a bit differently than what the general consensus seems to be. The ideal outcome would be for them to join this discussion and for us all to get on the same page about how place classification should work moving forward.

Some encouragement, not that it’s needed, but it can’t hurt!

I’ve been “rowing this boat” (of place=* names and their various ways to be categorized in the USA, admin_level=* discussions…) for years, along with many others in OSM. We have forged a lot of consensus about these things, we continue to do so. These can be truly challenging topics, as what state laws / constitutions / city charters… specify “what is” (legally), there are perceptions / understandings / perspectives that “we, the People” have about the communities we live and work in that might accurately resonate with the legal realities (or, as some call political entities like countries, states and so on: legal fictions), or they might not resonate. This sometimes-wide chasm of understanding getting to consensus can be difficult, yet OSM continues to prove that building bridges to such agreement is also achievable. That we have both local experts and seasoned participants (both of whom know when “we know this for certain” vs. “we are in deep water here”) is wonderful and part of what makes this work.

Through it all, people keep talking, resulting in greater understanding, better tagging, really, a better map. Keep up the good work, everybody: keep talking. I thank everybody for the hard work already done and the continuing, ongoing discussions that make our data even better: place=* names and admin_level=* are textbook examples of “community effort” and I nod my head as things improve — they really do. I am honored to be a Contributor here.

Thanks for fixing the cities in Maine!

It seems like there are more issues in Maine. I changed Augusta, the state capital, back to a city. There are other major population centers (by Maine standards) that also seem like they should be switched back, but I am hesitant to be the one to do so. Being relatively new to this, does the incorporation status of a municipality affect how we classify locations in New England at all, or do OSM guidelines trump that? It seems like if the incorporation didn’t matter, perhaps more generic terms could be used for the classifications to avoid confusion.

Here are some overpass queries showing before and after that may help:

It seems like there may have been a few too many place=city nodes in Maine beforehand, but currently things may be too sparse. For comparison, in Vermont we have only four place=city nodes and we’ve been discussing possibly cutting that back to just one (Burlington). In OSM terms, place=town is already a fairly significant settlement, but we’ve yet to come up with clear guidelines to distinguish between place=town and place=city in New England, or the US more broadly really.

1 Like

Agreed. Franklin, like Berlin, uses a city form of government but is just too small to be called place=city. I have reverted 149318265 except for Franklin.


edops blocked by woodpeck | OpenStreetMap 0-hour block on user edops requesting participation in this discussion.

1 Like

It’s great to see forward momentum here! Such things often (heh, case in point) take years to unravel…so far, so good. Thanks to all who positively contribute to OSM, Onward!

It seems to me that no one has any standards that they will ever decide or agree upon. People have inconsistently allowed settlements in Maine with fewer than 100 people be classified as towns, while allowing settlements with populations numbering a few thousand be classified as cities. Meanwhile, States like Mass, Conn, and Rhode Island remain equally inflated, while Vermont and New Hampshire are allowed to remain sparse. In the past, other mappers have come up with and implemented the 100,000+ (city), 10,000+ (town), 1,000+ (village) 1,000- (hamlet) standard. I do not care for this standard, but it is better than the lack there of that has been allowed, and simply removing any efforts to standardize settlement classification without coming up with your own solution is lazy and ineffective. Now what’s worse, is some of you have taken it upon yourselves to once again over inflate Maine place names. Rockland has been classified as a city, despite having a population of less than 7,000. Should towns with populations much larger, such as Berwick, or Windham, be called cities then? Other “cities” such as Bath, Brewer, and Belfast, are labeled as “towns” (rightfully so, due to their small size) and yet they both have similar populations to Rockland. What makes Rockland so special that it deserves city status? The same can be said for Presque Isle, which is, like Rockland, a city in official name only. It’s population is well under 10,000 people, and that’s only due to the massive land area it takes up. Presque Isle is one of the largest municipalities in the state by land area. If it had the same land area as Portland, its population would likely be cut in half. Brunswick, which has, and always will be, a town, has also been repeatedly classified as a city without any discussion. Brunswick is officially a town, and not viewed as a city by anyone who lives in Brunswick, or anyone who lives in Maine. The only true cities in Maine are arguably Portland, Lewiston, and Bangor. Augusta is notable enough to be classified as a city given that it is the state’s capital, despite its meager population, which is well under 20,000. Biddeford is a suburb of Portand, similar to Westbrook and South Portland, and Auburn to Lewiston, all of which have populations greater than Biddeford, but none are classified as cities given they function suburbs of actual, larger cities. It seems to me that no one going to actively come up with a solution, and instead undermine any efforts to implement one.


The way to a solution is discussion and collaboration, proposals, and experimentation, not unilateral editing. Hence your invitation to the forum here.

We came to a consensus similarly on highway classification, which was quite tricky. This topic is no less challenging.

Also, I would invite you to insert some line breaks every now and then for readability.


And yet no one in this forum has offered any proposals. If you need line breaks to read, that’s your problem, not mine. If you have no proposals to offer, why even respond? You only have a response for my lack of line breaks, but not to a single thing that I have discussed?


The main discussion is linked in the original post, where you can find several possible ways forward. I would however aim for a less combative approach if you expect to gain any traction in these discussions.


Hello @edops. Thank you for joining the discusssion. I appreciate your enthusiasm for improving place classification and I hope you will choose to participate in a productive way. Through community discussion and consensus building I’m sure we can make durable changes to the New England hierarchy of place classification. This is not something that can be implemented by a single mapper. Continued attempts to go it alone will only lead to more churn.

In addition to the previously linked discussion on New England place name inflation, there have also been several recent dicussions focused on other areas of the country. Lots of good ideas in these threads.

While no one has formally proposed a set of guidelines yet, we seem to be coalescing around using the Census Bureau’s Metropolitan and Micropolitan Statistical Areas as a starting point. Simple population thresholds for each class of place will not work. A 10,000 person settlement can be quite significant within a sparsely populated rural region, while a 50,000 person settlement can be fairly insignificant within in a densely populated metropolitan region. Place classification reflects the relative significance of a settlement and thus some significant places with low populations will be correctly classified above some insignificant places with higher populations. This is not easy to get right!


@edops please take note of the message in your first warning:

The discussion has to happen in a suitable community medium BEFORE you make the edit.
edops blocked by woodpeck | OpenStreetMap

Five new changes:

  1. Changeset: 150399846 | OpenStreetMap
  2. Changeset: 150403110 | OpenStreetMap
  3. Changeset: 150403267 | OpenStreetMap
  4. Changeset: 150403852 | OpenStreetMap
  5. Changeset: 150403957 | OpenStreetMap

Please hold off on these updates until consensus has been reached.