The problem I see is that you’re conflating visibility, difficulty, and maintenance. For example, when I tried hiking the Ragged Ridge trail back in 2016, peak incline was 30% (class 1), the surface was packed dirt (class 1), the trail was visible but had encroaching vegetation (class 2) except for one brief section where the precise routing of the trail was uncertain due to an absence of vegetation (class 3), there were numerous waist-high trees across the trail (class 4), and there was no fall risk (class 1).

So what level of difficulty is this, and does it really drop from class 4 to class 2 once the trail crew gets there?

#1 might not legally be a turn, but if I were driving that road at night, I’d sure want my GPS to say something to explain why I’m seeing oncoming headlights to my right.

You’re missing a critical step: verify the building footprints against what’s actually there. I need to do a proper writeup of it, but in my checking, the Microsoft’s building footprints have an error rate between 10% and 50% depending on the situation (and a fairly consistent 25% non-detection rate).

According to Copernicus Sentinel imagery, the most recent additions are generally (barely) outside the burn area. The ones in Malden very much weren’t, and I’ve been working on figuring out which ones did or didn’t survive.

23 of 26 Spokane County drop boxes mapped. They’re surprisingly easy to spot in aerial imagery.

The Washington Secretary of State has a complete list of ballot drop boxes at https://www.sos.wa.gov/_assets/elections/drop%20boxes%20and%20voting%20centers%20pdf.pdf

Looking at local examples, there are some disused:names for places that no longer exist, but which have left traces in things like business names. There are also some name:lang localities for places that aren’t significant (and don’t have a name) for the majority culture, but are important to a minority.

Blanket deletion is not a good idea.

In this particular case, the best thing you can do is ignore it. Based on the available imagery, this is a very minor stream that is mostly not visible from above. In order to get the routing correct, it needs to be mapped on the ground, by someone following it with a GPS.

I don’t think Guantanamo Bay Naval Base is a border dispute of the sort you’re trying to map here: the USA and Cuba both agree that the land is part of Cuba. The dispute is that the USA says they’re leasing land for a naval base from Cuba; Cuba says the lease was agreed to under duress and is void.

Let’s look at what the robots.txt file really does, rather than just counting up the number of lines:

Disallow: /traces/tag/
Disallow: /traces/page/

These are various alternate ways of searching the GPS traces that have been uploaded on the site. The main trace listing is still accessible.

Disallow: /trace/

This is the API endpoint for accessing GPS traces. It is not intended to be displayed in a web browser, and contains nothing useful for a search engine.

Disallow: /api/

This is the API endpoint for editing the map. It is not intended to be displayed in a web browser, and contains nothing useful for a search engine.

Disallow: /edit

This is the URL for the in-browser editor. Everything under this URL is behind a login barrier, and it contains nothing useful for a search engine.

Disallow: /message

This is the URL for the on-site PM system. Everything under this URL is behind a login barrier, and it contains nothing useful for a search engine.

Disallow: /login

This is the above-mentioned login barrier. It contains nothing useful for a search engine.

Disallow: /history

This is the visual history browser. The contents change far too rapidly to meaningfully index on a search engine.

Disallow: /geocoder

This is the on-site search system. Search engines searching search engines never ends well.

Disallow: /browse

Disallow: /*lat=
Disallow: /*node=
Disallow: /*way=
Disallow: /*relation=

These are obsolete URL hierarchies for browsing individual map elements. The current URL hierarchy, with URLs of the form https://www.openstreetmap.org/way/238241022, is still crawlable.

Disallow: /user/*/traces/
Disallow: /user/*/diary
Disallow: /diary

These are the only entries that block pieces of the site that might be of interest to a search engine.

You’ve got considerably more faith in peoples’ willingness to read than I do.

On the surface, your challenge looks like something that any armchair mapper could clear out in an hour or two. My expectation is that if you open it to the general userbase, someone will come along and see a nice, easy task, and proceed to pump up their leaderboard score, completely overlooking the bit about “locals only”.

I know what you shouldn’t do: MapRoulette. I’m still dealing with the aftermath of a MapRoulette “cleanup” of TIGER roads in Eastern Washington.

Proper cleanup will require research and probably local knowledge to tell the localities from the ghost towns from the (former) railway stops from the trailer parks from the unincorporated communities. All in all, far more effort than the average MapRoulette user is willing to put forth.

The problem with the Condorcet method is that any voting system that’s guaranteed to produce the Condorcet winner violates the participation criterion: that is, there are situations where you can cause someone to lose by voting for them.

Straight density doesn’t strike me as being an effective measure of “mappedness”: in a perfectly-mapped area, it should closely reflect population density.

Estimating the expected number of POIs should give a more accurate measure. For example, the ratio of the number of POIs to the length of roads, or the ratio of POI density to the above-mentioned population density.

Odd that you’re getting such a high loss rate with 0.6.11: I’d estimate my loss rate at less than 1% for images, and 0% for audio.

OSMI (https://tools.geofabrik.de/osmi/), Osmose (https://osmose.openstreetmap.fr/), and JOSM’s validator will all warn if a river segment’s downstream node doesn’t connect to a suitable point. I think there may also be one that uses elevation data to find wrong-way flows, but I can’t find it.

How does your system deal with https://www.openstreetmap.org/node/154309316 or https://www.openstreetmap.org/node/160368207?

At least in the US, on-ramps have destinations. For example, the interstate near me has signs indicating “I-90 west to Seattle” or “I-90 east to Spokane”. See https://www.google.com/maps/@47.393803,-117.8388722,3a,75y,10.56h,88.38t/data=!3m6!1e1!3m4!1s5EaWcp1m5ZYD5zzcXUtKtA!2e0!7i3328!8i1664 for one such sign.

Not only is the sign an exit speed limit, it’s the yellow of an advisory speed limit rather than the white of a mandatory speed limit.

Could be worse. Near me are roads with names like North North Road.

I don’t know how much use this will be. I checked all 15 detections in Eastern Washington, and only one of them was not already mapped through road geometry or turn restrictions.