What are the quality criteria for names in OSM?

In general, we work with the OTG rule - i.e. it should be verifiable on site. That doesn’t always work with names. In particular, natural objects such as mountains, rivers, lakes, etc. are often not signposted. If we use maps as a reference, the suspicion of copying other data may arise.

Some objects do not have official names but may have local, regional or nick names. Here it is difficult to prove whether and to what extent these names are in use.

In the German forum we are currently discussing a nickname for Autobahn 2 “Warschauer Allee” this name is said to be used by truckers. He got into Wikipedia without a source reference and was consequently picked up by the press for reports. After removal from Wikipedia, usage in the press also seems to be declining.

When is a name relevant for OSM and how should its (“common”) usage be proven? (It may be necessary to set different standards here depending on the importance of the objects and, for example, put the use of the “name” in relation to the frequency with which the object is mentioned.)

Or is quality completely irrelevant to us and anyone can invent, enter and spread (fake) names?

Please lets talk here about names in general. To discuss “Warschauer Allee” please use the other thread.


There’s also the case of hikers giving names to lakes, streams and peaks. In Italy I remember at least two cases of users giving names to nameless elements they “discovered”.

1 Like

Or is quality completely irrelevant to us and anyone can invent, enter and spread (fake) names?

a fake name is one that nobody knows of, we do not want this. IMHO the limit is “rarely used names” which can still be included, if usage is less (nobody knows it, no hits at all for a web search, etc.) we should not have them. Basically we shouldn’t be the only source, if you go to the place and “ask all locals”, more than one should know it.

1 Like

Personally, I expect that name appears on some sort of sign somewhere, while reg_name, or loc_name probably won’t. But I would then expect that a significant amount of people near the feature would know the loc_name or reg_name. How many is significant? I would say that if chances are, that when asking for directions, a local would use that name, are higher than 10%. But that’s just a gut-feeling and it could really be any number. But how do you know how many people know a feature by a certain name?

The only way to know, is very likely by being one of the “locals”. So loc_name and reg_name should probably only be added by people who live in that region/area, or have seen a sign with that name, saw that name used in a local magazine or in a local newspaper.

Other than that, I think loc_name and reg_name

  • do not work for things like other countries, highways, rivers - basically everything big enough that there are no “locals” to determine what the “local/regional” name is, because there are too many regions and names to consider.
  • should not be derogatory or even insulting (like people from one city giving the twin city names like “Gannover” instead of Hannover, to emphasize that there are issues with crime (Ganove = someone who steals))

But this only for features where the name is clear and unambiguous. Once you add things like the Camino de Santiago pilgrimage trails, you will see that there are tons of names in use, but none of these is really official, because they are derived from different historic sources. And depending on where you are, there will be different names on signs. I don’t think I could come up with quality criteria for those, that wouldn’t dismiss another name as invalid/inferior/unused, that might actually be widely used somewhere. I think you gotta be really careful and delicate when dealing with names like these. If in doubt, more is more.

But coming back to the original question: A name should be considered relevant, when it’s on a sign or map somewhere. If not, then my answer would depend on the type of feature we’re talking about (village vs. hill vs. road vs. school vs. art, and so on). I don’t mind adding many names to a feature, if it enables people to search for it, but then we should be very considerate about which name we want to show on a map. If there is no name=*, do we can to show alt_name / reg_name / loc_name, or better no name at all?

ask locals

(yes, this is trickier for nicknames for roads supposedly used by truckers - but in such case asking truckers or observing is name in use on their discussion forum is still possible)

I would not expect this. For start many natural=peak definitely have names but are not signed on the peak itself. Lets take K2 as an extreme example.


But then you have the problem of discrepancies betweenteh signpost & the apparent “official” name.

A couple I’ve seen in the last few weeks:

  • the sign at the highway bridge says that this stream is “Doctor Creek”, but the government map has it listed as Doctors Creek

  • the road sign says “Arcadia Valley Road”, but other maps & sources says “Arcadia Valley South Road”

Which should it be?

But you then have “This Highway” passing through a small country town. Where it goes through the town, it’s known & sign-posted as “That Street”. Should it be named Highway, with loc_name (or should it be alt_name?) Street?

Or should they be the other way around - name Street, alt_name Highway?

If in doubt, I would use the signed name for name and the others as official_name or alt_name, unless the signed name is obviously erroneous (like typos), in which case I would put the correct name into name and the other one in description or note, since name:signed is being used for something else. This is my interpretation of OTG truth.

I should have made myself more clear: I’m talking about something like Route 66. If that was known as “Freedom Route”, but only in Flagstaff, and “Mother of all Roads”, but only in Chicago, and other names in other cities/regions, and they would all be referring to the entire Route 66, then yes, these would all be regional names. The problem I see is that, of course, you could add all these to the route as regional names, but without specifying which region that name is used in, it’s losing significance. If I told you “The regional name of the Route 66 is Freedom Route”, you’d probably answer “Yes, but only in the flagstaff region”, unless we were in Flagstaff. So my point is: regional names for a feature larger/longer than the region, can be problematic. They don’t have to, but they could be misleading.

Okay, okay, what I meant was: signs, maps, literature, and so on. Don’t make me think for hours to come up with all possible sources :laughing: And I don’t mean signs like “This is the K3, prove me wrong” in someone’s backyard…

Okay, okay, what I meant was: signs, maps, literature, and so on.
Don’t make me think for hours to come up with all possible sources
:laughing: And I don’t mean signs like “This is the K3, prove me wrong”
in someone’s backyard…

It is indeed difficult to get right. I remember one case where a
persistent vandal, claiming that some hill in a forest was named
so-and-so, actually but up a sign in that forest naming that hill. So
while I totally support giving a lot of weight to verifiable
on-the-ground truth, some common sense has to go into that verification
process too.

1 Like

Just note that there are some people who actually demand in-place sign and reject local survey and asking people, general knowledge and so on (K2 is my favourite counterexample)

There are also some who expect that solely names in official database are mappable (which is also obviously wrong)


Just to add a silly (but all too real) example near me, bus stops have several valid names:

  1. The name on the bus stop.

  2. The name displayed on the TV screen on the bus itself.

  3. The name announced on the audio at the same time that the stop is displayed on the TV screen (yes, sometimes completely different to 2)

  4. The name in the data feed (presumably NaPTAN via traveline; this is the UK) used by apps that display the stop name (Google, Oeffi etc.)

1 Like

What I actually wanted to talk about was: if we suppose we have a really known (significant) object that we know what its official name is, or even that it doesn’t officially have a name (but maybe a reference (ref)).
Do we allow any other name for this object just because it appears in some internet source or someone claims to know it by this name or do we make certain quality requirements for the name here.

In a community, the object is known by a nickname, whether it’s the road among truckers or the climb among cyclists. It could also be a name that was supposed to be floated by a tourist board but never got accepted.

depends on its quality, joke website does not count

depends on its reliability

In both cases surveying this (asking multiple people living in a given area) could be helpful in case of doubt (though for minor named features you may need to ask many people)

1 Like