How to handle massive name deletion?

I stumbled over the following three changesets:

Where a user deleted around 200 tags name:hy from formerly Armenian, now Azerbaijani locations in the Nagorni Karabakh area.

As far as I am aware, OSM policy is that we use “reality on the ground” rules for this, which means that the name tag should change to Azerbaijani language after the takeover of the region. But I assume that even in this case, the old name should remain as a multilanguage tag?

Is that correct?

If yes, would it be justified to revert those three changesets?


Do the DWG know?

It is IMHO definitely vandalism. The Data Working Group should be informed and the user blocked in order not to continue and the changes reverted.

I was too quick in my previous answer. The correct approach would be: contact the user by commenting the dubious changesets and ask him for explanation and correction. If he does not react or correct and it is proven case of vandalism , then revert the changes. If the user continues in unwanted activity, contact the DWG for help.

hello… after the de-occupation of the Karabakh lands, the government decided to revise the names of administrative units that were under Armenian occupation and were willfully renamed by the leadership of the unrecognized Nagorno-Karabakh republic… I note that many settlements not directly related to the Nagorno-Karabakh zone, which are the occupied regions of Azerbaijan, were also renamed in the Armenian way. … I want to assure you that now all these settlements are returning their names in accordance with the legislation of the Republic of Azerbaijan… as for the pro-Armenian users Taktaal (I justified my answer to him in the comments) and the user Tomas, who unexpectedly supported him, I noted that any user can add an “old name” tag in any language, but not distort the real and historical names of a country to which he has no connection… I live in Azerbaijan, I have been an OSM user for almost 11 years, and 70% of the work is on Azerbaijan belong to me… I am far from self-praise, but I would like to note that I am here and know the situation much better than the same Taktaal… working in OSM is not just a hobby and passion for me… thank you, with best regards…

1 Like

As far as I am aware, you were correct in changing the “name” tag from Armenian to Azerbaijani after the change of ownership of the region, but you should not have deleted the “name:hy” tags. But, I don’t want to claim that I am an authority here, so I will refer this to DWG, they will be better able to formulate policy and make a decision.

By the way, first, I believed it only affected the 4 towns in Changeset: 142772453 | OpenStreetMap so I fixed those by manually editing them again because I believed it was just a small thing. But later on I found the three additional changesets listed above with 200 additional deletions, which is why I wanted to collect further feedback.


I agree with Taktaal in this. The best solution is imho to raise the case to the GWD now and to make it precedens for all such cases. Thanks. Tomas

1 Like

Curiously, not knowing what the OSM key for the Azerbaijani language is and if same as in “Karabakh lands”, if the local name:xy=* is not the same as the name=* it gets flagged by QA. Seen quite a few in Italy where what’s written in the name field is not exactly equal to the name:it field.

Italy isn’t unilangual, it has a French speaking region in the northwest (Aosta), and a German speaking one in the north (South Tyrol).

We have very obscure areas such as a little pocket where they speak an Albanian variant as well, up in the Dolomites there’s an area where most any all who speak German and in fact, go into southern Switzerland and they all speak Italian, but that’s not what I was pointing at. name:it represents the national designated tongue and OSM seems to want to match that to what’s written in the name field. Is the removed name:hy NK language or the Azerbajani? Checking a few, it really boggles the mind that any mapper would deem it appropriate to remove that language tag.

As far as I have seen, that is not the policy in OSM that a name has to represent a “national language”. Mostly because soo many countries don’t have one, I think definitely more than 50% of all countries in the world have regions with minority languages. And so far, those minority regions are mapped in OSM using the local language, not the majority language of the country.

The current rule in OSM is that the language of the name tag has to correspond to the language of the people living nearby.

But, all this discussion is not relevant here. The former Armenian population of the Nagorno-Karabakh region has been thoroughly displaced, and the reality on the ground is that the region is now Azerbaijani speaking. This is why I said I don’t disagree that the name tag should now be in Azerbaijani. My issue is whether in such a change of land ownership our policy should be:

  1. We change the name tag to the new language, and move the old name to the corresponding multilanguage tag (“name:xx”)
  2. We change the name tag to the new language, and delete the old name completely
1 Like

To be clear, this is what is being checked here, in Itay… if a bar has the name=Giorgina filled out and then name:it=Bar Giorgina, that is being flagged. So my guess is if name=xyz and name:az=xyz it would not raise an alert if equal today.

There a discussion on dual naming i.e. both in the name field with a separator, be it comma, semicolon, slash, whatever is acceptable.
Deleting the NK name or not transferring it to an ‘old’ tag or capturing it in name:hy is dubious to me. It’s like that picture where certain persons were disappeared from… politics.

Definitely option no. 1, not deleting anything.

“Note that even if Armenian population will be fully expelled from Nagorno-Karabakh (see Flight of Nagorno-Karabakh Armenians - Wikipedia ) then it is not valid to remove name:hy=* and in case of changing name=* tags then Armenian name must be kept in name:hy=* as Armenians names are still in use”


Hello again everyone…

  1. I don’t understand the link to the Wikipedia page, which was edited literally 5 days ago (!)… it is clear that this page was created by the Armenian side and cannot be impartial and neutral… I can put here similar links from Wikipedia created by the Azerbaijani side side where the exact opposite is noted… but I don’t do that…
  2. today in Karabakh (not only in the former Nagorno-Karabakh itself) a certain administrative division has been created and already exists, in which specific names of settlements are given… for example, the city of Agdam… this city is now called Azerbaijani Hiroshima - there are no Armenian occupiers there , they did not leave a single building, everything was destroyed to the ground… I was born in this city, and after the Armenian occupation I became a refugee, and Armenians have never lived in this city in history… in any source on the network this city is called Agdam, although after the occupation of this city the Armenians renamed it Anka, and as far as I know, this was previously displayed here in OSM (and by the way, then no one was outraged by the massive renaming of Azerbaijani names into Armenian ones)… the question is - what does the fictitious name Akna have to do with it? has the real name Agdam?.. after all, such a name never existed… the same is with the Azerbaijani city of Fizuli - the Armenians came up with the name Varanda for it… these cities were not even part of the former Nagorno-Karabakh autonomy… .and such settlements according to there are hundreds of data on Karabakh. …I’m just wondering which bus driver, taxi driver or which guide will understand when they tell him “take me to Varanda”? all foreign guests coming to Karabakh, journalists, officials write Azerbaijani names in official applications to visit the country, and not invented Armenian - and this is the demand of the official authorities of Azerbaijan…
  3. I quote Taktaal - “The current rule in OSM is that the language of the badge must correspond to the language of the people living nearby.”… i wonder who can be considered living nearby?.. based on the fact that there are no Armenians in Karabakh now (why are they not there , this is a different question, a political one, not an OSM question), and while there is no demarcation or demilitation of the border between Azerbaijan and Armenia, then what does the expression “people living nearby” mean?.. Iranians also live nearby - which means it turns out that fictitious Persian names can be added to all the names of Azerbaijani cities?.. where is the logic here?.. in such cases, links to official UN maps are important, but not to hastily edited Wikipedia pages…

Taking into account the above, I believe that the tag “name:hy” for the de-occupied settlements of Azerbaijan will be incorrect, since it will only aggravate the problem… as an alternative, it would probably be possible to apply the tag “alt_name:hy” (or “old_name:hy”) , which will rely not on invented names, but on transcription using the Armenian alphabet… the Azerbaijani language has a Latin alphabet, so Azerbaijani names are perceived adequately and are acceptable throughout the world…

I repeat, I have been editing in OSM for almost 11 years, and more than 70% of the map of Azerbaijan was compiled by me, due to my profession I travel all over the country and know every corner of it… and this is not just a hobby for me, but a desire to present my country to everyone wider and more accessible to the world, with nuances that reflect reality and facts… what are those who call for blocking me so simply trying to achieve?.. how should I understand this, and what will it give to others?..

Thank you for your patience and patience…

I will use the example of Agdam here to try to illustrate what our problem is and hope that you understand.

If you look at how Agdam was handled by OSM in the past by looking at Node: ‪Ağdam‬ (‪1434761991‬) | OpenStreetMap you will see that in the 2010s while it was inside Armenian territory, it was tagged as “name=Ակնա”, “name:en=Akna” (which is the transliteration of the name tag). This follows the rule that the name should always be of the local population. And in the 2010s, that was the Armenian language.

But! You will also notice that in all those years, the town always kept its “name:az=Ağdam”.

Then, in 2020 when ownership of the town changed to Azerbaijan, the tagging was adjusted. Now it’s “name=Ağdam”, and “name:en=Aghdam” (again, transliteration of the name tag.

But! What needs to remain is “name:hy=Ակնա”

This is the most neutral way to handle this. It’s not just your area that has this problem, there are similar issues in a lot of other conflict zones in the world, and this has proven to be the way that is most acceptable to everybody.


And by the way, a name is not ficticious because it’s in another language. i.e. the city of London is called Londres in french. We put that multilanguage name in the city Relation: ‪London‬ (‪65606‬) | OpenStreetMap to allow people who speak that language to search for it.

It doesn’t matter how much an English person might dislike it if a French person call it Londres, it’s still not ficticious, it’s just how it’s called in their language.

1 Like

I’m sorry… Agdam was on the territory of Armenia??.. are you seriously saying this?.. which international legal organization recorded this?..

yes… Londres… may be Londra, Londonus, Londis, Landen… but not Tegusigalpa for example…

may be “name:hy=Աղդամ”(Aghdam), or “alt_name:hy=Ակնա”(Akna)…

since before the occupation of Karabakh, even in Armenian sources it was written and called Agdam… Anka is an invented name for an Azerbaijani (not Armenian!) city, in which Armenians have never lived in history… do you understand the difference?

1 Like

put them in old_name:hy