If no one is publishing a copy of something with an open license, then that is the definition of de-facto closed.
de-facto means what is the reality vs what is the theory.
In theory you can get a copy because it has a license that says so.
In reality you can not get a copy because you are not one of the people with physical access to some existing copy.
It is de-facto closed while that set of facts is true.
There are these books that explain the meanings of terms. We don't have to guess and have random individual ideas about what words mean, and then wonder why no one can communicate.
Yes, and "in-facto" is not the same as "de-facto".
De-facto means "describes practices that exist in reality, regardless of whether they are officially recognized by laws or other formal norms". It is not a precise term.
I have some code on my server that I wrote with an MIT license. You have no way of accessing my server, and the code has not been made available elsewhere. You can't even know that the code exists, much less that it has an MIT license. To you, it is for practical purposes the same as closed-source software, in the sense that you cannot obtain the source code of the software, despite it having an open source license. This is an applicable situation for "de facto" by its common meaning.
"Open source" is not even a precisely defined term, although people who are persnickety about the definition refer to the OSI definition, which includes "Where some form of a product is not distributed with source code, there must be a well-publicized means of obtaining the source code". By this definition, code with the MIT license held secretly in a private repo is not open source. It became open source when Roman made it public, as was his right to do by the license. The license is one component of what makes software open source, but there are other components as well.
And then Roman made the repo public, effectively distributing the licence to people and thus making it open source, without coordinating with the other contributors?
I don't see any replies by Alexandr, but it feels like this could have easily been resolved with a less antagonistic response by Roman - but of course, I don't know what other history they have.
* I guess technically, it was given to the other contributors with access to the private repo - i.e. Roman did have the legal right to distribute it further under the MIT licence. Presumably, the original contributor (Alexandr) just applied that licence by mistake.
So there are people who contributed the code under the MIT license, so its not a sole work of Alexander and the license was not added as a mistake.
IMO from a legal standpoint Roman is absolutely in his right to redistribute the code.
I'm not sure if there is some distinction between software and map-data entailed in the discussion of this "server", "software repo", etc. but assuming it's all one thing:
if the content in question was MIT-licensed as specifed by the license in the repo, any one of the members of the project with access to the material would be within their rights to make copies public. There is no de-facto closed source wrt open licenses.