If one consults the key:boundary and key:admin_level wiki pages, they seem to insist that admin_level key should be used only for boundaries of type administrative.
However, admin_level tag is also commonly used to indicate the level in the hierarchy of a boundary=religious_administration, with German, Polish, French, Hungarian etc. wikis even providing a precise subdivision mapping standard.
This combined outcome is not only contradictory as a whole, but also creates further problems. For instance, basing on the assumption from the first paragraph, some applications render administrative borders basing only on the presence of the admin_level key, ignoring the value of the boundary key. Which in turn leads some mappers to remove this level value from religious boundaries in OSM altogether, in order to “fix the map”… As can be seen e.g. in this thread.
I don’t rememeber edit wars over this sort of thing - can you link to an example?
I suspect that most consumers of admin_level also look at boundary=administrative, so I’d be surprised if many got confused by the status quo. There is occasional confusion over historic boundaries of various sorts, but that is a different issue.
I think ignoring the value of the “boundary” tag and only looking at admin_level is creating the problem and those mapmakers who do it should be made aware.
Generally, the “admin_level” documentation should be amended to explain also other uses outside of the “primary use” together with boundary=administrative.
Consumers that ingest admin_level without looking at boundary are making a mistake. I say this from the perspective of being a data consumer of that tag. There’s all sorts of weird tagging out there and you want to be explicit when you’re consuming OSM dara.
Thank you for your replies. Since we seem to have a consensus on this matter, I’ll make appropriate edits to the wiki (that is, describing “other” uses of admin_level tag) if no one will voice their dissent shortly. I also agree that fixing the relevant applications or map styles will never hurt.
Well, maybe using the word “war” was too clickbait-like from my side, but I also said “can lead to” and not “led to”… I am unable to find an example in changeset history now, but there were some instances where admin_level tag was stripped from religious boundaries, and the logic (which was held by more than one user) behind restricting this tag to administrative borders was “if wiki says so and some applications do so then it must mean it is so” – hard to battle against such a rhetoric, even if it sometimes proves to be wrong… Also it is not hard to imagine such situation could easily repeat in the future with some other user if wiki was not to be corrected.
I am unable to find an example in changeset history now, but there were some instances where