geo hint "city" field is ambiguous

[npdoty]

RFC 8805 didn't specify what city meant, except that it's a free form string. There is a hint that it's equivalent to RFC 4119's A2 element, but that appears to be incorrect in practice. A2 is described as "county" and the common examples don't seem to be at that administrative level (most that I can see are city or city division).

If IETF is going to publish new geolocation format standards, I would recommend working from geopriv, or providing a precise mapping (like the sample XML in RFC 8805) or at least pointing to other standards that are less ambiguous.

[tfpauly]

The point here would be to include literally what is in a Geo IP feed file, and align with that rather than sharing new information that would not be available from the IP address itself. To that end, I don't think it makes sense to change this to be divergent? It would be better to revise 8805?