I'd like to point out that it only "gets pulled in" if the data is actually made available by the government, and there are dedicated volunteers who work on getting the data, massaging it into the right format and importing it into OSM, and that is not the case in many countries.
replyNot to mention the occasional drift between what's on paper and what's actually been built.
replyYeah, that's a whole different kettle of waspnests. :)
replyThe thing lacking even more is a good search engine. Doesn't matter how complete the database is when search keeps missing text appearing verbatim in a label.
replyThis is the problem with using vector search for everything: there is no “verbatim” when the corpus and the search are both converted to a vector, and not necessarily using exactly the same transform.
reply