[mkgmap-dev] Missing addr:place addresses

Mon Dec 3 01:34:41 GMT 2018

Lorenzo Mastrogiacomi <lomastrolo at gmail.com> writes:

> If I understand correctly I would say yes.
> unclassified roads should be tagged with name=* in osm if they have a
> name and should not take names from nearby roads.

I am really not following what's going on.

Most unclassified roads will have a name and some might not.

Some service roads (and tracks) will have names and most will not.

Addresses on building/nodes will have a variety of address components.
Typically, an address with a given addr:street will be located
reasonably close to some highway way with that same name.  But not
necessarily; the world is not designed by computer scientists :-)

I"m not really clear on addr:place; that sounds like a name from the
settlment hierarchy rather than the administrative hierarchy, and which
is used in addressing will depend on local customs.

I think this question is about putting address ranges on named highway
ways, because the garmin format doesn't represent individual addresses.
I can certainly see merging ways that are the same in tags that mkgmap
considers important (ignoring width changes, etc.).  But I don't
understand merging ways with names and ways without names.

A further complexity is address points far from the road in question,
and perhaps closer to some other road (e.g. a 1 km driveway).  There, I
could see a process of figuring out that the address point and the named
way are related by the service road and thus the address point should
perhaps be considered to be on the named way where the service road
connects.

I admit to never having taken the time to make addressing work in my
garmin maps so far; I mostly use them for hiking and I use OSMAnd for
car navigation.