OpenStreetMap
I've been adding some addresses.
I've been using nodes for the entrances tagged with the addr:*=* tags and building=entrance.
The reason for this is because some buildings have multiple "bits" in them. I've seen others dividing the building into smaller sections instead.
I'd say the node way is better, but I just wonder if it's good practice.
Discussion
Comment from Chaos99 on 27 January 2012 at 06:35
Nothing wrong with that.
Adding entrance nodes is the easier way, because it works every time, and it also adds the additional information of the location of the entry.
Splitting buildings is only a good idea if it *is* a separate building. (Like inner city houses build side-by-side along the street). You can mostly spot that by different kinds of roofs or wall decoration. It's not a good idea if it really is just one building with several entrances (like an apartment block), because the separation would be only *virtual* and doesn't represent any facts about the real world.
There is an ongoing discussion about this on the german mailing list, but it seems to not matter if you tag the entrance (or POI) node(s) or the building itself with the address if there is only one. Just don't tag the same address multiple times (like with multiple POIs inside the same building). Tag either just the building or use a relation to group them all together.
That's all just my opinion of course. At the end there is no *right* way to do things here.
Chaos99
Comment from SuborbitalPigeon on 27 January 2012 at 21:56
Yes of course. I always try and do something that others do for the sake of simplicity.
Comment from marscot on 28 January 2012 at 14:29
bruce that is the exact way I have been doing things, check shawhead coatbridge, I also use the wheelchair access tag too on the entrance if I know it, also access=private for houses