[Imports] NYC building + address import - to merge or not to merge?
frederik at remote.org
Mon Oct 14 18:43:06 UTC 2013
> ## Option 2: Always keep address points separate
My $.02c -
* address on house outline - common
* address on a node when no houe outline exists - common
* address on a node that is part of the house outline - frequently used
when people want to be able to locate the entrance
* address on a node somewhere within the building outline "near to" the
entrance - strange
I must admit though that I don't have numbers to back this up. We can
easily see from taginfo that we have 15m addresses on nodes and 9m
addresses on ways, but we can't see how many of the 15m nodes actually
have a building outline around them.
> a) this is the NYC GIS way, making it nicer for GIS folks to use OSM
GIS folks will likely have to be able to work with all kinds of
addressing anyway - even if you decided to do it one way in NYC, mappers
might still use other schemas later on.
> b) this is the generally applicable method. No matter whether we have
> one or multiple addresses you can expect to find a separate node
> carrying address information.
Yes but the only relationship between the building and its address will
be a spatial one - i.e. the operation "find the address of this
building" is much more complex than with option 1.
Placing the address node *on* the outline of the building might be a
good compromise. If you're not sure enough whether it is an entrance
then leave that out. (Care has to be taken though that over-eager use of
"make rectangular" then doesn't drop these nodes because they appear not
to be contributing to geometry.)
> c) retains useful information
> Note: it has been suggested to use the address location information to
> tag an entrance. Unfortunately the data is not consistent enough to do this.
This does not make much sense. What part of the information is useful
and worth retaining if it isn't even useful enough to assume that there
is an entrance there? Either the info is useful, then let's boldly claim
that there's an entrance where the building outline comes nearest to the
node, or such an assumption would be dangerous/stupid, in that case
there is no useful information to retain.
Frederik Ramm ## eMail frederik at remote.org ## N49°00'09" E008°23'33"
More information about the Imports