[OSM-talk] Launching OpenAddresses

Frederik Ramm frederik at remote.org
Tue Mar 25 23:43:30 UTC 2014


On 25.03.2014 23:14, Paul Johnson wrote:
> Am I wrong to assume that this would be considered a potential data
> source on a limited scale for folks who wish to do the conflation?

I think it would be great if someone were put a little work into
importing this data directly into a locally held Nominatim database,
just as can already be done with TIGER address data. This would mean
that anyone could set up their own geocoder with an address data set
larger than what is available from OSM, without having to go through a
proper import; no conflation is necessary in this case because Nominatim
can be made to resort to the secondary data set only if nothing is found
in OSM.

Please help spread the word that other than such "downstream mix-in",
copying more than a handful of addresses from the OpenAddresses data set
*is* an import and has to go through the same import process as
importing any other data set.

Most bad (un-discussed) imports are committed by people who didn't know
any better. OpenAddresses, if used properly, can be an excellent valve
to take out the "OMG free data, import immediately" pressure out of OSM
mapping. Let us make sure it actually has this positive effect.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"

More information about the talk mailing list