[OSM-dev] Proposal for OpenMetaMap - proper OSM import solution

Anthony osm at inbox.org
Tue Aug 30 00:03:34 BST 2011


On Mon, Aug 29, 2011 at 6:23 AM, Jaak Laineste <jaak.laineste at gmail.com> wrote:
>  I am a bit more worried about probability that some of the external
> sources are not available from time to time, temporarily or even
> permanently. To solve this I see three options:
> a) live with it, contact providers to provide stable API
> b) cache external data in OSM/OMM central cache/backup service. At
> least sources which allow caching
> c) create weekly planet-with-externals.osm.pbf file, which would be in
> principle same as cache (more precisely proxy), just all-in-one and
> easier/quicker to use. This would even work also with old osm2pgsql.

[I took the liberty of correcting your second b to c]

I think b) has a lot of huge advantages when the license permits it.
Three of the biggest are 1) the cache/backup can be stored in a format
which is easier to mix with OSM data (if it can be done without losing
information, a .osm version of the external data would be great); 2)
the stability of any keys is in our hands, not theirs; and 3) history
information can be kept even if/when the original data provider
doesn't keep it (or doesn't publish it).

The way I see it, you have three databases - the OSM database, a local
copy of the external database, and a local matchup database, which
links the two together and contains fixes for the external database
which haven't yet been accepted by the external data provider (or have
been rejected by them).



More information about the dev mailing list