[OSM-dev] Saparate "layer" for imports?

[Jaak Laineste]

What do you think about setting up separate "layer" for data imports?
The Imports Layer would hold 3rd party data so it could be manually
carefully reviewed and copied (using copy-paste or tracing) to the
main database. The principle similar to GPS traces, but with
structured, fully prepared (tagged, topologically cleaned,
relationificated etc) vector data.

Advantages:
 - reduces surprises from "bad" imports
 - better sharing and using good reference data without actual import
 - it would be possible to do some services (e.g. geocoding, search)
using this data also, license-wise it must be clean anyway.
 - easier updates of already done imports. Example: I have prepared
updated municipality borders, but cannot upload it directly, cannot
update all the data myself (too big).
 - can pre-view impact of import - Mapink import+osm tileset could be
applied. Once I have used dev.openstreetmap.org for this, but that has
really old base data now.

 Then the preferred way of any import would be through this import
layer. Direct bulk import would be possible with exceptional cases
only, if no similar data exists for the area or cannot be surveyd
(e.g. admin borders where they did not exist).

 It should not be too hard to implement, as minimum:
   * Set up separate database, instance of OSM API 0.6 using it. Some
data like users should be shared.
   * Plugins for JOSM (and others) to be able to read from multiple
OSM API servers.

-- 
Jaak Laineste