/**
* Splits osm data file into nodes, ways, relations files
*
* Relations and ways in osm doesn't have geometry and coordinates
* they only have a references to nodes.
* (Relations have members which are other relations, ways and nodes)
*
* So after we have filtered out which relations and ways we need by tags
* we should load nodes underneath.
*
* Data stored in linear way (as a list of xml nodes).
* So now it's faster to split data into three
* files neither read whole file for many times.
*
* Actual job is performed in
* me.osm.gazetter.split.Split
*
* Next slice phase: @see me.osm.gazetter.striper
* */
package me.osm.gazetter.split;