2 - warn/skip when node was found but siruta:code changes; possibly we
3 should ignore the old node in the situation and create a new one
4 - fuzzy match (lexical proximity with î closer to â than others)
8 - check consistency and enforce it on the Romanian data
9 - refs should be respect regex: (D[NJC]|[0-9]+[A-Z]|A[1-9])
10 - name should not be refs
11 - ways with DNxx refs and Exx/Exxx int_refs should be trunk
12 - ways with DNxx refs and no int_refs should probably be primary (or missing int_ref)
13 - DJxxx - secondary, DCxxx - tertiary
14 - names and corrections - see repo osm-helpers, file 'rules/inuse/ro.osmrules' in branch 'rules'