number of unaggregated class C's in swamp?

How did you account for longer prefixes that are covered by shorter ones?

Did you limit your study to the RADB, or did you use the whole IRR?