It does seem a bit odd to me that it uses a 9B and 7A rather than just identifying the 3 7A's independently since this would be simpler. My guess would be that this is related to performance and also not identifying the same cluster multiple times. It may be worth bench-marking this method in the future though.
10B is not wonderful but better than it was.
It does seem a bit odd to me that it uses a 9B and 7A rather than just identifying the 3 7A's independently since this would be simpler. My guess would be that this is related to performance and also not identifying the same cluster multiple times. It may be worth bench-marking this method in the future though.