What I love about bloom filters is that when checking to see if something is in a bloom filter, you can only get false positives, not false negatives. That property is just awesome to me.
Also the fact that you can use more bits per element and reduce the error rate for false positives. Using around 3 bytes per element, can bring the error rate down to 0.001%.
also the fact that you can use multiple hash functions to reduce the error rate even further is pretty nice. this is of course without incurring the overhead of extra storage space...
