Hacker News new | past | comments | ask | show | jobs | submit login

1. It has O(1) lookups (it has to probe at most 2 constant-size buckets) and O(1) amortised insertions (the probability of needing more than k swaps to find an empty slot decreases exponentially with k).

2. It's geared for low false positives, 3% or lower.

3. It does have false positive on lookups: signatures are lossy.

4. It has comparable space requirements to bloom filters for the same error rate and set (not universe) cardinality.




Amortized O(1) is very different from O(1). But I do appreciate the corrections. It seems that I missed or misread some of their paper.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: