4 ms·

^ as a self-taught dev, this is accurate. I still don't really know Big-O notation or Bloom Filters. But it's never come up as an issue.

• jedberg 6y ago
  I've found that Big-O comes up any time you're wracking your brain trying to figure out why something is taking so long, and then you realize you've created something n squared with some nested loops. And bloom filters are super cool, you should read up on them. You'll find an interesting set of problems that can be solved with them. The key to remember is that if you ask a bloom filter if something is in the set, and it says no, you are guaranteed that the answer is correct, and if it says yes, it may be there, but it's super fast at giving a no answer. So if you expect that the answer will be no most of the time, a bloom filter may be a good data structure. Here is one place it comes up: on reddit (or HN) when you load a page it shows you all the stuff you've already voted on. There are usually many things you could have voted on, and chances are you haven't voted on a lot of them. So when you ask "did user X vote on item Y", the answer is usually no. Storing the votes in Cassandra, which uses a bloom filter, makes answering that query super fast.
  • saagarjha 6y ago
    Of course, it's important to note that bloom filters by themselves can give false positives.
    • jedberg 6y ago
      Right, I said that. I said that it may return yes when it isn't in the set. But it never gives a false negative, which is key.