demerphq

6 ms·

1. 1.▲
   by demerphq 9y ago

   No you are right. Modern perls use the BSD code in File::Glob.
2. 2.▲
   by demerphq 10y ago

   I am only repeating what my tests say. I was surprised frankly, given FNV's reputation, to see it fail so many tests. Also I put a lot of effort into improving the quality of testing that smhasher does, so there are many hashes that wo
3. 3.▲
   by demerphq 10y ago

   Thanks for the correction!
4. 4.▲
   by demerphq 10y ago

   BeagleHash IMO is less interesting than Zaphod64 and StadtX. With all of them I used genetic algorithm techniques to search the space of possible minimal mix functions to find ones that minimized error from the strict avalanche test. I will
5. 5.▲
   by demerphq 10y ago

   It is only small in older hashes. Modern hashes tend to have a state larger than both the seed and the input block. See: https://github.com/demerphq/smhasher/blob/master/HashAnatomy... It actually turns
6. 6.▲
   by demerphq 10y ago

   You are welcome. I wanted add another point. rurban does have a bit of a point. The hash functions you referenced in your article are generally old and low quality, and slow. FNV is neither a good hash, nor is it fast. FNV_yt is very fast,
7. 7.▲
   by demerphq 10y ago

   I think you have confused "ideal hash function" and "perfect hash function". A perfect hash function is one that maps a finite set of keys into the same number of buckets with no collisions. They have to be tailor constr