7,019 reputation
1711
bio website
location Ontario
age 40
visits member for 3 years, 8 months
seen May 29 at 21:12

Feb
27
revised Which hashing algorithm is best for uniqueness and speed?
edited body
Oct
28
awarded  Yearling
Oct
24
comment Which hashing algorithm is best for uniqueness and speed?
@yaccz i chewed whatever gum i found laying around. This one i think it was
Oct
17
comment Which hashing algorithm is best for uniqueness and speed?
@JimR i could, but it's Delphi, using various code files of helper routines and classes that i've collected over the years. You're better off to start over in C# or something. The real work was implementing the algorithms. It would be a lot of pain to cull down the relevant files into the required routines; and even then it was only designed for Delphi 7 - using an ide from this decade would require code-rework. In other words: no. :)
Aug
14
awarded  Good Question
Dec
23
awarded  Popular Question
Dec
22
awarded  Announcer
Nov
27
comment How to implement a safe password history
@Mathew Foscarini The downside of generating a password for the user, and forcing them to use it, is that now it has to be written down - completely negating any of the security you were hoping to gain.
Oct
28
awarded  Yearling
May
25
comment Which hashing algorithm is best for uniqueness and speed?
@LouisWasserman CityHash has no 32-bit version; so....no you can't see it covered :)
May
25
revised Which hashing algorithm is best for uniqueness and speed?
added 15 characters in body
May
25
revised Which hashing algorithm is best for uniqueness and speed?
Added mention of old thing article
May
4
comment Which hashing algorithm is best for uniqueness and speed?
@maaartinus On the downside, as a practical matter (non-cryptographic) hash functions are usually used to implement a hash table. Hash tables are usually keyed by a string. Trying to pick our the correct 4 bytes of a guid so you can convert it to an integer, so you can convert it to a string adds needless complexity. And it also reduces the usefulness (e.g. if i want to find the item for CustomerGUID "{B5225B82-F3DE-41E7-A70B-91F5AB1F9709}")
Apr
28
awarded  Great Answer
Apr
28
revised Which hashing algorithm is best for uniqueness and speed?
Added link to mention of SuperFastHash having poor collision resistance
Apr
28
comment Which hashing algorithm is best for uniqueness and speed?
@maaartinus i've not hashed random data, nor would i see a value in doing so. On the other hand i do hash type 4 uuids (uuid's not tied to a MAC address). Storing customers in a hash table by their unique key does make sense. Thinking one can just take the first 32 bits of a 128-bit guid, and assume it has better collision resistance characteristics than an actual hash, is about as smart as me trying to invent my own hash algorithm. p.s. only 122 bits of type 4 random uuid are random.
Apr
28
comment Which hashing algorithm is best for uniqueness and speed?
@Secure i've tested known inputs where available. The implementations may be wrong; but they're the implementations i'd be using - so the results are accurate for me. You're free to test the algorithms yourself. In practice i'll likely never see a collision.
Apr
27
revised Which hashing algorithm is best for uniqueness and speed?
added 274 characters in body
Apr
26
revised Which hashing algorithm is best for uniqueness and speed?
Everything i said about crc was mistaken
Apr
26
comment Which hashing algorithm is best for uniqueness and speed?
Now i'm wondering if randomness is a useful thing at all when looking for a hashtable implementation. Given virtual memory paging, cache lines, L2 cache - wouldn't i want the first 100 items i hash to fall into the first 100 slots of my list? Collisions are bad, and cause a hash-table probe, and speed is tremendously important. But randomness - not so much.