On 22/06/2011 11:24 PM, clipka wrote:
> Am 22.06.2011 23:04, schrieb Darren New:
>> On 6/22/2011 13:23, Orchid XP v8 wrote:
>>> looked specifically at functions which are cheap to compute but still
>>> look
>>> "reasonably random".
>>
>> Yes. They're called "non-cryptographic hashes". ;-)
>
> ... but beware - there are some pretty simple algorithms out there that
> probably don't suit your needs. E.g., "hash(data) := data mod N"
> qualifies as a hash function, but probably not as "reasonably random".

Yeah, a hash function for a hash table just needs to avoid collisions if 
possible. The results don't need to "look random" at all.

Post a reply to this message