note 79422 deleted from function.microtime by tularis

Note Submitter: Anonymous ---- @zenofeller at zenofeller dot com: I was trying to point out that randomString() significantly reduces the "randomness-per-byte" compared to what mt_rand() and friends gives you to begin with - eg. making it more "predictable". If you're looking to generate hard-to-predict printable keys and cares about the length of the key (you perhaps need to stick in in a CHAR(32) field of some db-table), you'd be better off with either my rstr() or md5(uniqid(rand(), true)); which both will provide you with a much larger keyspace (and thus a smaller chance of a collision) in the same amount of bytes as randomString(). I should have said "more predictable" instead of just "predictable" in my previous post; you're indeed correct when you say that randomString() _will_ give you random data. It most definitely does, it just doesn't do so very effectively :)