Re: Random number access on the i5

It still isn't clear to me what you are trying to do. When you say "Pick 1 winner" I am uncertain what a "winner is. Could there be more than one winner, depending on the seed for the Random generator? I am guessing that you are trying to generate a unique key, but that doesn't quite fit the scenario, either. In other words, unless this is just an exercise, what is the real problem you are trying to solve? There may be another way to boil this cabbage??

Mark Villa wrote:

What isn't clear to me is what it is that he is doing.

Very....simple requirements
Pick 1 winner in a file.
You never know in advance how big a file is, but yes, you could obtain
that number at run time for the result bounds. But output bounds seems
to restrict the idea of random in ways that you could not explain. On
the other hand, without bounds you may never get a match you can use
and it could run for days.

I listed the original question( to get the responses I got) because I
found the random numbers were not so random after all using the RAND
DB function on a PC side of an application. Shame on me, I did not
realize the topic was threaded so many times here. This and other
threads confirmed there is not a source for a random number on the i5
or any computer not using a special box feeding analog inputs from
random observation. And we (I) should access an industry standard
pseudo-random number API unless it really really must be random. It is
obvious that all computer picked contest winners are pseudo-random and
so it must be ok with the rule-makers.