Re: 63/64 bit version of Delphi's function random( aRange : integer ) : integer; ?

"Joseph Ashwood" <ashwood@msn.com> wrote in message
news:z7Y4g.64043$_S7.12057@newssvr14.news.prodigy.com...
> "Skybuck Flying" <spam@hotmail.com> wrote in message
> news:44544b08$0$710$5fc3050@dreader2.news.tiscali.nl...
>> Seems to me you just throwing some numbers together and then modding
>> them...
>
> Since you were looking for a complete 2^63 space, it will have precisely 2
> values that mod to each result in each 64-bit space.

I don't think so.

Your method is just as flawed as the rest.

First of all you do not provide a parameter to the three random calls.

Second of all you seem to forget or not know how random works in delphi.

It can only returns 31 bits, not 32 bits.

Finally it will be ass slow too.

> Since XOR is entropy maintaining the processing I gave will not introduce
> biases, and can only serve to maintain or increase the entropy density.
> Across 2^63 the distribution will be at least as smooth as the original
> Random() function.
>
> Now that you have a (apparently heavily flawed since you've posted
> countless adjustments of it) design, a smooth design, a simple design, and
> at least one other that I don't immediately recall it's advantage, might I
> suggest that you stop babbling?

Yes,

I am done with all these flawed methods.

Bye,
  Skybuck.
Received on Mon May 1 02:06:35 2006