Getting a *more* random value
Wed Mar 29 10:18:21 CEST 2006
You could always bury the first value?
Also you could probably seed the pseudo random generator with the
results of another pseudo random sequence? Basically just ensure each
process has a different PRNG seed and you should be golden.
Anyone want to speak to quality of the PNRG implented in uniform? I
know the man page cites a journal entry, but I really dont have access
to it. Clearly its good enough for generating test cases, no?
On 3/28/06, Pupeno <> wrote:
> I have an application where new processes are launched which calculate one
> random number between 0 and 512 and then do something with that number and
> I am currently using random:uniform(513) - 1 but that gives me the same random
> value for all the processes for each run of the application. I see that
> consecutive calls would get other numbers but I only need one.
> How am I supposed to get a number that is not the same on every process ?
> Thank you.
> Pupeno <> (http://pupeno.com)
More information about the erlang-questions