Puzzled about random initialisation

Heather Coppersmith me at privacy.net
Fri Jul 9 18:10:39 CEST 2004


On Fri, 9 Jul 2004 08:43:11 -0400,
Christopher T King <squirrel at WPI.EDU> wrote:

> On Fri, 9 Jul 2004, Robin Becker wrote:

>> thanks for all this it makes things a lot clearer. Using
>> time.time() makes it harder, but still not impossible to guess
>> what values could have been generated.  The application should
>> probably be using /dev/urandom to seed generators.

> Perhaps you could try something like time.time()*os.getpid()?
> This will give different processes starting at the same time
> different random seeds.

That (probably) doesn't help:  If I know what time a process
started (down to the kernel-clock tick), then I (probably) know
that process' pid, too.

Regards,
Heather

-- 
Heather Coppersmith
That's not right; that's not even wrong. -- Wolfgang Pauli



More information about the Python-list mailing list