W
Wolfgang Draxinger
Joachim said:But that then would create very predictable 'random' numbers.
That's not the point, if the OPs intention is to resolve network
collisions. He'll be fine if the method ensures, that no two
stations are likely to start transmission at the same time.
Given a good PRNG (like mersene twister) it's sufficient to start
each PRNG with it's own seed. In the OPs applications it even is
no problem if single PRNGs get reinitialized by a system reset.
OTOH he's implementing it on a DSP. I bet there's some hardware
attached, that delivers some entropy. Like the lowest
significant bit of an A/D conversion.
Wolfgang Draxinger