[Q]:Generate Unique ID's

[achrist at easystreet.com]

"Martin v. Löwis" wrote:
> 
> Steven Taschuk wrote:
> 
> > When you say "no exceptions", do you mean "some very small number
> > of exceptions, close enough to zero for all practical purposes"?
> 
> He meant what he said: forever unique, no exceptions. That, of course,
> means that he has fallen to marketing hype. What the algorithm really
> guarantees is "close enough to zero for all practical purposes".
> 

That IETF doc that was mentioned a ways back in this thread did mention
that there would be some kind of a rollower in about 1400 years, so
forever isn't quite what it used to be, but it never has been.

My back-of-the envelope calculations, done without any envelope, 
suggest that if you could generate truly random p=0.5 bits,
then if you generate a trillion row database every second for
a period of time equal to the current age of the universe, and give
each row a 128-bit random ID, it's about even money whether or not
a duplicate key error would occur once over all those databases
over that time period.  

"close enough to zero for all practical purposes"?

Al