On Tuesday 04 May 2004 07:37, Gavin Jamie wrote:
> > were any at all - I've never heard anyone claim to have actually found a
> > collision) and secondly that the solution proposed could be to start
> > again with an alternative system - I do hope nobody has suggested the
> > latter.
>
> Actually a collision is likely for once 21000 people are on the system.
> Birthday paradox.
But I've never heard anyone claim to have found one... even when it was being
re-ordered it was only a possibility that some would need changing.
How?
I know the birthday paradox, and I'm not sure it applies.
People are born randomly, whereas NHS IDs were assigned with an effort to
avoid duplication, in each Executive Ctee (and successor org) area.
Each of those areas probably had a unique starter section or block of starting
parts of the "numbers" it issued.
So there may only have been 100 (number of EC/FPC/FHSA/HAs) members in the
group, or there may have been 110 000 people in each of 100 groups.
But it was not random, so the birthday paradox doesn't apply.
--
Adrian Midgley (Linux desktop)
GP, Exeter
http://www.defoam.net/
|