[Gllug] NHS opt out
Nix
nix at esperi.org.uk
Sat Jan 27 20:23:37 UTC 2007
On 27 Jan 2007, Mike Brodbelt outgrape:
> Nix wrote:
>> If you find a biometric that works and has affordable and convenient
>> readers
>
> <snip>
>
> Define "works". I tend to think people get hung up on perfection here -
Well, yes, of course, because of the statistics of rare events.
> although I confess to not knowing the details, I'm sure implementing a
> biometric with measurable superiority of current forms of identification
> should be possible at a reasonable cost level. It doesn't have to be
> perfect, just better.
Unfortunately, as long as the vast majority of people are not fraudsters,
it *does* have to be nearly perfect, or the rare true positives will be
vastly outnumbered by the false positives, because even though the error
rate may be low in relative terms, when multiplied by the vast numbers
of true negatives, it'll still lead to a large absolute number of false
positives.
(This is basic statistics, although also somewhat counterintuitive if
people don't know about it. I think Ben Goldacre had something about it
in the Guardian's Bad Science column a while back)
> It would be good if people got used to the idea
> that the expectation of a perfect solution is unrealistic, and that they
In that case, the expectation of ID cards that will magically spot, um,
whatever the hell the threat is this week, is also unrealistic.
What they want isn't an ID card, anyway. Looking at the things they've
said ID cards will defend against, what they want is an I Am A Bad Guy
card. And you'll never get the bad guys to carry one of *those*.
--
`The serial comma, however, is correct and proper, and abandoning it will
surely lead to chaos, anarchy, rioting in the streets, the Terrorists
taking over, and possibly the complete collapse of Human Civilization.'
-------------- next part --------------
--
Gllug mailing list - Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug
More information about the GLLUG
mailing list