In message , Stefek Zaba
writes[i]
Mary Fisher wrote:

................ want to see *each* such connection evaluated
on its *merits*, balancing gains in administrative convenience
against risks to privacy.
And you really believe that will happen?

Well, good software system design practice requires it; and as far as
handling personal data's concerned, the law of the land - the Data
Protection Act - requires it too.

shamefully, two such bent public servant got a small fine and a
suspended sentence. That wasn't exactly a message about strong
enforcement.
One swallow ...

positively indicates that claims of a swallow-free land are false. One
swallow sighting per person allows a guesstimate of the swallow
population, allowing for whether they fly about a lot or generally
hide. In the case of people who are bent, or are simply helpful on the
phone (look up 'social engineering' if you're not sure what I'm on
about), there's a lot more than just one or two working away.

As pointed out above, it's not that there will be more errors
necessarily, but that the consequences of errors will be greater.
To have consequences you have to have the errors.

You're serious that your initial intuition is that the National
Identity Register's data would be error-free? Would it shake that
intuition to hear that the error rate on the Driver Vehicle Licensing
database is 30% - i.e. that a little under one in three entries have an
inaccuracy? (Would you believe me anyway?)


Having just given Mary an example of someone I know in another NG who
had major problems due to an admin error which she twisted into a
problem of him having a debt, I'm coming to the conclusion that she
doesn't seem to be able to come to terms with evidence which doesn't fit
into her established mindset


--
geoff