Almost entirely off-topic rant - apologies in advance:::

[Brian Morrison]

On Fri, 26 Jul 2002 21:35:47 +0100
"Dave Howe" <DHowe@hawkswing.demon.co.uk> wrote:

> Brian Morrison wrote:
> || Well, not to pour cold water on your theory, but in the same
> article|| it does say that MLS state that the software does not store
> || fingerprints in their original form, it instead stores an encrypted
> || version of the 'points' in the fingerprint so that reconstructing
> || the fingerprint is supposed to be fairly hard. I can't say that I
> || know what the encryption used is, nor how the fingerprint is
> || distilled into recognition points, but this is a little different
> || from the gist of your posting.
> IIRC though - that is exactly how the matching library in the national
> computer works - by looking for significant points, making a map of
> them, and comparing maps.
> once they had it down to one or two possibles, they could probably get
> a warrant on the strength of that to get the real fingerprints.

OK, I stand corrected then, and I don't like the MLS method much, even
less so because it seems to be completely unnecessary.

-- 
Brian Morrison

bdm at fenrir dot org dot uk