J. Andrew Rogers
andrew at jarbox.org
Fri Aug 31 09:11:10 PDT 2012
On Aug 31, 2012, at 8:43 AM, Stephen Williams <sdw at lig.net> wrote:
> The point was that there are multiple cases, unfamiliar to many of us, where database correlation (of apparent fraud, unification of identities, etc.) is going to come to the wrong conclusion in many real-world cases.
The probability of this is rapidly becoming vanishingly small.
Entity analytics has become astonishingly good over the last decade and we now live in a time when the data available makes it work very well, even for people who aren't wired. We just don't identify people this way (yet) for official purposes. Nobody who really cares about identity reads the label on the papers, it is all about duck typing now.
More information about the FoRK