[extropy-chat] Many eyes

[Dan Clemmensen]

John Kozubik wrote:

>Dan,
>
>On Sun, 10 Jul 2005, Dan Clemmensen wrote:
>
>  
>
>>If every Londoner with a cell-phone camera had taken a 10-shot panorama
>>at the time of the bombing, we would almost certainly have a picture of
>>at least one of the bombers.
>>
>>To speed the analysis, we should also add a volunteer analytic
>>infrastructure. If every relevant Londoner made panoramic pictures,
>>there would be far more pictures than police analysts could process
>>quickly. But each photographer could add the pictures to a distributed
>>database, and each photographer (plus innumerable volunteers) could do a
>>preliminary analysis.
>>    
>>
>
>
>Ok.  But how do you distinguish fake data and analysis from real data and
>analysis ?  What if the criminals involved created hundreds or thousands
>of CGI photos of real and/or fake persons at the scene and submitted them
>as part of the broader public response ?
>
>Such a barrage of misleading information could hamper the investigation at
>best, and falsely redirect it at worse.
>
>Comments ?
>
>  
>
MD5 signatures. Each photographer can either join a pre-existing chain 
of trust. No need to
actually divulge the photographer's name to all, but the police can find 
it if an analyst
decides something is wrong.

Similarly, analysts are identifiable if necessary, and analysis is 
overlapped, so many analysts
will cross-check each other.