Encryption and risk

[Matthew Gillen]

On 10/06/2009 09:43 AM, markw-FJ05HQ0HCKaWd6l5hS35sQ at public.gmane.org wrote:
>>> There will always be an exploit. If not through the encryption algorithm
>>> itself, through the implementation.
>>
>> not exactly, that makes it sound pointless to strive for improvement
> 
> No not at all, just because it is "true" that there will always be an
> exploit, something is "safe" until it isn't. When it isn't, you fix what's
> broken. Then you have another period of "safe." That mind set really gets
> to security people, but it is a zen thing. We try for the best, but humans
> are imperfect thus everything we make, no matter how good, is imperfect.

Even more blasphemous is the notion that all you're really trying to do is
raise the difficulty bar high enough to keep the bulk of your problems at bay.
 A perfect security "system", even if it could exist, would always be
undermined by the humans in the loop (case in point, the failure here was not
in the security mechanisms/system, but in intentionally allowing the attacker
to bypass them as a measure of good faith:
http://www.nytimes.com/2009/08/29/world/middleeast/29saudi.html?_r=1
)

Someone with the right tool can break into your locked car in 11 seconds (I
timed the guy from AAA once).  But locking your car doors is still usually
worthwhile.  Likewise, homebrew, broken encryption provides some defense
against casual attackers.  (and since it took me so long to write this, dsr
made my point already, albeit with a harder edge).

Matt