[Cryptography] Secure erasure

[Kent Borg]

On 09/09/2016 01:47 PM, John Denker wrote:
> All the «secure erasure» routines I've seen appear just as fatally flawed

Most of the "secure erase can't work" in this thread seems equivalent to 
"memory can't be trusted". Indeed, there is a lot of reason to mistrust 
hardware. Okay then, do you dare do any cryptography at all in your 
hardware? If so (big if), then how to do a secure erase becomes a 
sensible question.

Whether the hardware is trustworthy maybe isn't completely binary, 
systems like to cheat for efficiency sake and not because they are evil, 
which is why I like writing something more interesting than than just 
zeros. Give the hardware more to chew on to try to get it to actually do 
the chewing.

> Here's the image that comes to mind:
>     http://haha-business.com/i/_interface/epic-fail-photos-security-fail.jpg

Cute picture, and though it might be a security fail, it might not be a 
legal fail. A little like locking just a screen door: it won't stop a 
forced entry but it turns it from something that might be no offense at 
all into "breaking and entering", possibly a full crime. Legal 
deterrence does have value.

Trespass, similarly, can turn on whether notice is given. People like to 
tear down no trespassing signs and say they didn't know. Tearing down 
this would take a little time and probably hit a dollar threshold 
designed to elevate the vandalism offense. And having a path might be 
pretty clever, a way to lead would-be trespassers to the prominent 
notice, without littering signs every few meters--each then easy to 
remove and again claim "Oh, I didn't know."

-kb