I reviewed the new batch of changes on the cold boot attack
documentation. We're almost there!
That's still in doc/cold-boot-attack.
A few comments on what I did:
- « When using a computer, everything done during a session is stored in
RAM ». I find the use of "everything" and "stored" very ambiguous here.
I tried to find a more precise expression with « all the data
manipulated is written temporarily in RAM »
- I felt your doc was missing a sentence to explain that the data in RAM
always disappears automatically after several minutes. You took that for
granted I think. So I added:
« After a computer is powered off, the information in RAM disappears
rapidly, but it can remain in RAM up to several minutes after shutdown.»
- Use "may" to denote permission, and "might" to denote uncertainty.
From the Chicago Manual of Style:
« May denotes permission {you may go to the movies}. May sometimes
connotes an uncertain possibility {you may find that assignment too
difficult} and often becomes might {you might find that assignment too
difficult}. »in
- I flipped around the two sentences describing attacks, and made them
more active by including the figure of the attacker. There is no cold
boot attack without attacker so let's make this explicit.
« A physical access to the computer you used may enable to recover
everything you've been achieving during the session. » becomes « An
attacker having access to a computer before it disappears could recover
important data from your session. »
« Moreover, a physical access to the computer you're using Tails on,
enable as well recovering data from RAM. » becomes « Moreover, an
attacker having physical access to the computer *while Tails is running*
can recover data from RAM as well. »
- I used "data in RAM" instead of just "RAM" in several places: the RAM
itself is not erased (hopefully) but the data in it.
- I use '-' instead of '*' for list because it wraps better with vi :)
- I used a bullet list to list the two bugs that affect this feature.
And added a <div class="bug"> to make it clear that this should be
worked on.
- I'm not sure about the use of "standard" in "standard procedure for
data recovery", shall we use "common" instead?
