As for the "homeland security" argument, do you have any idea how much
raw data they'd have to sift through before coming to something
appearing to be a password? This really wanders into the realm of "only
A lots, but finding sequence of user/password from the same source is
not that hard in huge data. Just logical if you think about it.
============
garbage data...
user<cr>
password<cr>
garbage data....
user<cr>
password<cr>
garbage data....days worth of it...
user<cr>
password<cr>
garbage data....
So, only need to parse to find two lines that repeat itself a few times
in a few days and you have high probability to have user/password
combination.
Isn't it the most majority of users use user name as single word and
password as single word, may be with weird characters in them granted,
but still... space???
So, process one line at a time and discard all that have more then one
word, if you find two consecutive lines with one word only, save them
and see if you find more instance of it.
Sure look quick to process to me.... Just stream it live and save single
double sequence of words and that's about it you have to do.
So, look for sequence of single word in two lines that repeat itself and
you have high probability to have it.
In the end, if you don't trust the keyboard, don't use it.
Really, not as hard as one might think!
Daniel
