Sometime on Oct 31, Sameer D. Sahasrabuddhe assembled some asciibets to say:

We can see that it involves rather simple replacement of letters with alphabets in a fairly common way ... how easy is it to crack

^^^^^^^^^ letters of the alphabet ;)

such a password? Do brute force dictionary attacks (the only ones I know) take such things into considerations?

Some of them do carry out l33t replacement of letters with digits, so we get things like: 0 == [oO0] 1 == [liI] 3 == [eE] 4 == [A] 5 == [sS] 6 == [G] 7 == [tT] 9 == [g]

With this kind of replacement, the given password might have been cracked, but not as fast as if we add certain diphthong replacements that I've been thinking of (in addition to the above):

1 == w | n | un | wun | won 2 == to{1,2} 3 == th | tr | r 4 == for | four | fore 5 == [fF] 6 == [xXsS] | sx 8 == ate? 9 == nn | n X == ck Z == S

So, figure out this password: 491c81d3c163n70Xn6 (which is how long your password should be anyway). It's pronounceable, and seemingly hard to crack, but made up of dictionary words. My system should get close.

Philip