[Discuss] deadmanish login?
Kent Borg
kentborg at borg.org
Thu Feb 2 17:15:14 EST 2017
On 02/02/2017 04:38 PM, Richard Pieri wrote:
> On 2/2/2017 2:51 PM, Kent Borg wrote:
>> Does have 40-bits of entropy, that is.
> Not really:
> https://www.schneier.com/blog/archives/2014/03/choosing_secure_1.html
>
A rather terse posting, there.
You seem to be saying "Schneier agrees with me, go read this.".
And I have read that. I presume you are referring to:
"This is why the oft-cited XKCD scheme for generating passwords --
string together individual words like "correcthorsebatterystaple" -- is
no longer good advice. The password crackers are on to this trick."
It depends on where those words came from. I am not relying on some
trick, I am relying on raw combinations.
For example, let's just take 16-bits: from /dev/urandom:
d85e
Uh, oh! Did I somehow reduce the number of possible combinations because
I represented it in hex? Would I somehow reduce the number of possible
combinations of I wrote it as "dee-eight-five-ee"?
No.
In my case I am using a simple program called mnencode that was
specifically designed for making binary data pronounceable. I can run
32-bits into mnencode and get "trade-medical-episode" and I can run
"trade-medical-episode" through mndecode and get back the original
32-bits. It didn't change anything, it is just a coding.
I could tell you what those original 32-bits were--but they are binary,
to put them in an e-mail I would need to code them some other way. Would
"d7c1 271f" be okay, or does that somehow remove entropy? Octal: 153701
023437 okay? mnencode: trade-medical-episode? Oh, no, you don't like
that one.
It's just another coding. But it is easy to remember and easy to type on
an ASCII keyboard.
If someone wants to brute force it someone is going to have to run 2^32
combinations to try them all.
How do I get 40-bits in my examples? To make it extra good I prepend
two-hex digits: f1-sultan-joker-editor. (Also because some stupid
systems will silently truncate passwords, it packs a little more entropy
at the beginning.)
-kb
More information about the Discuss
mailing list