[tor-talk] Let's make Onion Addresses Meaningful To Humans

[Ahmed Hassan]

Well,..

The according to the onion wiki, the length of the onion address is 80
bits.

The largest number the onion address can get is:
1208925819614629174706175 

That's because FF:FF:FF:FF:FF:FF:FF:FF:FF:FF is the largest number
(unsigned) in hex for 80 bits key length.

If we assume we have a dictionary that has 50K words, the maximum number
of words in the onion address will be 6 words.

Wolframa link:

http://www.wolframalpha.com/input/?i=1208925819614629174706175++convert
+to+base+50000

For a 100K words dictionary, it will be 5 words

http://www.wolframalpha.com/input/?i=1208925819614629174706175++convert
+to+base+100000 

The average length of a word in English dictionary is 5.1 characters
according to this http://www.wolframalpha.com/input/?i=average+english
+word+length 
 

The larger number of words in a dictionary we use, the shorter the
address we get.


The end result will be something like this:

xxxxx-xxxxx-xxxxx-xxxxx-xxxxx.onion 

 

On Fri, 2012-02-24 at 15:03 +0100, Andreas Krey wrote:
> On Fri, 24 Feb 2012 13:36:45 +0000, Robert Ransom wrote:
> ...
> > Which languages do you want us to ship a dictionary for in every Tor
> > client?  (Please specify the exact dictionaries you want us to use as
> > well.)
> 
> Left as an exercise for later.
> 
> > How large are these dictionaries (in bytes)?
> 
> The last one I tried is 16655 words, 91445 bytes (null-terminated strings).
> 
> ...
> > Have you tried this using the actual dictionaries that you want us to
> > use?  Are the resulting addresses really memorable?
> 
>   goric-edema-Alces-rune-pan-coost
>   feign-crig-plane-tret-balli-chela
> 
> => Slightly.
> 
> (I admit that I did not look up what base the *.onion names are
>  in, so the number of bits and thus words may be off.)
> 
> > How long are the
> > resulting addresses?
> 
> Longer, of course.
> 
> > Can they be entered into a computer as
> > efficiently as addresses in the current format?
> 
> Depends on the meaning of 'efficient'. Being longer it's more obvious work
> to type, but...
> 
> > Can a human proofread
> > addresses in this form for errors as efficiently as addresses in the
> > current format?
> 
> ...easier to proofread or spell over the phone. But then, the proofread
> part may be eased by adding a few minus signs into the usual onion names
> just as well.
> 
> That said, the real problem is deployment of anything like this.
> 
> Andreas
>