On 25/12/15 06:19, Ineiev wrote: > I assume the amount of entropy is what really matters. for instance, > if on every next step you are free to choose any of 4 random words > taken from 60000-word dictionary, you may put it in a grammatically > correct form[*], then you must get a certain entropy per step.
Yes, however, this characterization seems mathematically incorrect. Let's assume one in four words in the dictionary fits the grammar. I hope this concurs broadly with what you assumed. Rather than pick four random words of the full list, and then pick one of those, you pick one out of a quarter of the wordlist size. So that's 2 bits per word you're losing, a lot more than if you were free to pick one of four random words. And there is a lot more structure to the sentence given by Matthias than just its grammatical soundness. HTH, Peter. -- I use the GNU Privacy Guard (GnuPG) in combination with Enigmail. You can send me encrypted mail if you want some privacy. My key is available at <http://digitalbrains.com/2012/openpgp-key-peter> _______________________________________________ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users
