tag:blogger.com,1999:blog-2752058700154467226.post1919001932228338012..comments2024-03-09T10:23:37.496+00:00Comments on Bristol Cryptography Blog: Random Number Generation, RevisitedBogdanhttp://www.blogger.com/profile/13266116282208635140noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-2752058700154467226.post-76644160162942601992014-01-24T13:32:06.284+00:002014-01-24T13:32:06.284+00:00I slightly disagree with the conclusion of this ar...I slightly disagree with the conclusion of this article : <br />In my opinion Dodis cleverly showed that we don't need entropy estimation to prove a PRNG design secure. We can merely assume statements like "as long as the environnement provides me with so much entropy during this amount of time I know I'll recover"<br />It is a beautiful result and I'm looking forward having the time to read the details of the proof but from a practical point of view it could only bring a cleaner design to the linux kernel rng (which is already not nothing) because somehow we'll still need to be sure that the above statement is true which to my knowledge can only be done through entropy estimation. Anonymoushttps://www.blogger.com/profile/09360810877910199189noreply@blogger.com