On Sun, 5 Mar 2000, Tim May wrote: > This is starting to look like a variant of the old "we share a large set of > numbers--"one time pad"--and then the message includes a set of bits > specifying an entry point into this set of numbers." In this case, the > shared PRNG is not only security through obscurity, which doesn't work, it > is also dependent on the entropy of the seed. Which is probably a whole lot > less than the entropy implicit in our usual several hundred decimal > digit--equivalent work factors. > > And it vaguely resembles some of the many variants of the "pseudo one time > pad." The zero knowledge stuff seems ancillary, an afterthought, to use the > PRNG and seeds as the main carriers of information. Tim seems to have deeper insights into the psychology of my design decisions than I do. I foolishly assumed that I was proposing a perfectly valid completely technical method of thwarting a completely technical attack, but it seems my freudian subconcious had a hand in things. For those of you who are interested in actually reading and evaluating algorithms, instead of just commenting on words used in them, the algorithm mentioned earlier this thread is at http://www.gawth.com/bram/essays/simple_public_key.html I'll be changing that page to reflect the modification I mentioned earlier this thread in a bit. -Bram
