On 2004-08-13T09:07:44-0400, Adam Back wrote: > Yes. It's just the 21st bit of a 20 bit stamp with type B would > _have_ to be 1 whereas now (type A) it can be either 0 or 1. Yes, but I think you miss my meaning. The new scheme (B) is indistinguishable computationally (or wrt entropy or probability distributions) from a 21-bit stamp. You don't gain any advantage. You could specify any pattern you wanted for collisions. For a 20-bit stamp, some could be: 01010101010101010101 10101010101010101010 00000000000000000000 (current) 00010000000100000001 What you propose is 000000000000000000001, which is a 21 bit pattern, therefore it is the equivalent of a 21-bit stamp under the current validation scheme. > On Fri, Aug 13, 2004 at 12:46:47PM +0000, Justin wrote: > > On 2004-08-13T05:47:07-0400, Adam Back wrote: > > > A. (as in v1) value = claimed_bits if claimed_bits <= measured_bits > > > > > > B. or value = claimed bits if claimed_bits == measured_bits > > > > > > What happens to the std. deviation -- is it better with type B or type > > > A? > > > > A mandatory 20-bit stamp is no different than a 21-bit stamp under the > > current scheme, right? The 21st bit just has to be 1 instead of 0. > -- "When in our age we hear these words: It will be judged by the result--then we know at once with whom we have the honor of speaking. Those who talk this way are a numerous type whom I shall designate under the common name of assistant professors." -- Kierkegaard, Fear and Trembling (Wong tr.), III, 112