On Sat, Aug 24, 2002 at 01:10:03AM +0200, Benja Fallenstein wrote:
Hi,
I think one thing we should write an article about is the new diff
mechanism. The idea is terribly simple (even if the implementation needs
some careful work in order to be really robust), yet I don't know of it
being used elsewhere (does someone else here?)-- and it is quite
interesting, because it marries the permanence and security (if not
completely the robustness) of Storm with the efficiency of storing
diffs, storing backward diffs, and transmitting only the current version
(when downloading the current version of something from the 'net).
Should this be before or after urn-5? Probably we should limit to
actively writing one, and having the others in the queue.