Lee writes: > I wouldn't count Twofish out yet; in fact I'd be surprized in Rijndael > won by itself. Twofish will also have better performance in Java. > > But I think the correct answer to the question of "which algorithm > do we implement now" is "XOR"; i.e., let's make sure we can plug one > in when needed (perhaps even have code for serval available to plug), > make sure they all work, but not make the decision yet. There's no > need to paint ourselves into a corner yet until everything is working, > and by that time AES will probably be out.
The AES process required all entrants to submit a Java implementation coded to a specified API. If we made our crypto interface call the crypto code using that API, we could plug in any one of the AES candidates (including the eventual winner) without changing a line of code, just changing which module we link with. Hal _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
