> I'm inclined to think that the simplest solution is the best. That is to > say the one that divides the _appearance_ of the different content > addressing types as little as possible. We should also keep in mind the > chances of a collison are equal to 1 in 2 to the size of the hash we are > using, i.e. pretty damned remote.
I am confused as to whether you are supporting or disagreeing with my comments. :-) You have a good point about the remoteness of collision. The system assumes no collisions will occur. However, if we keep dividing our key space into smaller and smaller subsections, collisions become more likely. They will still be very unlikely, of course. If we store different types in different tables and access them with different messages, then the chance of key collision stays the same as we increase the number of message types. _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
