Ian Clarke wrote: > There are a few small things standing between us and our 0.5 release:
> FProxy :- > It is looking good, it does the basics, along with security, and > apparently it even supports split files although I have yet to test this > personally Not looking so good for me. First I had trouble because I didn't have the javax servlet stuff. When I included that I got still troubles as it couldn't get some German localization files for the javax stuff. I will not go through the hazzle and offer localization files for all kinds of Freenet installers.... Do we really have to rely on that rather external stuff? > Windows Installer :- > Strangely, the installer that has been available from "Techgrounds" > works fine for me, but the installer in the snapshot release didn't work > for me when I tested it on W2K. I will get a better bug report, but I > am sure it is a minor problem. David's configuration tool is great, > although I think it still needs some usability work which won't take > long. Sure, it needs some minor tweks, but works basically fine. Please send me a more detailed bug report, as I had only good feedback so far for the new Win installer (besides the included FCPProxy being broken, which will be fixed in the next snapshot). > Seed nodes :- > The current mechanism for providing seed nodes is probably a security > risk. It will be very difficult for users to find seed nodes > themselves, but relying on a centralized list is undesirable. > The optimal solution is probably to set up an obscure node which > "harvests" seed nodes from the network, and to provide a URL which > allows access to a list of some of these harvested nodes. The harvester could either change the seed.ref file frequently, which means we don't have to change anything on the client/installer side. > The more conservative among us would rather that users are instructed to > find seed nodes themselves, but I think that this would make > installation too difficult for 99.9% of our potential userbase (hell, I > would have trouble finding a list of seed nodes without asking on > #freenet or something, which is hardly a solution for every user). One valid point is security, the other is Freenet architecture: besides security reasons it would be really bad for Freenet to give everybody the same starting points and have them initialize the same way: -> very centralized freenet -> very predictive behavior of new users -> few, easy to attack starting points -> monocultures are never really robust _______________________________________________ Devl mailing list Devl at freenetproject.org http://lists.freenetproject.org/mailman/listinfo/devl
