Exeem > hopfully I can clear things up

hopfully I can clear things up

Exeem's biggest hurdle (though it may already be overcome, for all I know) is verified content publication. Suprnova was great because when you downloaded a file you knew it was what it said it was. A web site, such as Suprnova.org or Sharereactor.com can provide links to verified content, which is what makes P2p usable. Sure, you can search within Emule for a certain file but you're likely to see bad versions, fakes, several versions of the same file (with more than one being valid) and generally a lot of crap that's has nothing to do with what you want. When you visit a well maintained web site and you do a search, you get a very small number of results with a "standardized" version that makes for a larger peer sharing group.

Having verified content by its nature requires some kind of centralization. At the very least a moderation group is needed, and the simplest solution is distributing this verification information through a centralized service, such as a web site. The best I can imagine that would make Exeem allow decentralized, verified content would be "moderation groups" managed by private/public cryptographic keys.

The group moderator maintains a list of approved moderators that submit/approve of content. This list is nothing more than a list of other users with their public key information plus hash information to find the sub-moderator list file, which in turn themselves may be group moderators underneath this higher moderators. It's possible to give (via the group moderator's list) to give sub-moderators authority to have sub-moderators of their own based on trustworthiness.

Each of these moderators has a list of content they have "published" and of possible submoderators. The effect is once a person grabs this group admin file (which must be found via external methods, much like someone must discover a web site address by word of mouth), the admin file then queues to download moderator and sub-moderator lists. The program parses these lists to create a database of verified files, sorted by whatever criterea deemed appropriate by the moderators. As long as a moderator is considered trustworthy, they stay on the admin's list and their files are published automatically. If a moderator begins to publish fake files, they can be removed from the admin's list (and perhaps their non-fake files and their responsibility given to another moderator) to remove the corruption. This method can be applied to any P2P network, though it would greatly help to have support by the P2P client or at the very least a third party program to manage the database of verified files and check for the authenticity based on the cryptographic keys provided.

The net effect of this is that something such as Suprnova.org can be now condensed in its entirety to something as tiny as 100 bytes or so, such as an ed2k link that need only be clicked once. The public keys stored inside the original file allows the individual user to search for files generated by the same admin with a later timestamp. This updated admin file isn't even too necessary, as the moderator lists contain the actual files. As long as the moderators stay trustworthy, the user only needs to search for files generated by these moderators to update their database.

Exactly Mylon, That is the major hurdle. It will be just as fast as bittorrent. The reason why bittorrent is so much faster then other p2p's is because of its ability to prioritize uploading and downloading. When you are on a torrent you become a part of a group that only participates in that torrent.. You don't have like a share directory where any peer can just grab pieces of your files... Everybody on the torrent is focusing directly on the torrent itself. That's why its so fast. Decentralizing the tracker is only going to change the way the clients are told what and where to download the info... The main concept will be the same.

But as Mylon pointed out, content verification becomes an issue.. And the idea that you have just pointed out with the moderators is a great idea.. However.. That sort of centralizes it a tiny bit.

And last point.. The major issue i have with exeem is its just going to be a way for suprnova peeps to make money. They will likely place banner ads on it and keep the source closed. Im not down with that. We need an open source decentralized bit torrent so that it can evolve and different clients can be made.

exeem = Peace of shit. It wont be the same as bit-torrent at all. Bit-torrent = anti-kazaa mentality. I will not use some crappy ass program that all these lamers will end up using spreading crappy content with banner ad's and all kinds of shit. That's for the main stream. Bit-torrent is not main stream.

reply to Mylon
I agree. There still has to be some way to verify that you're getting a valid file, otherwise, this all turns into another kazaa/edonkey/etc... network (which is all together futile). Unless I'm just not understanding correctly how this works, It is my opinion that this will turn into another kazaa with high unreliability (never knowing if you're going to get a file or not due to the lack of tracker supervision/rules). Those "associations" out there will continue to pursue more and more people for data that they are sharing making it virtually impossible to share confidently no matter what the file is. But if they can come up with a way to verify the contents of a particular file before downloading it, that would be cool...kind of like nforce.nl but with additional file information.

reply to Donny1
actually, BT doesn't punish leachers and reward seeders. that is the biggest problem with BT. if i choke out my uploading by capping it at 3kb/s, then my download speed dramatically increase - exactly the opposite of what you are saying.

reply to slacker22

a cool research study of BT and Suprnova