Owner-Free File System

[Verbal Mike]

http://wiki.offdev.org/Main_Page
This is either an elaborate mindfuck or something that will change the whole filesharing world.
Took me a bit of reading on their wiki to wrap my head around it, but I think I get it. Basically, the system breaks down any file you put in into 128KB blocks, but instead of these actually being parts of the file, they are random chunks that form the original blocks only when put together correctly using XOR. The blocks are distributed randomly across different servers. They are reused for different files (as one example puts it, if you have blocks A, B, C, A+B could be part of oopsididitagain.mp3 and A+C could be part of rootsrockreggae.mp3 - block A itself is meaningless) and the only person potentially prosecutable for copyright infringement is whoever actually assembles a copyrighted file off of the OFF system. The system is designed so you can "stream" distributed files without needing to create your own copy (it is a virtual file system.)

So basically, if this ever becomes the predominant P2P technology, it will be *impossible* for ISPs and RIAAs to catch people using it for piracy. Not difficult, actually impossible. The only way you could do it is by actually checking what is on people's hard-drives, and with the proliferation of high-grade HD encryption, that will soon become a rather impractical thing to do.

[Cramulus]

PIRATES WIN AGAIN!

[PeregrineBF]

Ok, a few problems: Wouldn't the URL be considered a tool for circumventing copyright protection under the DMCA, and thus mean you are still liable? There's no high-anonymity routing like Freenet. If this could be implemented within Freenet it would be great, and really much closer to actually "impossible" to sue. There are ways one could be caught for using OFF. While it provides plausible deniability that you wanted to assemble a specific file it does not provide deniability that you downloaded the parts that make up that file, and the individual set of blocks used in a file is likely not going to be small enough that they will ALL be shared among multiple files. It's a neat trick, but making Freenet faster and then adding this on top of it seems like the best bet.

[Verbal Mike]

Well the point of this is it's a "brightnet" - nothing has to be hidden because it's all perfectly legal. Like I said, either this is an elaborate mindfuck because it's impossible, or they've figure it out.
I think the idea is the URL does not let you download the file, but rather just access it. It's a virtual file system. So the URL itself is just a pointer to a file that's technically not there. And if you copy it, well, the only way they can find out is by checking your HD. The key is to think about it as a virtual file system.

[PeregrineBF]

I know what it does. The data is legal, the URLs aren't. Legality of the data is questionable, but probable.

[Verbal Mike]

The URL is only illegal if the data is copyrighted. You can only find out if the data is copyrighted if you have the URL. Seems safe enough to me.

[PeregrineBF]

Well, yes, assuming copyrighted material.
This actually transfers liability. With a torrent the URL indicates where the data is, but is not copyrighted in and of itself. Thus, the tracker (host of URLs) is not liable, but the downloaders are. (NOTE: US and AU laws (may, in most cases do) make the tracker liable as well. Other countries may do the same.) This system stores the copyrighted data in the URL, making the tracker liable, but not the downloaders.
If you say no copyrighted data is sored, you're wrong. It's been ruled that a format-shifted form of some data is still copyrighted, you can't distribute music just because you changed it from CD to MP3 form. All possible encodings of any work are copyrighted when the work is created. Since the URL encodes the work (well, how to re-assemble it from the data blocks) the URL is copyrighted.
It's easy to find if the data is copyrighted: download it. If the MAFIAA downloads it, and finds a work they own the copyright on then they can simply sue the URL provider. The users are safe, but the central server is screwed. Napster also had a central point of responsibility, and look what happened to it. What is needed is not more centralized stuff, but more decentralization and anonymization. This provides neither, though it could be modified to provide them.
A simple way is to combine it with bittorrent. Create a normal tracker webpage and enable DHT (decentralized tracking), that tracks OFF URLs. Now everyone in the DHT swarm is liable, but that's a big swarm. It's impossible to sue them all, just as it's impossible to sue all the bittorrent users. (actually, just incredibly impractical.) The problem, of course, is that this provides only added complexity, and no more safety for users.  The system needs onion-routing or some similar method added in to provide plausible deniability to whoever is hosting the OFF URLs. This will make it slow.
Now, I think this is stupid and a bunch of bad laws, but that doesn't mean it isn't the way things work. Sadly.

[Verbal Mike]

I think the whole point is there is no URL provider and there is no central server. I'm guessing the only person who has the URL is one who uploads the data - and without the URL it is entirely impossible to make any sense of the data.
Anyway, your critique would be better placed on the OFF Wiki, where people who know how it works could respond, and possibly learn from what you say.

[PeregrineBF]

Yeah, should post there.
If only the uploader has the URL, no one can download it. This defeats the purpose, you could just use a normal hard drive.

[Voodoo]

PIRATES WIN AGAIN!

sexy pirates always win.


is it just like packeting?  Thats what I am getting.

[Cain]

I'm sure I've seen that chick dressed as a witch as well.

Name check?