DVD Talk Forum

DVD Talk Forum (https://forum.dvdtalk.com/)
-   Tech Talk (https://forum.dvdtalk.com/tech-talk-10/)
-   -   The Big Hack (https://forum.dvdtalk.com/tech-talk/338534-big-hack.html)

Otto 12-31-03 07:32 PM

The Big Hack

Anybody seen this site yet? It's kind of a neat concept.

As near as I can figure, they're developing a file transfer application. Kinda P2P, kinda not. Anyway, the concept is that you take any given file, break it into 128k blocks, then XOR each block with another random block from the network. Then you distrubute the resulting random blocks. Since every block can be parts of any file whatsoever, none of them can be said to be copyrighted. So in the process of getting a file from the network, you are not transmitting copyrighted material over the network. There's no way to tell if a block is part of an MP3, or a movie, or the US Constitution.

Kinda nifty, anyway. :)

cheater512 12-31-03 07:42 PM

If you want a better understanding of the OFF System please check out this first.

Gunshy 01-01-04 05:56 PM

Oof, after reading that I'm even more confuzzled :hscratch:


Deftones 01-01-04 06:05 PM

that is very interesting. wouldn't, though, as soon as you "combine" the blocks it be illegal again?

cheater512 01-02-04 02:53 AM

Technically yes but noone knows that you even downloaded a file.
How can they catch you if noone knows you did something wrong?

This is unlike current Darknets where the RIAA just needs to setup a supernode and watch everyone's downloads.

Otto 01-02-04 03:31 AM

Deftones: As I see it, assuming copyrighted files are put into the system, it's the act of putting them together again that would be the illegal part, since the blocks themselves obviously cannot be copyrighted. Since every block is a part of several different files, who could own the copyright on it? So transferring them would be perfectly legal, as you can't tell if it's being transferred for the latest chart topping MP3 or a public domain work or even the US Constitution. Any given chunk of data could be for all three, there's no way to tell.

I think it's the act of transferring files that the RIAA has been suing people for. In this case, the files you are actually transferring cannot possibly be copyrighted, so that's out. Uploading copyrighted material into the system is probably legal too, because without the key (in this case, the special URL containing reassembly data), the file cannot be extracted from the system either. It's secure even though it's distributed, as nowhere is there anything allowing you to even know where to begin reassembly of the file except with that key piece of information.

Where it gets iffy is a) reassembling copyrighted data and b) passing the key around to others. The legality of those two isn't something I can fully answer, or indeed is fully answered under existing law.

I don't know law nor long term ramifications of such a system. I simply see something technically nifty. :)

Otto 01-02-04 03:38 AM

Originally posted by Gunshy
Oof, after reading that I'm even more confuzzled :hscratch:
Gunshy: I've gotten into the development side of it and looked at some code, and here's the basic breakdown:

You take a file. You "store" it into the system. This means breaking it down into 128k chunks, and XOR'ing each chunk with two other chunks already in the system, picked at random. The resulting chunk is still 128k, but now suitably munged up, and you stick it into the system as well. A new block or set of blocks is created which holds a list of all the chunks and how to put them back together again. This is called the descriptor block(s). This gets XOR'd with two random blocks as well, and then is also uploaded into the system.

To get something back out, you need the identifiers of the three (or more) blocks that make up the descriptor set. You put them together to get the descriptor block, then get all the blocks it lists and put them together too. Then you build the file.

By combining every block with two other blocks, you basically make it so that every block is a) no longer identifiable in any possible way and b) probably a part of several different files. Thus eliminating copyright on each block itself, because each block is a part of potentially any file in the system, and thus who would own the copyright to it?

The downside is that to get a file, you probably have to download three times as much data as the file itself contains. Still, kinda neat.

All times are GMT -5. The time now is 05:58 AM.

Copyright 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.