2006-03-13 · in Ideas · 148 words

The infringe program should exist: given a collection of data files and a source file, it should document how the source file can be constructed from fragments of the data files, and should do so -- thus proving that any digital file can be construed as a copyright infringement of another.

Piet Delport pointed me at Monolith, which has much the same purpose, but takes a different approach: it generates an random data stream and XORs it with a source file to give two files, neither of which are useful on their own, but can be combined to give the original work. There's also the Owner-Free Filesystem, in which each data block is used to reconstruct multiple files (which in itself is an interesting idea).

There's some discussion of the important idea here -- that information content isn't everything in copyright law -- by Matthew Skala: "What Colour are your bits?".