Blind Publication: A Copyright Library Without Publication or Trust

摘要

Good morning. The title, Blind Publication, may not convey too much about what I’m actually going to talk about: it’s a simple idea which I hope will have some merit, and it really started when I started thinking about copyright. Now all sorts of people are thinking about copyright, digital rights, etc, but they’re looking at movies, images, all the sorts of stuff that Hollywood thinks is valuable, and I thought, well let’s look at something much simpler, let’s look at plain text and see if we can get any good ideas by looking at a much simpler situation. So that’s the area that I want to look at, but I thought we’d start with a picture as an analogy of what I want to do. Essentially what I want to do is to look at these three pictures, but you can’t see them properly, they’re partially hidden, they’ve not been fully published, and then to identify that two of these pictures are similar in some sense, and one is different. And you get kind of a clue that there are some features in the first picture which are also present in the second. They’re clearly very different, they’re clearly not the same picture, but there’s some commonality: there’s a feature (the smile) in image 1 and image 2 which is common. Now what I’ve discovered, through looking at text, is that perhaps if can we extract features from copyright documents we may be able to find a way of comparing documents without actually publishing them. This is the basic idea. Essentially I want to leak just enough information from one copyright document so that you can tell whether or not the other copyright document is the same, without actually looking at the entire copyright document.