No.
> 2)Can you mark the information such that if it is distributed, you can
> track the one who distributed it.
No.
My opinion, and echoed by people around here: content providers are just
going to have to deal with these realities. It's *unnatural* to the
medium to try to do 1 and 2. For every step you can take towards trying
to enforce this, someone can write a program to break that down, probably
with less effort than you spent, too. That said, it is certainly
possible to work against sites committing massive copyright fraud, but
trying to prevent the average user from passing along some bits to
another user just isn't going to work.
> In each area one must consider audio, video, text, and images.
"bits is bits"... there's no functional difference, really, other than
some types take up a lot more space than others.
> Area 1:
>
> If the content provider encrypts the information with a user's public
> key and provides a viewer that decrypts the information using the
> user's private key, then ostensibly the information will be somewhat
> secure if there is no way to save a deciphered version of the
> information from the viewer.
>
> Of course, other programs can grab the information off whatever output device
> is being used. Image grabs are trivial to do, video grabs and audio
> grabs are somewhat harder. A text grab would take some effort and
> some OCR, but is possible. However, one can imagine stumbling blocks
> being put in place that would deter some of these, e.g. ignore certain
> X events etc.
But it would be relatively easy, especially if I had source to the
decrypting viewers, to write a "debabelizer" to convert the encrypted
data to ordinary data. No screen grabbing or X events needed. Of
course, this presumes that the encryption format is a standard or well
known - if it's not, we've got more things to worry about than copyright.
> Area 2:
>
> Well, how do you protect information if someone can grab it? Chiefly,
> I'm looking to see if anyone has investigated schemes to sign an
> image/video/audio with the user's key that are not perceptible and
> can't be removed from the information.
There has been talk here, I think, that one solution is to imbed into the
data the content provider sends some sort of information about who
received it - then, if a rogue archive of the CP's data appeared
somewhere, you could trace it to a particular recipient and act
accordingly. That's the way many commercial software packages work -
"this copy licensed to *blah*" appears on the startup screen and is
unchangeable (well, without hacking at the bit level I suppose). The
problem is in finding embedding methods, particularly for plain text.
> If anybody knows of people working on this problem, of papers that
> exist, etc. Please give me some pointers to the info. Also, if
> anyone knows of products that are being developed/exist in Area 1, I'd
> be interested in this information also. (are the commercial web
> browser people looking at this?)
I hope not.
Brian