Others argue that by pretending that we can't censor our nodes, we avoid responsibility for failing to do so. Well, I regard copyright infringement (by an individual in the privacy of his home) as a civil matter and will not attempt to censor infringing content and will go to trial if necessary.
But permitting KP to be distributed without making a reasonable effort
to prevent it may be a criminal matter, and it is beside the point of
freenet, IMHO, and not worth going to trial over. So I point out how to
censor the easily accessible KP from your node. I also argue that the
law enforcement people who run freenet nodes ought to publish lists of
KP CHKs that deserve censorship.
On Sat, 2004-08-07 at 08:51, Toad wrote:
> On Fri, Aug 06, 2004 at 11:25:46PM -0400, Edward J. Huff wrote:
> > Kiddie porn is a minor issue compared to the copyright problem.
> > For that, I think there will be a technological solution which
> > will make it less easy to find on Freenet, so that no one without an ax
> > to grind will think node operators are liable. Regarding the copyright
> > problem, come and get us. We aren't going to give up without a fight.
>
> Uh, like what?
I think everyone agrees it's not hard to convert a base 64 CHK key to
the hex representation used in filenames, and to delete matching files
from the datastore. (I don't see why SSK's would be any more difficult
but I haven't actually looked at the details there).
It is not hard to find out the "well known" kiddie porn CHK's. Monitor
the KP Frost boards; also pick out KP from the descriptions in TFE.
In order to avoid decrypting actual KP, use freenet.client.cli.Main
with --noredirect and use /dev/null as the output filename. The
metadata comes out on stdout, and includes the CHK's of any splitfile
blocks.
So, without any help from developers, users could set up scripts which
delete all "well known" KP CHK's from their datastores. Also, given
patches and scripts, they could compile a version of fred (or distribute
a jar which goes ahead of freenet.jar on classpath) which gives
immediate RNF to such keys.
If they at least delete KP CHK's, I think no jury would find such a node
operator criminally liable for carrying KP which he could not find out
about without devoting large amounts of time looking for the keys.
Hopefully, no prosecutor would go after such node operators, since being
dragged into court is a considerable inconvenience even if the verdict
is innocent. (The police rarely offer apologies, let alone
compensation).
On the other hand, investigators running freenet nodes themselves can
easily find nodes which fail to censor well known KP CHK's. There is no
need to "break" freenet to be able to do this.
If the investigators are decent people, they will give such node
operators a chance to voluntarily censor the KP, and even provide a list
of CHK's which should be censored. The investigators themselves are
able to look at the actual content and decide if it is KP without risk
of illegal possession of decrypted KP.
---- file keyBlacklist/KeyBlacklist.java ----
package keyBlacklist;
import freenet.client.Base64;
import freenet.support.HexUtil;
public class KeyBlacklist
{
public static void main(String[] args)
throws freenet.client.IllegalBase64Exception
{
int i;
for (i = 0; i < args.length; i++) {
if (args[i].startsWith("CHK@")) {
byte[] key = Base64.decode(args[i].substring(4));
System.out.println(HexUtil.bytesToHex(key));
} else {
byte[] key = HexUtil.hexToBytes(args[i]);
System.out.println(Base64.encode(key));
}
}
}
}
---- compile and run script ----
/usr/java/j2sdk1.4.2_05/bin/javac -classpath \
freenet.jar keyBlacklist/KeyBlacklist.java
/usr/java/j2sdk1.4.2_05/bin/jar cvf keyBlacklist.jar \
keyBlacklist/KeyBlacklist.class
/usr/java/j2sdk1.4.2_05/bin/java -classpath \
freenet.jar:keyBlacklist.jar \
keyBlacklist.KeyBlacklist [EMAIL PROTECTED]
signature.asc
Description: This is a digitally signed message part
_______________________________________________ chat mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.general
