Amend federal law to remove common carrier in civil cases.
Of all the legal options, this one will be the most effective, as it gets around the problem that the US laws would have no jurisdiction outside the US (and that, even if enough nations ban the sending of spam to anywhere from within their borders, that will only make it more worth it for nations like Luxembourg or Costa Rica or whatnot to keep it legal). How would this work?
US ISP receives spam
US ISP sues their upstream for carrying it. In all probability, the upstream won't bother fighting it, so they'll pay up and accept the log of the email headers.
The upstream ISP will then be able to sue the next cog in the chain up (pass-through liability), much as insurance companies are able to do.
Lather, rinse repeat.
At some point, you hit one of:
the spammer himself, actually sued by his ISP.
some ISP decides that the buck stops here and eats the cost; however you can bet that they won't be willing to give the client who put them in this position a great deal on his service.
an international backbone provider from $COUNTRY to the USA that pays up and has no recourse in $COUNTRY.
How does this work with say, a spam sent from Korea to AOL through C&W?
AOL sues C&W. AOL wins. C&W pays up. Assume that C&W can't go after one of their customers in Korea. This fact would lead them to offload the risk of doing business in Korea to insurance companies. The insurance premiums then get built into their connectivity prices from Korea to the US. Eventually every operator of a route from Korea to the US will either give up or stay in with dramatically higher prices. A thousandfold increase in bandwidth prices from Korea to the US is not out of the question. At this point, after some relay operator in Korea gets a bill from his ISP that's 50 times higher than normal, one suspects that he would close the relay. Problem solved.
Extend this to DDoS zombies, copyright infringement, and other civil offenses.
I imagine DDoS attacks will become a lot more difficult when people whose machines became zombies get a letter from Comcast or Verizon terminating their service with a demand of $1,000 to turn it back on.
Once you've established the liability for these activities in monetary terms, insurance companies can move in. You would have the option of buying insurance for your internet connection protecting you from liability of this type. If you're running up to date software, not participating in file-sharing networks, and not running an anonymous proxy or open relay, you're premiums will be low and since you're properly insured, the ISPs will give you service at the normal current rates.
Just wondering how useful "only show images from originating server" really is. After all, there's a lot of non-banner uses for a separate image server (the main one being if you're running a dynamic site and want to mod_mmap the graphics for even better performance).
Then you turn the current model on its head. The producer tracks down the advertisers who pay the producer. The producer then pays NBC to put the show on.
And you'll see more of a movement to sports programming where two things come into play:
The proportion of viewership watching on a time-skipping basis is likely less (especially when communal viewing (e.g. bars) is taken into account, though current audience measurements do a piss-poor job of that)
It's trivial to integrate the advertising into the content (beyond event-produced ads like boards on sidelines and sponsorship patches on clothing); CBS, for instance, was periodically digitally painting AOL 9.0 ads on the field during the Florida/Florida State game Saturday.
In this case the theory that it's to get chicks can probably be ruled out, as:
It's a chick who solved it
Even if she is a lesbian, I doubt there's a single hetero- or homosexual female on this earth that will sleep with someone because they solved a math problem.
I was working in a convenience store a few years ago (third shift in the middle of nowhere). The power goes out, causing the ATM to reboot. Lo and behold, an OS/2 boot screen...
OS/2 is still quite possibly the most bulletproof PC operating system yet devised.
Unless you actually modified the source code the developer uploads, you're not going to do anything.
For Mandrake, in order to upload a package, the developer builds a source RPM on their box, scp's it to Mandrake's build cluster, then ssh's in and builds the binary there (I myself install the SRPM, then build a new SRPM and the binary RPM on the cluster). The binary thus built is then uploaded to the main repository. This is the way that Mandrake policy specifies for building packages for inclusion (and I believe the upload script checks for this). This build cluster is, one would assume, well-monitored and so forth.
An attacker could, for instance, hijack an existing patch being maintained on the developer's box. That's about the only thing that such a h4x0ring would do.
Use Mandrake. From a admin perspective, your RH knowledge (and scripts) will transfer over, and urpmi trumps apt as far as security is concerned (all official Mandrake packages are signed). urpmi also, from what I've heard from debian users who have jumped to Mandrake, has less of a tendency to trash a system than apt.
Go with Mandrake. You get all the power of debian thanks to urpmi (and there's work being done on being able to roll-back upgrades through urpmi as well).
Adjust the score of bayes_99. Every few months or so, I increase the scores of the bayesian tests by 10% or so, as the training from an expanded corpus makes the bayesian scores more reliable.
I've been thinking about implementing my own spamassassin derivative that, rather than assign scores to distinct regexps and then run through a bayesian scanner, uses the regexps matched as extra tokens for the bayesian scanner to chew on. Because the regexps would be crafted to look at certain non-tokenized data (such as a gap of more than 6-12 hours in the Received: headers, or similar domains in the To: or Cc: addresses, or indications of a dictionary attack, etc.) this would undoubtedly be more effective than a simple bayesian scanner. But I'd actually have to learn perl before doing that...
The RIAA is not pursuing criminal cases. They're going through the civil courts. As a result they can't jail you (unless you're found in contempt, but that's a separate offense); they can only deprive you of property. This in turn means that the standard of proof is not "beyond reasonable doubt", but "preponderance of probability". If they can show that there's at least a 50% chance that you're liable (note that it is not whether you did it or not) for the actions being contested, then you lose.
Think back to the OJ case. The criminal charges resulted in "not guilty". The civil charges resulted in "liable".
FWIW, listen to the St. Anger DVD... normal snare tunings there (and generally better production as Bob Rock was not involved with the DVD). If there was a decent way to only rip the audio from one chapter of a DVD, I'd use that as my canonical version of St. Anger.
Indeed, Mr. Robertson likely believes (as do many born-agains) that the Catholic Church isn't even Christian and is in league with the Freemasons and the Trilateral Commission to impose the New World Order...
Slashdot Mirror
How about this:
Amend federal law to remove common carrier in civil cases.
Of all the legal options, this one will be the most effective, as it gets around the problem that the US laws would have no jurisdiction outside the US (and that, even if enough nations ban the sending of spam to anywhere from within their borders, that will only make it more worth it for nations like Luxembourg or Costa Rica or whatnot to keep it legal). How would this work?
US ISP receives spam
US ISP sues their upstream for carrying it. In all probability, the upstream won't bother fighting it, so they'll pay up and accept the log of the email headers.
The upstream ISP will then be able to sue the next cog in the chain up (pass-through liability), much as insurance companies are able to do.
Lather, rinse repeat.
At some point, you hit one of:
How does this work with say, a spam sent from Korea to AOL through C&W?
AOL sues C&W. AOL wins. C&W pays up. Assume that C&W can't go after one of their customers in Korea. This fact would lead them to offload the risk of doing business in Korea to insurance companies. The insurance premiums then get built into their connectivity prices from Korea to the US. Eventually every operator of a route from Korea to the US will either give up or stay in with dramatically higher prices. A thousandfold increase in bandwidth prices from Korea to the US is not out of the question. At this point, after some relay operator in Korea gets a bill from his ISP that's 50 times higher than normal, one suspects that he would close the relay. Problem solved.
Extend this to DDoS zombies, copyright infringement, and other civil offenses.
I imagine DDoS attacks will become a lot more difficult when people whose machines became zombies get a letter from Comcast or Verizon terminating their service with a demand of $1,000 to turn it back on.
Once you've established the liability for these activities in monetary terms, insurance companies can move in. You would have the option of buying insurance for your internet connection protecting you from liability of this type. If you're running up to date software, not participating in file-sharing networks, and not running an anonymous proxy or open relay, you're premiums will be low and since you're properly insured, the ISPs will give you service at the normal current rates.
Any wm that uses GTK+2 should support AA fonts. Sawfish is an example.
Still, how many non-banners are you missing?
Indeed, I'd probably say that Mandrake is the best RPM-based server distro out there (especially for the price!).
Just wondering how useful "only show images from originating server" really is. After all, there's a lot of non-banner uses for a separate image server (the main one being if you're running a dynamic site and want to mod_mmap the graphics for even better performance).
Then you turn the current model on its head. The producer tracks down the advertisers who pay the producer. The producer then pays NBC to put the show on.
And you'll see more of a movement to sports programming where two things come into play:
In this case the theory that it's to get chicks can probably be ruled out, as:
Also, Mandrake's configuration tools are all GTK+, not Qt.
Sir, is it just me, or are you completely illiterate?
I was working in a convenience store a few years ago (third shift in the middle of nowhere). The power goes out, causing the ATM to reboot. Lo and behold, an OS/2 boot screen...
OS/2 is still quite possibly the most bulletproof PC operating system yet devised.
The real problem is that the "ogg" plugin for xmms only recognizes vorbis...
The Japanese have long had a history of whitewashing WWII in their history education. The books basically go from Midway to Hiroshima.
Christ, only Austria is less willing to admit any sort of wrongdoing in that war.
He was just stating the simple truth that it's the winners who write the history books.
Opera for one.
I think some Moz derivatives support this as well.
Unless you actually modified the source code the developer uploads, you're not going to do anything.
For Mandrake, in order to upload a package, the developer builds a source RPM on their box, scp's it to Mandrake's build cluster, then ssh's in and builds the binary there (I myself install the SRPM, then build a new SRPM and the binary RPM on the cluster). The binary thus built is then uploaded to the main repository. This is the way that Mandrake policy specifies for building packages for inclusion (and I believe the upload script checks for this). This build cluster is, one would assume, well-monitored and so forth.
An attacker could, for instance, hijack an existing patch being maintained on the developer's box. That's about the only thing that such a h4x0ring would do.
Use Mandrake. From a admin perspective, your RH knowledge (and scripts) will transfer over, and urpmi trumps apt as far as security is concerned (all official Mandrake packages are signed). urpmi also, from what I've heard from debian users who have jumped to Mandrake, has less of a tendency to trash a system than apt.
If Mandrake got h4x0r3d, urpmi would still work, as all RPMs supported by Mandrake are signed.
Nice to know, that.
Go with Mandrake. You get all the power of debian thanks to urpmi (and there's work being done on being able to roll-back upgrades through urpmi as well).
Adjust the score of bayes_99. Every few months or so, I increase the scores of the bayesian tests by 10% or so, as the training from an expanded corpus makes the bayesian scores more reliable.
I've been thinking about implementing my own spamassassin derivative that, rather than assign scores to distinct regexps and then run through a bayesian scanner, uses the regexps matched as extra tokens for the bayesian scanner to chew on. Because the regexps would be crafted to look at certain non-tokenized data (such as a gap of more than 6-12 hours in the Received: headers, or similar domains in the To: or Cc: addresses, or indications of a dictionary attack, etc.) this would undoubtedly be more effective than a simple bayesian scanner. But I'd actually have to learn perl before doing that...
The RIAA is not pursuing criminal cases. They're going through the civil courts. As a result they can't jail you (unless you're found in contempt, but that's a separate offense); they can only deprive you of property. This in turn means that the standard of proof is not "beyond reasonable doubt", but "preponderance of probability". If they can show that there's at least a 50% chance that you're liable (note that it is not whether you did it or not) for the actions being contested, then you lose.
Think back to the OJ case. The criminal charges resulted in "not guilty". The civil charges resulted in "liable".
What likely happens is that the homeowner is sued and loses and then turns around and sues the fence manufacturer.
I'm not sure that hax0ring should be a criminal offense, but it should be at least a civil offense (in which case the issue gets muddier).
Perhaps only pay for the songs that are actually redeemed...
FWIW, listen to the St. Anger DVD... normal snare tunings there (and generally better production as Bob Rock was not involved with the DVD). If there was a decent way to only rip the audio from one chapter of a DVD, I'd use that as my canonical version of St. Anger.
Indeed, Mr. Robertson likely believes (as do many born-agains) that the Catholic Church isn't even Christian and is in league with the Freemasons and the Trilateral Commission to impose the New World Order...