However, an OS crash due to an application failure is generally considered to be the fault of the OS. Where the fault lies for an OS crash due to a driver bug depends on whether you're a microkernel zealot or not.
Third-party code that could reasonably take the blame for "crashing Windows" would almost certainly be limited to drivers or other in-kernel code. Since 99% of third-party kernel code tends to be drivers, most people just round up and say "drivers".:-)
So the ideal goal, then, is to not contribute to a defense fund, but rather, to make it so expensive for DTV to pursue this broken line of reasoning thatn they throw their hands up in disgust. Everyone on SlashDot should simply buy one of these unprogrammed readers. We should develop some sort of use for them as a security device for accessing our computers, and use precisely the same encoding firmware that DTV coding would require, just to be even more evil.
Then, suddenly they are forced with several million "suspects", the vast majority of whom are using the technology for something entirely legal. I figure at twenty minutes per C&D letter, that's still over a million dollars of legal expenses, plus the cost of maintaining enough staff to handle the "bite me" responses that come back from the overwhelming number of legitimate users.
It gives you powerful control over every aspect of the score. On the other hand, if it got the bloody placement of articulations and other markings right to begin with, you wouldn't need such minute control. There's nothing quite like having a dynamic marking, a slur marking, and a note articulation all simultaneously occupy the same space....
If only Igor Engraver would move to Mac OS X.... Sigh.
Yeah, and the 802.11 band (2.4 GHz) is supposed to be specifically allocated to unlicensed transmitters globally, or so the marketdroids keep saying. You mean it isn't? Say it ain't so! Say it ain't so!
Yeah, but you're driving a giant faraday cage.... It will leak like a sieve in your house, but in your car, it shouldn't be much of a nuisance except maybe to the person beside you on the M25.
Hmm. I don't know which is scarier... the thought of SBC winning or the thought of the RIAA winning. The lesser of two evils... the enemy of my enemy is my enemy? Aargh, the confusion!
Here are some helpful techniques that don't (quite) involve incineration. Kids, don't try this at home.
1. Degauss the drive with a tape demagnetizer, making sure to thoroughly hit each side of each platter. This does a pretty good job for the most part, at least against most typical end-user attempts to recover the data.
2. Demagnetize as above, then put it in your oven at the hottest temperature for an hour or several. While it won't demagnetize it, it will cause significant random changes in polarization (since your hard drive is only stable at a few tens of degrees above room temperature due to the high bit density). This should defeat pretty much any attempts to recover the data, but there might still be random parts of the platters that are recoverable.
3. Place it in your kiln. Heat to above the Curie point of Cobalt (about 1500C). It should be noted that this is, coincidentally, about the average melting point of the class substrate, though depending on composition, glass could potentially melt as low as 500C. If you are not able to raise the Cobalt alloy to its curie point, melting the glass substrate is a relatively effective alternative.
4. Cobalt metal will (slowly) dissolve in a dilute sulfuric acid mixture, and dissolves easily in hydrochloric acid or nitric acid. In any case, be careful not to get the acid on your skin (or anything else, for that matter). This is quite possibly the easiest effective solution for destroying data on the hard drive.
That's the chip vendor's policy. There's nothing preventing Abit from making a duplicate. I doubt they do, as that would seem like a really big headache, but if they wanted to, they could.
No, the only potentially completely secure solution is for the users to generate the key data themselves.
It probably is. The courts have always held that "reasonable" depends on the circumstances, which includes to some degree, the nature of the locality, whether you're talking about negligence or time/place/manner restrictions on speech. One thing's for sure, though, if somebody walks into an unlocked house, steals car keys, and drives it into a crowded bus stop, mowing down a busload of school children before shooting himself, people in your area probably won't leave their houses unlocked for very much longer.
As to whether you'd be held liable for their deaths (financially/civilly), it depends on not only whether the people in your area really leave their doors unlocked as a matter of course, but also on what their reaction is to the heinousness of the action committed by the criminal. If that happens, you'll definitely want to make a motion that the trial occur outside the area, because you're gonna have a hard time getting a fair trial locally. They might all know in their hearts that it was the norm, but they'll still be looking for someone to blame.
Umm... the original version only did 40-bit DES, and you're right, even I could probably crack that faster than realtime.
More recent versions of the X-Wall chipset, however, do both DES and triple-DES at 40*/64/128/192-bit strengths. It's unclear from the abit website what version of the chipset will be included in this motherboard.
(* indicates mode only available on the low-end version of the chipset)
IANAD either, but I find that caffine helps me in moderation. Moderation being what most/.'ers probably call insanely low levels;-)
<snip>
Bear in mind, I do have ADD (official diagnosis), so your milage may vary.
There's something to be said about that. In fact, it isn't really all that surprising.
Many drugs that are used to treat people with hyperactivity-like disorders result in a calming effect for those folks, but result in making normal people totally bounce off the walls. For some reason, some people's brains react to those sorts of chemicals in the opposite way from the way that a normal brain would react.
To make matters worse, at least in a traditional UFS-like environment (EXT*, FFS, etc.), "/." = "/", so slashdot starts off with a Score -1: Redundant....:-)
A device driver can't be isolated from the hardware, but device drivers are not necessarily part of the kernel, at least not in a modern operating system.
In a modern OS, it should be possible to write a driver for any arbitrary device largely (or in some cases entirely) in user space unless there is already a kernel driver. However, if the driver's primary client is the kernel, then it is a bad idea to do so for performance reasons.
For example, in Mac OS X, device drivers for HID (human interface) devices, such as mice, keyboards, trackpads, graphics tablets, and so on, reside in user space. Similarly, printer drivers reside in user space in almost all operating systems.
In the most extreme case, that of a traditional microkernel architecture such as CMU Mach, device drivers existed entirely in user space, as did file systems and a lot of other things that you would assume are always part of the kernel. I don't recall the details of how they were granted access to the hardware, but IIRC the microkernel mapped the device into the address space of the relevant Mach task.
So no, device drivers are not inherently part of the kernel. That having been said, you're right that there are some drivers that would be very hard to deal with outside the kernel---video card drivers, for example---and frankly, if your video driver crashes, unless we're talking about a server, you're pretty much going to have to reboot anyway, so from a stability standpoint, having such a driver in the kernel is no great loss.:-)
Can you imagine if there were thirty different ways to lay bricks?
You could do only one color of bricks with each row shifting by half.
You could do longer bricks and shift by a third the length.
You could put a brick of a different color in place of every third brick.
You could turn every other brick sideways so it would stick out the front.
You could tile the bricks alternating between two colors.
You might add a column of bricks, in which case you could have some of the bricks in a vertical orientation.
You might do a pattern of bricks on the ground consisting of two bricks sideway followed by two bricks turned 90 degrees.
You might use a short brick for every fourth brick, and you might lay every other row in opposite directions so that you would form a zig-zag pattern.
With all of these, you might use a light mortar or a dark mortar.
With all of these, you might choose to use a smooth brick or a rough brick.
That's 8 * 2 * 2 = 32 ways to lay bricks. Those are just the first few off the top of my head.
Just as there are infinitely many ways to write most non-trivial pieces of software---indeed, as with any art form---there are certain rules that must be met in order to get something that resembles usable output, but there are infinitely many ways to lay bricks. That having been said, no matter how complex the pattern, if you put the wrong brick in the wrong place, it's still a screw-up.
The only real difference is that screw-ups in software (i.e. bugs) are generally somewhat easier to fix. However, this in no way excuses the apathy that many programmers seem to feel about the existence of such mistakes.
Actually, it's very simple. If you were negligent---if you left your house unlocked and someone walked in, picked up the set of keys hanging by the door, and ran over 14 kids with your car, then yeah, you're liable. You aren't criminally guilty (though you might be held criminally negligent). You are, however, partially responsible for the crime in a civil sense, and thus could be held liable in a civil suit.
The question is whether the parents could reasonably have prevented the illegal music downloads. The answer in about 99% of cases is "no".
However, that doesn't have anything to do with whether they can be served with a subpoena. They aren't being charged with a crime. They're being asked to surrender property that may have been used to commit a crime. If the law and the subpoena are both legal (which is another debate entirely), then serving them with such a subpoena is reasonable, regardless of whether they're serving someone's parents, someone's roommate, or the guy who runs the self-storage shed down the block.
What troubles me is that the $40 is in addition to the $35 physical pressed CD distribution contract that you already have to be bound by before you can even start into this digital distribution agreement. But wait... before you can enter into that $35 contract, you have to either do duplication or replication of the CD. Or you could be really cheap and do homemake CD-Rs at a substantial investment in time, then keep the channel full at an even bigger investment in time if people buy it.
Not to mention that I believe I've read that the iTMS requires a barcode for every CD. Unless CDBaby is providing those, that pretty much means that anybody who wants their music distributed through iTMS basically has to submit a professionally duplicated disc (or pay someone who has a set of bar codes to sell them one). In either case, you're probably talking about a significant investment.
If the cost of entry were really $40, it would be a way of deterring people from submitting random crap. However, with it structured in the way it is, I rather suspect that the cost of entry is measured in hundreds of dollars, in which case, the $40 is a fee, regardless of how you want to describe it.
What I'd like to see would be CDBaby doing this without any requirement for physical distribution. This would thus provide a way for the really small labels (the 1 and 2 band labels that are probably too small to get on iTMS) to use it as a way of getting into iTMS, but without the need for them to handle your CD distribution (or for that matter, without the need for any physical CDs at all).
What's funny is that it's the directors and actors who are the only ones who make residuals, and thus all those other people are gonna get paid whether the movie sells or not. That is, of course, unless so many movies fail to sell that they end up losing their jobs entirely.
Almost all of the cost results from the airtime -only- if you are doing a for-tv commercial. Given that they're doing it on film for theater consumption, the numbers are probably much less skewed in that direction. Air time may not be cheap, but neither are 10-30 prints each for thousands upon thousands of theaters.:-)
I'd be very surprised if that were the case. An iBook 900 can easily move well over 2 megs/second between its hard drive and the optical drive. A 1x DVD-R burn only requires about 1.3 megs/second.
Now, I wouldn't even think about actually using a G3 machine to do the encoding for obvious speed reasons, but I can't imagine such a machine not being fast enough to burn one.:-)
I assume you meant "into a G4 motherboard". I'm not sure about being below G4 speeds, but it would definitely cripple it.
A more interesting upgrade card design would be similar to the Sonnet G3 upgrade for the PowerMac 7200. It's basically a single-board computer with its own RAM, etc. on the card, but using the MLB for peripherals via the PCI bus. An upgrade card design like that could easily bring reasonable single-processor G5 performance to a G4 or G3 machine, albeit probably at a much higher price than a G4 upgrade card (and possibly at a higher price than a G5 machine itself).
The real question becomes the line between possible and practical, and I'm not really sure where that line might be.
Intent to damage character is implicit in the action. I believe the word you're looking for here is "malice", not "intent". Intent, at least in cases of defamation (libel, slander), can be unintentional, as strange as that sounds.
For a public figure to claim libel or slander, you have to prove malice, i.e. that you knew the statement to be false and made it to deliberately attack the reputation of the person or company.
If you are a private individual, i.e. not a politician, celebrity, etc., you have to do nothing of the sort. You must simply prove that your reputation was injured, that the statement was false, and that at least one other person saw it. Of course, if you can prove malice, then your damages will be much greater.
Actually, it does -not- have to be public. Most people make this mistake. In fact, if only a single person other than the person prosecuting, that person's lawyer(s), and the defendant and his/her lawyer(s) sees it, then the prosecution has engaged in libel, according to the courts.
Put another way, if a secretary stuck the letter in the envelope, the company might as well have put a gun to their heads and pulled the trigger, assuming that these claims really are false in a majority of cases. Of course, if most of these people really were stealing service, then that's another issue. This still strikes me as legally questionable behavior for a company to do something like this, though....
Slashdot Mirror
Third-party code that could reasonably take the blame for "crashing Windows" would almost certainly be limited to drivers or other in-kernel code. Since 99% of third-party kernel code tends to be drivers, most people just round up and say "drivers". :-)
Then, suddenly they are forced with several million "suspects", the vast majority of whom are using the technology for something entirely legal. I figure at twenty minutes per C&D letter, that's still over a million dollars of legal expenses, plus the cost of maintaining enough staff to handle the "bite me" responses that come back from the overwhelming number of legitimate users.
Hey, a guy can dream, can't he?
If only Igor Engraver would move to Mac OS X.... Sigh.
I don't believe the OS X version is shipping for another month....
1. Degauss the drive with a tape demagnetizer, making sure to thoroughly hit each side of each platter. This does a pretty good job for the most part, at least against most typical end-user attempts to recover the data.
2. Demagnetize as above, then put it in your oven at the hottest temperature for an hour or several. While it won't demagnetize it, it will cause significant random changes in polarization (since your hard drive is only stable at a few tens of degrees above room temperature due to the high bit density). This should defeat pretty much any attempts to recover the data, but there might still be random parts of the platters that are recoverable.
3. Place it in your kiln. Heat to above the Curie point of Cobalt (about 1500C). It should be noted that this is, coincidentally, about the average melting point of the class substrate, though depending on composition, glass could potentially melt as low as 500C. If you are not able to raise the Cobalt alloy to its curie point, melting the glass substrate is a relatively effective alternative.
4. Cobalt metal will (slowly) dissolve in a dilute sulfuric acid mixture, and dissolves easily in hydrochloric acid or nitric acid. In any case, be careful not to get the acid on your skin (or anything else, for that matter). This is quite possibly the easiest effective solution for destroying data on the hard drive.
And remember, kids, don't try these at home.
No, the only potentially completely secure solution is for the users to generate the key data themselves.
As to whether you'd be held liable for their deaths (financially/civilly), it depends on not only whether the people in your area really leave their doors unlocked as a matter of course, but also on what their reaction is to the heinousness of the action committed by the criminal. If that happens, you'll definitely want to make a motion that the trial occur outside the area, because you're gonna have a hard time getting a fair trial locally. They might all know in their hearts that it was the norm, but they'll still be looking for someone to blame.
More recent versions of the X-Wall chipset, however, do both DES and triple-DES at 40*/64/128/192-bit strengths. It's unclear from the abit website what version of the chipset will be included in this motherboard.
(* indicates mode only available on the low-end version of the chipset)
<snip>
Bear in mind, I do have ADD (official diagnosis), so your milage may vary.
There's something to be said about that. In fact, it isn't really all that surprising.
Many drugs that are used to treat people with hyperactivity-like disorders result in a calming effect for those folks, but result in making normal people totally bounce off the walls. For some reason, some people's brains react to those sorts of chemicals in the opposite way from the way that a normal brain would react.
Doctorr: So if you press this button, the life support system will dispense Morphine.
Patient: Well, what does this button do?
Doctor: Oh, don't press that. That's the caffeine button.
(Patient pushes button).
Patient: Ah yes, that's the stuff.
Doctor: Jeez! Are you crazy!?! In your condition?
(Click, click, click, click, click, click, click, click, click, beep, beep, beep, beep, beeeeeeeeeeeeeeeeeeeeeeeeeeee.....)
In a modern OS, it should be possible to write a driver for any arbitrary device largely (or in some cases entirely) in user space unless there is already a kernel driver. However, if the driver's primary client is the kernel, then it is a bad idea to do so for performance reasons.
For example, in Mac OS X, device drivers for HID (human interface) devices, such as mice, keyboards, trackpads, graphics tablets, and so on, reside in user space. Similarly, printer drivers reside in user space in almost all operating systems.
In the most extreme case, that of a traditional microkernel architecture such as CMU Mach, device drivers existed entirely in user space, as did file systems and a lot of other things that you would assume are always part of the kernel. I don't recall the details of how they were granted access to the hardware, but IIRC the microkernel mapped the device into the address space of the relevant Mach task.
So no, device drivers are not inherently part of the kernel. That having been said, you're right that there are some drivers that would be very hard to deal with outside the kernel---video card drivers, for example---and frankly, if your video driver crashes, unless we're talking about a server, you're pretty much going to have to reboot anyway, so from a stability standpoint, having such a driver in the kernel is no great loss. :-)
- You could do only one color of bricks with each row shifting by half.
- You could do longer bricks and shift by a third the length.
- You could put a brick of a different color in place of every third brick.
- You could turn every other brick sideways so it would stick out the front.
- You could tile the bricks alternating between two colors.
- You might add a column of bricks, in which case you could have some of the bricks in a vertical orientation.
- You might do a pattern of bricks on the ground consisting of two bricks sideway followed by two bricks turned 90 degrees.
- You might use a short brick for every fourth brick, and you might lay every other row in opposite directions so that you would form a zig-zag pattern.
With all of these, you might use a light mortar or a dark mortar.With all of these, you might choose to use a smooth brick or a rough brick.
That's 8 * 2 * 2 = 32 ways to lay bricks. Those are just the first few off the top of my head.
Just as there are infinitely many ways to write most non-trivial pieces of software---indeed, as with any art form---there are certain rules that must be met in order to get something that resembles usable output, but there are infinitely many ways to lay bricks. That having been said, no matter how complex the pattern, if you put the wrong brick in the wrong place, it's still a screw-up.
The only real difference is that screw-ups in software (i.e. bugs) are generally somewhat easier to fix. However, this in no way excuses the apathy that many programmers seem to feel about the existence of such mistakes.
The question is whether the parents could reasonably have prevented the illegal music downloads. The answer in about 99% of cases is "no".
However, that doesn't have anything to do with whether they can be served with a subpoena. They aren't being charged with a crime. They're being asked to surrender property that may have been used to commit a crime. If the law and the subpoena are both legal (which is another debate entirely), then serving them with such a subpoena is reasonable, regardless of whether they're serving someone's parents, someone's roommate, or the guy who runs the self-storage shed down the block.
Not to mention that I believe I've read that the iTMS requires a barcode for every CD. Unless CDBaby is providing those, that pretty much means that anybody who wants their music distributed through iTMS basically has to submit a professionally duplicated disc (or pay someone who has a set of bar codes to sell them one). In either case, you're probably talking about a significant investment.
If the cost of entry were really $40, it would be a way of deterring people from submitting random crap. However, with it structured in the way it is, I rather suspect that the cost of entry is measured in hundreds of dollars, in which case, the $40 is a fee, regardless of how you want to describe it.
What I'd like to see would be CDBaby doing this without any requirement for physical distribution. This would thus provide a way for the really small labels (the 1 and 2 band labels that are probably too small to get on iTMS) to use it as a way of getting into iTMS, but without the need for them to handle your CD distribution (or for that matter, without the need for any physical CDs at all).
Now, I wouldn't even think about actually using a G3 machine to do the encoding for obvious speed reasons, but I can't imagine such a machine not being fast enough to burn one. :-)
A more interesting upgrade card design would be similar to the Sonnet G3 upgrade for the PowerMac 7200. It's basically a single-board computer with its own RAM, etc. on the card, but using the MLB for peripherals via the PCI bus. An upgrade card design like that could easily bring reasonable single-processor G5 performance to a G4 or G3 machine, albeit probably at a much higher price than a G4 upgrade card (and possibly at a higher price than a G5 machine itself).
The real question becomes the line between possible and practical, and I'm not really sure where that line might be.
For a public figure to claim libel or slander, you have to prove malice, i.e. that you knew the statement to be false and made it to deliberately attack the reputation of the person or company.
If you are a private individual, i.e. not a politician, celebrity, etc., you have to do nothing of the sort. You must simply prove that your reputation was injured, that the statement was false, and that at least one other person saw it. Of course, if you can prove malice, then your damages will be much greater.
Of course, IANAL....
Put another way, if a secretary stuck the letter in the envelope, the company might as well have put a gun to their heads and pulled the trigger, assuming that these claims really are false in a majority of cases. Of course, if most of these people really were stealing service, then that's another issue. This still strikes me as legally questionable behavior for a company to do something like this, though....