They can't compile the source code they have seen and install it on the computers of an offending government
They don't have to - it already is installed on a lot of interesting systems. The point here is, that they have seen the source, so they might have knowledge of bugs nobody else have noticed. The real question is, what would they do if their audit really revealed exploitable bugs? Refuse to use the software for anything important? Tell the world about their findings? Use their acquired knowledge for their own benefit?
If they blindly increase their warranty period, they will end up losing a lot of money on support and replacements.
Replacing a drive that fails after more than three years shouldn't cost much. Consider how much the price will decrease in that time. Of course it is possible to increase the cost of performing a replacement, by spending a lot of time on testing the deffective drive, and by sending the deffective drive back to the manufacturer. Just don't do that. Let the shop that sold the drive replace it without making a big deal of it, and Seagate will just have to pay whatever the price amounts to at that time, and they will still have made money on the entire deal.
I have to count picoseconds for the kind of stuff I do
Unless you are working with individual gates inside a chip, I doubt picoseconds really matters. On ethernet we are certainly not talking picoseconds. We are still limited by the speed of light, so it would take the signal 100 picoseconds just to get through the RJ45 connector. With a 1.5m ethernet cable there will be at least 10 nanoseconds of roundtrip time, because that is the time it takes light to travel 3m.
I would rather pay a nominal fee for a easy to use, secure linux distro that get windows XP for free.
I agree with that. I didn't start using Linux because of the price, but because it was the OS that could satisify my needs. Windows just would have been incompatible with all the systems I had to work with. Actually when I recently bought a laptop, Windows XP was preinstalled. Of course the first thing I did was to wipe the harddisk and install Linux.
it says that in exchange for adding a warranty you can charge for the product
Yes, but that applies only to the binary code. Once you have bought the binary code, they must provide you the source for a charge no more than your cost of physically performing source distribution. If they sell binaries for $10 but charge $50 for the source, that is clearly a violation. If it was the other way around it might have been better. But I don't believe their costs every time somebody download the source is even $1. It would be different if the source was provided on a CD. But where is the written offer required by the GPL? Actually there might also be another problem. Since source is not immediately available to everybody who buy the binary code they don't comply with 3a, so they must comply with 3b, which says the source must be available to any third party. So you are entitled to get the source even if you didn't buy the binary. There are some pretty hard requirements in 3b, so I don't understand why anybody would chose to use it, since complying with 3a hardly doesn't require anything.
No, but RedHat's automatic update thing at least is almost certainly in violation...
Now I'm really happy that I long ago wrote a script to do it from my command line. It is also easier to do remote that way. And I just download updates once to one of my computers and rsync the directory to the rest. (Would be so great if this could force people back to the command line, I could be the ubergeek again, like in the good old days before crappy GUIs (-:.)
If they are the same, other circuitry in the CPU takes the same input value the comparitor saw -- from the exact same electrical 'pins' -- and passes it along.
And this is where something can go wrong. If the right data goes to the comparision unit, but on the other path to where the data is actually to be used, an error happens, it is not detected.
Having an error that matters, then also having an
error in the comparision unit, would amount to
having two unlikely errors at the exact same moment.
That is not what I had in mind. I think about a situation where both pipelines work correctly, but an error happens in the comparision unit, such that you end up sending the incorrect data to RAM.
Re-generating the CPU
registers should be possible as long as memory on
which they depend has gone unmodified (or can be
rolled back) since the registers were last saved.
Memory doesn't go unmodified for long time, how many instructions does programs execute on average between memory writes? (Less than 10 I would think). Rolling back memory contents is not an option if you have threads communicating through shared memory. And normally registers are saved and restored only by the scheduler. But what if an error happened while the scheduler was running.
I wonder if the consensus impossibility result with a single crash can be adopted to show something about a system of this kind.
Fire up apache and then post a link to it here on slashdot.
I don't think servers get slashdotted due to lack of CPU/RAM resources, but rather due to insufficient networkbandwidth or bad programming/configuration. A good programmer can make a C64 survive a slashdotting, it might be slow, but it keeps responding. With a 100mbit/s internet connection and a page with not too many pictures, I think any decent server should survive a slashdotting.
If the results differ, the CPU is taken down without corrupting memory as it dies.
A few questions:
What if an error happens in the comparison unit?
What happens to the program that was running on the CPU as it is taken down? (The CPU registers is part of the program state, so you cannot just continue on another CPU).
Until the vanilla Linux kernel accepts these changes and scale, Solaris still has a big edge in this area.
I wouldn't be surprised to see these changes in the 2.8 kernel. And what will people do until then I hear some people ask. I can tell you that right now it is very few people that actually have the need to scale to 1024 CPUs. And that will probably also be true by the time Linux 2.8.0 is released. AFAIK Linux 2.6 does scale well to 128 CPUs, but I don't have hardware to test it, neither does any of my friends. So I'd say there is no need for a rush to get this in mainstream, the few people that need this can patch their kernels. My guess is that in the time from now until 2.8.0 is released, we will see less than 1000 such machines worldwide.
Hmm, scan word docs looking for legalese adding and removing the word "not" at appropriate points.
An interesting variant would be to not do it to documents on disk, but rather do it only when printing. Imagine you print three or four copies of a contract, that had the word not removed in different places, and then they gets signed.
Rights simply cannot by obtained "step-by-step", as there can be reason under the sun that says killing someone at step #6,123 is murder, the most heinous crime you can commit against another person, but step #6,122 is not
murder.
Step-by-step makes the most sense. No matter where you try to put the dividing line, it will be a problem. The actual moment of birth definitely is too late. The conception OTOH is too early. If a fertilized egg have any rights, then the use of a coil would be murder, wouldn't it? I see no reason why you couldn't graduate the crime, don't call it murder, but come up with another word. And the later you do it, the higher the punishment.
It was never installed on any of my computers. But some surveys says it is about 95% of the users. Of course those surveys are not 100% reliable. Anybody still using IE should seriously consider switching now.
Joe Sixpack needs free hands-on technical support from a Linux professional to manage a Linux desktop
That is no different from the average Windows users. I have people asking me about Windows problems all the time. The main difference is, that if they come with a Linux problem I can help them. Windows just doesn't give me the tools to debug the problems.
and you wonder why OEM Windows still sells at the rate of ten million a month?
Because they force anybody who buy a computer to buy a copy of Windows as well.
I thought Linux was supposed to be all about choice.
True, but in this case it is not Linux that impose restrictions on you. It is NVidia that impose restrictions on you. When you buy undocumented hardware you get those kinds of trouble. In other words if you want freedom there are certain pieces of hardware you have to avoid.
I would need to upgrade the drivers sometime anyway.
If NVidia would allow the drivers to go mainstream, that wouldn't be a problem. You would simply get the new drivers with your distribution's update system.
I also had trouble with other things, like getting Mplayer to work so I could watch DVDs.
I use Fedora Core which does not by default include mplayer (cowards), but I can download it from freshrpms.net, and that works just great. But for watching DVDs I prefer ogle (also from freshrpms.net).
I couldn't install the newest Linux NVidia drivers (needed recompiling of the kernal or something).
If NVidia had been a litle more cooperative, you wouldn't even had needed to do that on your own, it would have come with the install. So I would have liked to advice you to chose better hardware, but unfortunately I don't know which gfx hardware to suggest. Personally I don't have much need for 3Dgfx, so I didn't investigate it.
I tried installing a few FPS demos, but couldn't.
I never tried it, so I cannot say for sure what problems there could be. But could it be, that they simply wouldn't work without the right drivers?
Slashdot Mirror
They can't compile the source code they have seen and install it on the computers of an offending government
They don't have to - it already is installed on a lot of interesting systems. The point here is, that they have seen the source, so they might have knowledge of bugs nobody else have noticed. The real question is, what would they do if their audit really revealed exploitable bugs? Refuse to use the software for anything important? Tell the world about their findings? Use their acquired knowledge for their own benefit?
which was forced to change its name to ZipZoomFly.
Ouch, what a horrible name. If that is the kind of punishment you get, you'd better avoid infringing on anybody's trademark to start with.
If they blindly increase their warranty period, they will end up losing a lot of money on support and replacements.
Replacing a drive that fails after more than three years shouldn't cost much. Consider how much the price will decrease in that time. Of course it is possible to increase the cost of performing a replacement, by spending a lot of time on testing the deffective drive, and by sending the deffective drive back to the manufacturer. Just don't do that. Let the shop that sold the drive replace it without making a big deal of it, and Seagate will just have to pay whatever the price amounts to at that time, and they will still have made money on the entire deal.
Maybe that's because I live in France ? /etc/resolve.conf file.
More likely it has something to do with the contents of your
http://fr/ too.
I tried it, but it doesn't resolve.
I have to count picoseconds for the kind of stuff I do
Unless you are working with individual gates inside a chip, I doubt picoseconds really matters. On ethernet we are certainly not talking picoseconds. We are still limited by the speed of light, so it would take the signal 100 picoseconds just to get through the RJ45 connector. With a 1.5m ethernet cable there will be at least 10 nanoseconds of roundtrip time, because that is the time it takes light to travel 3m.
Screaming video cards.
I think you should replace the fan.
ai and dk should work.
I just create a list with all the short ones I could find that actually resolve.
I would rather pay a nominal fee for a easy to use, secure linux distro that get windows XP for free.
I agree with that. I didn't start using Linux because of the price, but because it was the OS that could satisify my needs. Windows just would have been incompatible with all the systems I had to work with. Actually when I recently bought a laptop, Windows XP was preinstalled. Of course the first thing I did was to wipe the harddisk and install Linux.
it says that in exchange for adding a warranty you can charge for the product
Yes, but that applies only to the binary code. Once you have bought the binary code, they must provide you the source for a charge no more than your cost of physically performing source distribution. If they sell binaries for $10 but charge $50 for the source, that is clearly a violation. If it was the other way around it might have been better. But I don't believe their costs every time somebody download the source is even $1. It would be different if the source was provided on a CD. But where is the written offer required by the GPL? Actually there might also be another problem. Since source is not immediately available to everybody who buy the binary code they don't comply with 3a, so they must comply with 3b, which says the source must be available to any third party. So you are entitled to get the source even if you didn't buy the binary. There are some pretty hard requirements in 3b, so I don't understand why anybody would chose to use it, since complying with 3a hardly doesn't require anything.
No, but RedHat's automatic update thing at least is almost certainly in violation...
Now I'm really happy that I long ago wrote a script to do it from my command line. It is also easier to do remote that way. And I just download updates once to one of my computers and rsync the directory to the rest. (Would be so great if this could force people back to the command line, I could be the ubergeek again, like in the good old days before crappy GUIs (-:.)
If they are the same, other circuitry in the CPU takes the same input value the comparitor saw -- from the exact same electrical 'pins' -- and passes it along.
And this is where something can go wrong. If the right data goes to the comparision unit, but on the other path to where the data is actually to be used, an error happens, it is not detected.
Having an error that matters, then also having an error in the comparision unit, would amount to having two unlikely errors at the exact same moment.
That is not what I had in mind. I think about a situation where both pipelines work correctly, but an error happens in the comparision unit, such that you end up sending the incorrect data to RAM.
Re-generating the CPU registers should be possible as long as memory on which they depend has gone unmodified (or can be rolled back) since the registers were last saved.
Memory doesn't go unmodified for long time, how many instructions does programs execute on average between memory writes? (Less than 10 I would think). Rolling back memory contents is not an option if you have threads communicating through shared memory. And normally registers are saved and restored only by the scheduler. But what if an error happened while the scheduler was running.
I wonder if the consensus impossibility result with a single crash can be adopted to show something about a system of this kind.
Fire up apache and then post a link to it here on slashdot.
I don't think servers get slashdotted due to lack of CPU/RAM resources, but rather due to insufficient networkbandwidth or bad programming/configuration. A good programmer can make a C64 survive a slashdotting, it might be slow, but it keeps responding. With a 100mbit/s internet connection and a page with not too many pictures, I think any decent server should survive a slashdotting.
A few questions:
Until the vanilla Linux kernel accepts these changes and scale, Solaris still has a big edge in this area.
I wouldn't be surprised to see these changes in the 2.8 kernel. And what will people do until then I hear some people ask. I can tell you that right now it is very few people that actually have the need to scale to 1024 CPUs. And that will probably also be true by the time Linux 2.8.0 is released. AFAIK Linux 2.6 does scale well to 128 CPUs, but I don't have hardware to test it, neither does any of my friends. So I'd say there is no need for a rush to get this in mainstream, the few people that need this can patch their kernels. My guess is that in the time from now until 2.8.0 is released, we will see less than 1000 such machines worldwide.
Hmm, scan word docs looking for legalese adding and removing the word "not" at appropriate points.
An interesting variant would be to not do it to documents on disk, but rather do it only when printing. Imagine you print three or four copies of a contract, that had the word not removed in different places, and then they gets signed.
Also, the idea of selling a hard drive with linux preinstalled is really cool.
It is not a new idea. I read about Linux On A Disk back in 1999. But the site looks dead. Does LOAD still exist?
The song One Night in Bangkok from the musical Chess.
.ogg file would be hilarious.
Now if somebody with a bit of talent (unlike me) would record this version, I believe listening to the
Rights simply cannot by obtained "step-by-step", as there can be reason under the sun that says killing someone at step #6,123 is murder, the most heinous crime you can commit against another person, but step #6,122 is not murder.
Step-by-step makes the most sense. No matter where you try to put the dividing line, it will be a problem. The actual moment of birth definitely is too late. The conception OTOH is too early. If a fertilized egg have any rights, then the use of a coil would be murder, wouldn't it? I see no reason why you couldn't graduate the crime, don't call it murder, but come up with another word. And the later you do it, the higher the punishment.
does anyone still uses IE?
It was never installed on any of my computers. But some surveys says it is about 95% of the users. Of course those surveys are not 100% reliable. Anybody still using IE should seriously consider switching now.
Joe Sixpack needs free hands-on technical support from a Linux professional to manage a Linux desktop
That is no different from the average Windows users. I have people asking me about Windows problems all the time. The main difference is, that if they come with a Linux problem I can help them. Windows just doesn't give me the tools to debug the problems.
and you wonder why OEM Windows still sells at the rate of ten million a month?
Because they force anybody who buy a computer to buy a copy of Windows as well.
I thought Linux was supposed to be all about choice.
True, but in this case it is not Linux that impose restrictions on you. It is NVidia that impose restrictions on you. When you buy undocumented hardware you get those kinds of trouble. In other words if you want freedom there are certain pieces of hardware you have to avoid.
I would need to upgrade the drivers sometime anyway.
If NVidia would allow the drivers to go mainstream, that wouldn't be a problem. You would simply get the new drivers with your distribution's update system.
I also had trouble with other things, like getting Mplayer to work so I could watch DVDs.
I use Fedora Core which does not by default include mplayer (cowards), but I can download it from freshrpms.net, and that works just great. But for watching DVDs I prefer ogle (also from freshrpms.net).
I couldn't install the newest Linux NVidia drivers (needed recompiling of the kernal or something).
If NVidia had been a litle more cooperative, you wouldn't even had needed to do that on your own, it would have come with the install. So I would have liked to advice you to chose better hardware, but unfortunately I don't know which gfx hardware to suggest. Personally I don't have much need for 3Dgfx, so I didn't investigate it.
I tried installing a few FPS demos, but couldn't.
I never tried it, so I cannot say for sure what problems there could be. But could it be, that they simply wouldn't work without the right drivers?