A recent data recovery operation I did on a Windows XP system:
My friend had installed VMWare, and I think he somehow managed to set VMWare to use his physical primary drive as the drive the VM would use. He then installed Windows 95a (FAT16) onto his 40GB drive (which he thought was a virtual drive) which surprisingly worked until he tried to exit the VM and it wouldn't let him. After he rebooted, his machine booted into Windows 95a -- it had overwritten his NTFS drive! He was horrified at all the data he had lost, so I told him I'd try to recover it, although I didn't have much confidence in the situation. Rebuilding the MBR was straightforward -- I just used an MBR from another Win2K machine and modified the partition table to match his drive size (this is something I have a lot of experience with). The boot sector was also straightforward -- NTFS keeps a backup copy of the boot sector at the end of the partition. My main concern was the MFT. The mirror at the center of the drive would be intact (the FAT16 partition used by Win95a was 2GB and wouldn't reach 20GB into the drive), but I didn't think that the MFT mirror was a complete mirror. However, after reading the location of the MFT in the boot sector, it turned out that his MFT was located 3GB into the drive! This surprised me quite a bit. I later checked into this and found that a friend's XP machine had the MFT at about the same spot. My Win2k machine at work had the MFT about 200-400mb into the drive, and my machine at home had the MFT less than 1MB into the drive. Anyway, after a reboot (into an alternate installation of WinXP) and a scan, he was able to recover most of his data, which was a great relief. Some day though, I'll have to try out the MFT mirror thing and see if it would have been enough to recover an NTFS partition...
I remember reading that stores were left over with thousands of those Jar-Jar dolls because no one wanted them. They didn't sell very well, and if you've seen the movie, it would've been pretty obvious.
The parent poster makes an excellent point. I sometimes wonder about SCO's employees, the ones that actually wrote their software. I wonder how they feel about SCO attacking Linux. It would be pretty weird if one day you came to work and found out that your new CEO has decided to attack other companies using baseless accusations because your funding comes from a parasitical IP company. It's the sorta thing that would make you want to quit -- if you weren't already being laid off because the company didn't need to actually produce anything.
I think Baystar wants to keep SCO low profile and is telling McBride to shut his mouth because if more people were aware of their parasitical business model, I think it would hurt them financially and they might suffer some backlash from the public at large. As it stands now, probably only the IT community, and maybe just a subset of that, is aware of the implications of this lawsuit.
True. But the distinction I think he's trying to describe is like comparing Newtonian physics and Einsteinian physics. For most everyday stuff involving objects we can see and speeds easily measured, Newtonian physics work well (e.g. using F=ma to measure acceleration of a car). But as you approach the speed of light or supermassive objects, Newtonian physics' inaccuracies appear. The more extreme the conditions are, simpler models show their inadequacies and a more detailed accurate model is constructed. The same thing applies with electrons -- the basic model of an electron works reasonably well for things such as building simple electronic circuits and maybe particle collisions (I can't really say for sure, IANAP), but as more extreme conditions are explored, a more detailed model may be needed to explain electron behavior. Maybe it's like a fractal -- the closer you look at the edges, the more details that appear.
I'd also like to note that switching to jdk1.4.2 (I think we were using 1.3.1 previously) under Linux gave our servers an enormous boost in speed -- a lot of CPU/disk/network intensive code ran 5 times as fast. Btw, the JDK we used was from Blackdown.
I think they're blocked now, but they used to not be. I used to go to GRC.com and run that Shields-Up test to see what ports were visible from the outside, and the usual suspects were visible. However, you can simply unbind TCP with File & Printer Sharing (and even Client for MS Networks), and though the ports are visible, they're closed or good as closed.
I think a real big advantage for ATI is the fact that their card doesn't take up two slots, require a monstrosity of a heat sink and fan, and recommend/require a 450W power supply like the '6800 does. Even if the new ATI card wasn't as fast as the 6800, I wouldn't consider buying a video card like that. And I've always considered myself a fan of Nvidia cards (I used to hate the "ATI OS" that ATI's old drivers used to install -- it was very invasive). ATI has produced a very competitive card performance-wise, while keeping the same form factor and with a reasonable (relatively speaking) level of power consumption and heat dissipation.
This is a slippery slope... stretch the chain of events out, and maybe that cop that pulled you over for speeding was delayed just long enough that he wasn't able to respond in time to a home invasion and now the murder is blamed on you. I think that the worm deserves more than a misdemeanor, but not capital punishment.
I just installed XP SP1 last night on a new computer. Installation from scratch and all updates took a little over an hour. We had to download the drivers for the SBLive card. The ATI video driver was picked up by WindowsUpdate. Why doesn't Windows make all drivers available? Because they probably can't. There are a couple factors here that I can think of. One is that hardware manufacturers may not want MS to offer their drivers -- they either want people to have the driver installation CD, or they want people to download the drivers from their website. Secondly, even if the drivers were made available to MS, I think MS wants to certify them before offering them, which is probably an arduous task for a mfr. It may be more trouble than its worth for them. These are both areas that OSS has an advantage -- drivers can be freely distributed, and they don't have to pass a certification process by the distributions.
A quick way around this is to go into WINNT\System32\dllcache and replace notepad.exe with the desired file. Then delete WINNT\System32\notepad.exe. When SFP restores the file, it will generally pull a copy from the dllcache folder, and it should use the new file in dllcache. I did a quick test of this replacing notepad.exe with the command prompt, and it seemed to work.
A couple things to note: Linux distributions update the software that comes with the distribution, which is usually quite a bit. I suppose if Windows came with a lot more software (i.e., a "Windows" distribution), Microsoft would provide a way to update this all automatically, but for some reason I don't think most the people here would like that. Secondly, Linux distributions can do what they do because of the nature of the software licenses. With most software designed to run under Windows, MS would have to get permission from each individual company to provide updates. The GPL gives Linux distributions a nice edge when it comes to updating the system.
I have two Windows 2000 machines that I run with absolutely no firewall. They have IPs provided by the cable modem, and they're not NAT'ed. I keep them updated, and in their several years of continuous operation, I have not been infected with anything, nor has my machine been remotely crashed/compromised. Which isn't to say that this is wise, but it just goes to show that it can be done. I'd also point out that Linux can do the same thing, and I'd feel even safer (esp. if I configured iptables).
We have a Win2K machine at work that works at a webserver, and aside from periodic updates, I've seen uptimes over half a year. It would be longer if I didn't have to do updates. Win2K can be stable. I still trust Linux to remain up longer, but given good hardware and software, Win2K is solid. I'm betting you aren't keeping up with kernel updates on your RH box. If I didn't keep up with updates, I'd have an uptime that long too.
...the hardest part of IT work is getting the business people to decide what they want, and keeping what they want the same for the rest of the project.
In many ways, this is actually an important part of an IT job. Figuring out what people really want, and adapting as project requirements evolve. This can be abused, but it goes with the territory. If your job is a modular plug-in job that can be done by anyone, then you're not really seeing the whole picture, or you're not being used to your full capacity.
Having the IT staff train their own out-sourced replacements -- you gotta hand it to them, that's gotta be one of the most bone-headed moves I've heard of. Training your replacement if you're planning on leaving is one thing; training the person who would enable them to lay you off is just plain stupid.
I'll have to try this at home later -- I have a VM with XP Pro installed on it for testing. I've used Remote Desktop before, but I guess I've never actually seen what the host did when I logged in. With TS I know for sure that multiple people can work (kinda like Citrix, SSH, X-Windows tunneling, etc).
I have Win XP Pro, and, while 2 users can be logged in at one time (one at the console, the other over a network), they both cannot be working at the same time.
Not true. Two users using Fast User Switching can't be working at the same time even though both are logged in, but if you're logged in via Terminal Services (aka Remote Desktop), multiple people can work at the same time.
Not true, they say "three times the power", which is 200% more power. Also note that they list efficiencies around 90%, not 300%+ like this guy trying to hawk perpetual motion machines. Sheesh -- I'd have thought more people would understand what a scam that ANY device that gets more than 100% efficiency is. Want to know more, the hard science behind it? Check out http://www.lhup.edu/~dsimanek/museum/unwork.htm
No, I think they'll laugh at him like everyone else here should be. This is basically a free-energy, perpetual-motion-machine hoax.
Minato assures us that he hasn't transcended the laws of physics. The force supplying the unexplained extra power out is generated by the magnetic strength of the permanent magnets embedded in the rotor. "I'm simply harnessing one of the four fundamental forces of nature," he says.
Yeah... the power of magnets. Magic.
But a few months later they were forced to recant their decision after the US patent office recognized his invention and gave him the first of two patents.
US patent office okayed it? It must work!
But if you can accept the idea that Minato's device is able to create motion and torque through its unique, sustainable permanent magnet propulsion system, then it makes sense that he is able to get more out of the unit than he puts in in terms of elctrical power.
Most (if not all) motors use permanent magnets, but there's nothing magical about them. I don't think this is being said enough here, so I'll say it: This is total bullshit. If you want to know what this is really about, go here: http://www.lhup.edu/~dsimanek/museum/unwork.htm
And sometimes you need high performance on code that doesn't have fine-grained parallelism, meaning a Cray vector machine would be slow. So I guess Crays aren't High Performance Computers either.
That statement doesn't make any sense. Code that "doesn't have find-grained parallelism" is just code that can be ran in parallel in large chunks of work (taken to an extreme, you get distributed computing projects like FAH, SETI, UD, etc.), which a Cray would be good at as well, it just wouldn't be using the main advantages of a Cray. The thing that supercomputers like Cray have that clusters don't is high-speed interconnects, and for many supercomputing applications, this is _very_ important. If your interconnect is a LAN or other slow piece of crap, you're bottlenecked horribly when the various nodes need to communicate with each other, and your performance will be shit.
Slashdot Mirror
A recent data recovery operation I did on a Windows XP system: My friend had installed VMWare, and I think he somehow managed to set VMWare to use his physical primary drive as the drive the VM would use. He then installed Windows 95a (FAT16) onto his 40GB drive (which he thought was a virtual drive) which surprisingly worked until he tried to exit the VM and it wouldn't let him. After he rebooted, his machine booted into Windows 95a -- it had overwritten his NTFS drive! He was horrified at all the data he had lost, so I told him I'd try to recover it, although I didn't have much confidence in the situation. Rebuilding the MBR was straightforward -- I just used an MBR from another Win2K machine and modified the partition table to match his drive size (this is something I have a lot of experience with). The boot sector was also straightforward -- NTFS keeps a backup copy of the boot sector at the end of the partition. My main concern was the MFT. The mirror at the center of the drive would be intact (the FAT16 partition used by Win95a was 2GB and wouldn't reach 20GB into the drive), but I didn't think that the MFT mirror was a complete mirror. However, after reading the location of the MFT in the boot sector, it turned out that his MFT was located 3GB into the drive! This surprised me quite a bit. I later checked into this and found that a friend's XP machine had the MFT at about the same spot. My Win2k machine at work had the MFT about 200-400mb into the drive, and my machine at home had the MFT less than 1MB into the drive. Anyway, after a reboot (into an alternate installation of WinXP) and a scan, he was able to recover most of his data, which was a great relief. Some day though, I'll have to try out the MFT mirror thing and see if it would have been enough to recover an NTFS partition...
I remember reading that stores were left over with thousands of those Jar-Jar dolls because no one wanted them. They didn't sell very well, and if you've seen the movie, it would've been pretty obvious.
The parent poster makes an excellent point. I sometimes wonder about SCO's employees, the ones that actually wrote their software. I wonder how they feel about SCO attacking Linux. It would be pretty weird if one day you came to work and found out that your new CEO has decided to attack other companies using baseless accusations because your funding comes from a parasitical IP company. It's the sorta thing that would make you want to quit -- if you weren't already being laid off because the company didn't need to actually produce anything.
I think Baystar wants to keep SCO low profile and is telling McBride to shut his mouth because if more people were aware of their parasitical business model, I think it would hurt them financially and they might suffer some backlash from the public at large. As it stands now, probably only the IT community, and maybe just a subset of that, is aware of the implications of this lawsuit.
Mod parent up. I don't know if I agree, but I think that was the most lucid explanation I've yet seen of what this multi-universe stuff is.
True. But the distinction I think he's trying to describe is like comparing Newtonian physics and Einsteinian physics. For most everyday stuff involving objects we can see and speeds easily measured, Newtonian physics work well (e.g. using F=ma to measure acceleration of a car). But as you approach the speed of light or supermassive objects, Newtonian physics' inaccuracies appear. The more extreme the conditions are, simpler models show their inadequacies and a more detailed accurate model is constructed. The same thing applies with electrons -- the basic model of an electron works reasonably well for things such as building simple electronic circuits and maybe particle collisions (I can't really say for sure, IANAP), but as more extreme conditions are explored, a more detailed model may be needed to explain electron behavior. Maybe it's like a fractal -- the closer you look at the edges, the more details that appear.
I'd also like to note that switching to jdk1.4.2 (I think we were using 1.3.1 previously) under Linux gave our servers an enormous boost in speed -- a lot of CPU/disk/network intensive code ran 5 times as fast. Btw, the JDK we used was from Blackdown.
I think they're blocked now, but they used to not be. I used to go to GRC.com and run that Shields-Up test to see what ports were visible from the outside, and the usual suspects were visible. However, you can simply unbind TCP with File & Printer Sharing (and even Client for MS Networks), and though the ports are visible, they're closed or good as closed.
I think a real big advantage for ATI is the fact that their card doesn't take up two slots, require a monstrosity of a heat sink and fan, and recommend/require a 450W power supply like the '6800 does. Even if the new ATI card wasn't as fast as the 6800, I wouldn't consider buying a video card like that. And I've always considered myself a fan of Nvidia cards (I used to hate the "ATI OS" that ATI's old drivers used to install -- it was very invasive). ATI has produced a very competitive card performance-wise, while keeping the same form factor and with a reasonable (relatively speaking) level of power consumption and heat dissipation.
This is a slippery slope... stretch the chain of events out, and maybe that cop that pulled you over for speeding was delayed just long enough that he wasn't able to respond in time to a home invasion and now the murder is blamed on you. I think that the worm deserves more than a misdemeanor, but not capital punishment.
I just installed XP SP1 last night on a new computer. Installation from scratch and all updates took a little over an hour. We had to download the drivers for the SBLive card. The ATI video driver was picked up by WindowsUpdate. Why doesn't Windows make all drivers available? Because they probably can't. There are a couple factors here that I can think of. One is that hardware manufacturers may not want MS to offer their drivers -- they either want people to have the driver installation CD, or they want people to download the drivers from their website. Secondly, even if the drivers were made available to MS, I think MS wants to certify them before offering them, which is probably an arduous task for a mfr. It may be more trouble than its worth for them. These are both areas that OSS has an advantage -- drivers can be freely distributed, and they don't have to pass a certification process by the distributions.
So when you install Windows, it should come with all the applications you need. Hmmm...
A quick way around this is to go into WINNT\System32\dllcache and replace notepad.exe with the desired file. Then delete WINNT\System32\notepad.exe. When SFP restores the file, it will generally pull a copy from the dllcache folder, and it should use the new file in dllcache. I did a quick test of this replacing notepad.exe with the command prompt, and it seemed to work.
A couple things to note: Linux distributions update the software that comes with the distribution, which is usually quite a bit. I suppose if Windows came with a lot more software (i.e., a "Windows" distribution), Microsoft would provide a way to update this all automatically, but for some reason I don't think most the people here would like that. Secondly, Linux distributions can do what they do because of the nature of the software licenses. With most software designed to run under Windows, MS would have to get permission from each individual company to provide updates. The GPL gives Linux distributions a nice edge when it comes to updating the system.
I have two Windows 2000 machines that I run with absolutely no firewall. They have IPs provided by the cable modem, and they're not NAT'ed. I keep them updated, and in their several years of continuous operation, I have not been infected with anything, nor has my machine been remotely crashed/compromised. Which isn't to say that this is wise, but it just goes to show that it can be done. I'd also point out that Linux can do the same thing, and I'd feel even safer (esp. if I configured iptables).
I just read about this... on Forbe's list of the top ten worst cars ever. Apparently the system was failure prone and sucked ass.
We have a Win2K machine at work that works at a webserver, and aside from periodic updates, I've seen uptimes over half a year. It would be longer if I didn't have to do updates. Win2K can be stable. I still trust Linux to remain up longer, but given good hardware and software, Win2K is solid. I'm betting you aren't keeping up with kernel updates on your RH box. If I didn't keep up with updates, I'd have an uptime that long too.
Having the IT staff train their own out-sourced replacements -- you gotta hand it to them, that's gotta be one of the most bone-headed moves I've heard of. Training your replacement if you're planning on leaving is one thing; training the person who would enable them to lay you off is just plain stupid.
I'll have to try this at home later -- I have a VM with XP Pro installed on it for testing. I've used Remote Desktop before, but I guess I've never actually seen what the host did when I logged in. With TS I know for sure that multiple people can work (kinda like Citrix, SSH, X-Windows tunneling, etc).
Yeah good for him, trying to scam his investors. If it's difficult to buy a Raser, it's probably because it's new and expensive.