GPS chips, at least those designed in the US, are required to not function above 18km altitude or 515m/s for specifically this reason. Otherwise, they are classed as a munition. It's possible to get unrestricted GPS chips from other countries, but you couldn't just strap on any old common GPS chip.
Under this law you could make an incredibly abusive TOS:
It is illegal to "intentionally accesses a computer without authorization or exceeds authorized access, and thereby obtains... information from any protected computer." A definition of "protected computer" is anything connected to the Internet. You could easily write a TOS that forbid any access to a website, and merely loading the homepage would be illegal the way the law is written. Bypassing authentication is not required to break the law. Even worse, if there were no TOS at all, there would be no authorization given. Although it's way too broad, it's apparent why authorization was made the only consideration: If, conversely, if any data a server offered freely to public requests was termed authorized, then injection attacks could be said to be doing exactly what machine was programmed to do unless a TOS specifically detailed what was authorized input. That brings things right back to the TOS being able to define what is authorized. The law needs to be much more detailed to avoid being too broad, yet avoid "it's not a bug, it's a feature" type defenses.
Even WPA/WPA2 Personal isn't too hard to crack in most cases because most people don't use secure passwords. When I set up my wireless network again after moving I decided to go all the way and set up RADIUS using TLS (though I'm not paranoid enough to enable additional OTP auth). Although configuring FreeRADIUS is far more than a typical user would be able to do, most of it could be automated fairly well with reasonable defaults and automatic key generation.
Even if the names don't get changed, they still get upgraded a lot. The power costs are so significant (several million dollars a year) that running a system that's more than a couple years old is completely unfeasible. For example, I have an account on HECToR, which has gone through three/four upgrades since it was first built in 2008: 11,328 2.8GHz cores to 22,646 2.3GHz cores to 44544 cores to 90,112 2.3GHz cores (and ram upgrades along the way for a total of 90TB now). One of those was a two-part upgrade.
Fiber optics can easily contain photon energy of 3 megawats/cm^2 and have a low enough loss to allow for quantum communication, which relies on individual photons being delivered successfully. Yes, photons really are orders of magnitude more reliable in terms of transmission. The hard part about photons is actually the same thing: they're hard to make interact with each other, which is why designing optical quantum gates is hard.
The vast majority of science throughout history has been government funded. Newton, for example, was a member of the University of Cambridge, which has always been funded by the Crown. Individuals contributing to science were largely aristocrats.
Theoretical science simply isn't profitable. Devices using the new knowledge may be, but that's what separates engineering from science. Our research into the Higgs Boson will take a long time, if ever, to have any practical ramifications. However, it is still important, because it lays the groundwork for technology that needs a better understanding of how things works than what we know now. Things like manipulation of spacetime for FTL drives would allow for efficient space trade, but we need to bridge general relativity and quantum physics to even know what, if any, of the proposed possibilities would even work.
TL;DR: There's an almost completely opposing economical spectrum between hard science and technology.
If it's a communications device, it's already had a ton of testing. If you look at a devices packaging and often the back of the product itself, you will see symbols from the FCC and other national telecommunications agencies. Devices are not allowed to emit radiation that may cause interference with non-public frequencies, especially not frequencies as critical as those used for airplanes.
No FCC approved consumer device is allowed to transmit or cause problems on the restricted frequencies used by airplanes to begin with. Devices that do transmit at non-public frequencies require licenses to use. Source: A friend with a PhD in aerospace engineering who has worked for Boeing and Airbus.
Linux has a much sharper definition of privileges than Windows. If UAC was only needed to do important things, and less programs needed root-type privileges to even run, the security would improve a lot. Unfortunately, requiring elevated privileges just for a program to run is not a good idea, and setting a program to always run with elevated privileges is risky similar to how the setuid bit is risky. Microsoft needs to enforce that elevated privileges be required only when absolutely necessary.
Another way of doing things would be to implement a model similar to FreeBSD Jails. This would allow for users to install software safely, as the software would only be able to alter its own files and see its own information. PCBSD's (a derivitive of FreeBSD) graphical package manager does something similar to this, but there isn't a 1:1 jail:program setup. I find this to be quite an elegant solution for home use, where you often want application installation, not just usage, to be multi-user. It wouldn't stop botnet type programs, but it could be implemented to block access to private data files.
IE doesn't even ask about "accelerators" any more. I got a copy of Win7 SP1 and was annoyed on my way to install Firefox, that the only choice I had was Bing and wasn't presented with any other options. Talk about leveraging market share to gain unfair advantage.
The whole idea of PCs started as a way to get away from walled gardens. It seems Microsoft, et al. have forgotten that the very reason for their existence is that people want more freedom than what you get from a rented terminal client. Cloud computing would be a different story if people needed more computing power than what they could reasonably afford to own in full; however, the exact opposite is the case.
One of the problems is that it's called Windows RT. To this point "Windows compatible" has never had such a serious compatibility regression. Sure, some programs broke between windows versions in the past, but ALL applications not working hasn't been the case.
There's also the issue that Microsoft is pushing their monopolist practices really hard: a forced walled garden with WinRT, and a purposefully non-user-friendly standard of EUFI SecureBoot (there were ways proposed to make it just as secure, but muti-os friendly, and microsoft chose the monopolistic method). It seems PCs are becoming a niche. I can only hope it will remain possible to run whatever software you want at all.
I agree. 32 bit a PGAS memory model is silly. Giving each core its own 32 bit address space and using MPI for communication would be much more useful. Then, it could at least be a good learning tool for HPC programming techniques. Right now, it looks pretty useless.
Even GPGPU is limited for what it can do for HPC. There's a lot more to HPC than raw mathematical power. Memory is often the bottleneck, not the FPUs. The reason we even deal with multiple processors is that the performance increase of single cores has nearly stalled, forcing the use of multiple processors. Communication between multiple cores/processors is a very complicated thing, as well, and getting good performance is a lot more complicated than hooking up a bunch of processors in a grid. For example, the supercomputer I work with has 90,112 2.3GHz cores and 90TB ram; 16 cores per chip in 704 blades, interconnected with a 3d torus network topology. It's the memory/cache size and speed and network topology that makes it a supercomputer. You could get the 800TFLOP/s in a much smaller package using GPUs, but the performance would be drastically less. Even with the 64 cores parallella could have, distributing the workload on a 64 core grid isn't easy. GPGPUs use work groups of smaller numbers of cores to make this sharing a bit more easy to manage. They should have at least made the interconnects a 2d torus rather than a grid, thereby reducing the maximum path length in half. In order to do stuff like quantum mechanics, a 5d torus is optimal. Memory access is the key. This is a bit like comparing apples to oranges, but that's exactly my point: the thing is not a supercomputer.
The main problem is that the variety that is sold in the US is bread for size, not taste. The little ones you get in the tropics are way tastier. Cash crop bananas are about the most altered fruit there is.
Bitcoins != Bitcoin wallets. The breakage wasn't with bitcoin, but people deciding to store PRIVATE KEYS online. There's a big difference between a protocol being broken and people being stupid with private keys.
Of course, the problem is that he's trying to use lack of certainty as an excuse to to avoid taking any action, despite the fact that the science doesn't say anything at all about the best way to fix the issue (or indeed whether it needs fixing...)
This is exactly the problem I have with his response. We went to war with Iraq over the possibility that Saddam had WMDs. Now, the entire world is confirmed to be affected by global warming, and the Republicans want to wait to know exactly how fucked we really are before doing anything about it.
Yeah, nastiest bug I've encountered was from early kdevelop4. I canceled the "make project" dialog, and the project directory hadn't been entered, and was ~ (and not marked invalid due to no project name). Canceling it did an rm -rf ~/*.
KDevelop is starting to get there as well. It's still a bit buggy and needs some usability improvements (such as more automatic ctags stuff), but it's made quite a bit of progress and keeps getting better. I personally quite like the feature that gives each local variable its own tint (saturation can be adjusted in the config, mine's at ~10%); It takes a few hours to get used to and therefore not be a distraction, but once comfortable, the color coding makes visually scanning code easier to track. It's not quite up to Qt Creator level, yet, but it's something to keep an eye on.
As a Gentoo user who likes to cut out bloat, I HATE systemd, and now udev. udev now has networkmanager, systemd, and udev all as part of the same project, now. This huge bloat now makes it so that you can't even boot without an initramdisk (something I don't use) if you mount/usr on a different partition (which I do, because my/usr lives on an ssd). Not to mention, networkmanager uses a binary format for config. It's especially bad for embedded systems. I really don't want anything remotely like a Registry in linux. udev just keeps growing. Things like hotplug were perfectly fine as separate services. There's no reason to make everything monolithic, and is very contrary to Unix design philosophy.
If anything, I'd like to see things moving the direction of plan9. For example: wayland. A great way to make it very unix-y would be to make a virtual filesystem to handle creating windows in the style of plan9's/net. Have a command/syscall mkopengl, which would create/dev/opengl/{context,control,input,monitor,fb,data/}. context and control would have things like the window location and size. Updating those would just mean writing to those files. OpenGL instructions would be written to input. Reading from monitor would stream OpenGL commands written to that context. Reading from fb would grab the fully rendered frames (just cat into a file/encode from to record). Finally, data could have file descriptors to textures in video memory. Not only would a setup like this be powerful in terms of being able to chain programs together, but network transparency would simply be a matter of union mounting the remote/dev/opengl on top of the local/dev/opengl (the way NAT, etc works on plan9), or cat-ing from monitor into a local input "file".
fdupes is in Cygwin. No need for a livecd or vm. Cygwin can handle ntfs hardlinks well, too (I used a python/sh script in cygwin just a few weeks ago to hard link a bunch of photographs into a series of dvd-sized folders).
Also, in the future, I recommend using rsync to back up files (either remotely or to another directory). Rsync can do incremental backups using deltas (a binary version of diffs), which saves a ton of space. Rsync also works well in cygwin, and there are cron daemons available that will install themselves as windows services for automatic backup.
Re:I Guess This Is What Happens When I Don't Watch
on
The Case Against DNA
·
· Score: 2
Indeed, and it's not exactly hard for someone to do the opposite and significantly reduce their own dna output if they were trying to frame someone. The semiconductor industry has been improving techniques to keep the rate of human contamination from technicians/researchers down in their clean rooms for decades, and a couple orders of magnitude difference is considered basic.
Only our lowest courts use juries. Panels of experts (multiple judges) has generally worked quite well in higher courts. Stuff like this and the Apple vs Samsung case make me wonder how great of an idea trial by jury really is. Sure, it's better than trial by fire or combat, but we have a representative democracy precisely because the average person lacks the time and/or training to comprehend the framework they are trying to make decisions from. There are so many things wrong with our criminal justice system.
Re:I Guess This Is What Happens When I Don't Watch
on
The Case Against DNA
·
· Score: 5, Insightful
Plus, if they had any idea as to how science works, they would know that one datum doesn't constitute reliable evidence at all. Even finding the same DNA in multiple locations doesn't rule out systematic contamination. Multiple types of evidence are needed to confirm causation with anything. It scares me that people's entire lives can depend on methods that would get any scientist laughed out of the room.
Slashdot Mirror
I just hope this doesn't end up like ACPI, where everything is broken and only companies with secret specs can be made to work easily.
GPS chips, at least those designed in the US, are required to not function above 18km altitude or 515m/s for specifically this reason. Otherwise, they are classed as a munition. It's possible to get unrestricted GPS chips from other countries, but you couldn't just strap on any old common GPS chip.
Under this law you could make an incredibly abusive TOS: It is illegal to "intentionally accesses a computer without authorization or exceeds authorized access, and thereby obtains ... information from any protected computer." A definition of "protected computer" is anything connected to the Internet. You could easily write a TOS that forbid any access to a website, and merely loading the homepage would be illegal the way the law is written. Bypassing authentication is not required to break the law. Even worse, if there were no TOS at all, there would be no authorization given. Although it's way too broad, it's apparent why authorization was made the only consideration: If, conversely, if any data a server offered freely to public requests was termed authorized, then injection attacks could be said to be doing exactly what machine was programmed to do unless a TOS specifically detailed what was authorized input. That brings things right back to the TOS being able to define what is authorized. The law needs to be much more detailed to avoid being too broad, yet avoid "it's not a bug, it's a feature" type defenses.
Even WPA/WPA2 Personal isn't too hard to crack in most cases because most people don't use secure passwords. When I set up my wireless network again after moving I decided to go all the way and set up RADIUS using TLS (though I'm not paranoid enough to enable additional OTP auth). Although configuring FreeRADIUS is far more than a typical user would be able to do, most of it could be automated fairly well with reasonable defaults and automatic key generation.
Even if the names don't get changed, they still get upgraded a lot. The power costs are so significant (several million dollars a year) that running a system that's more than a couple years old is completely unfeasible. For example, I have an account on HECToR, which has gone through three/four upgrades since it was first built in 2008: 11,328 2.8GHz cores to 22,646 2.3GHz cores to 44544 cores to 90,112 2.3GHz cores (and ram upgrades along the way for a total of 90TB now). One of those was a two-part upgrade.
Fiber optics can easily contain photon energy of 3 megawats/cm^2 and have a low enough loss to allow for quantum communication, which relies on individual photons being delivered successfully. Yes, photons really are orders of magnitude more reliable in terms of transmission. The hard part about photons is actually the same thing: they're hard to make interact with each other, which is why designing optical quantum gates is hard.
Just because the humans are ignorant of it doesn't mean the dolphins are.
The vast majority of science throughout history has been government funded. Newton, for example, was a member of the University of Cambridge, which has always been funded by the Crown. Individuals contributing to science were largely aristocrats.
Theoretical science simply isn't profitable. Devices using the new knowledge may be, but that's what separates engineering from science. Our research into the Higgs Boson will take a long time, if ever, to have any practical ramifications. However, it is still important, because it lays the groundwork for technology that needs a better understanding of how things works than what we know now. Things like manipulation of spacetime for FTL drives would allow for efficient space trade, but we need to bridge general relativity and quantum physics to even know what, if any, of the proposed possibilities would even work.
TL;DR: There's an almost completely opposing economical spectrum between hard science and technology.
If it's a communications device, it's already had a ton of testing. If you look at a devices packaging and often the back of the product itself, you will see symbols from the FCC and other national telecommunications agencies. Devices are not allowed to emit radiation that may cause interference with non-public frequencies, especially not frequencies as critical as those used for airplanes.
No FCC approved consumer device is allowed to transmit or cause problems on the restricted frequencies used by airplanes to begin with. Devices that do transmit at non-public frequencies require licenses to use. Source: A friend with a PhD in aerospace engineering who has worked for Boeing and Airbus.
Linux has a much sharper definition of privileges than Windows. If UAC was only needed to do important things, and less programs needed root-type privileges to even run, the security would improve a lot. Unfortunately, requiring elevated privileges just for a program to run is not a good idea, and setting a program to always run with elevated privileges is risky similar to how the setuid bit is risky. Microsoft needs to enforce that elevated privileges be required only when absolutely necessary.
Another way of doing things would be to implement a model similar to FreeBSD Jails. This would allow for users to install software safely, as the software would only be able to alter its own files and see its own information. PCBSD's (a derivitive of FreeBSD) graphical package manager does something similar to this, but there isn't a 1:1 jail:program setup. I find this to be quite an elegant solution for home use, where you often want application installation, not just usage, to be multi-user. It wouldn't stop botnet type programs, but it could be implemented to block access to private data files.
IE doesn't even ask about "accelerators" any more. I got a copy of Win7 SP1 and was annoyed on my way to install Firefox, that the only choice I had was Bing and wasn't presented with any other options. Talk about leveraging market share to gain unfair advantage.
The whole idea of PCs started as a way to get away from walled gardens. It seems Microsoft, et al. have forgotten that the very reason for their existence is that people want more freedom than what you get from a rented terminal client. Cloud computing would be a different story if people needed more computing power than what they could reasonably afford to own in full; however, the exact opposite is the case.
One of the problems is that it's called Windows RT. To this point "Windows compatible" has never had such a serious compatibility regression. Sure, some programs broke between windows versions in the past, but ALL applications not working hasn't been the case.
There's also the issue that Microsoft is pushing their monopolist practices really hard: a forced walled garden with WinRT, and a purposefully non-user-friendly standard of EUFI SecureBoot (there were ways proposed to make it just as secure, but muti-os friendly, and microsoft chose the monopolistic method). It seems PCs are becoming a niche. I can only hope it will remain possible to run whatever software you want at all.
I agree. 32 bit a PGAS memory model is silly. Giving each core its own 32 bit address space and using MPI for communication would be much more useful. Then, it could at least be a good learning tool for HPC programming techniques. Right now, it looks pretty useless.
Even GPGPU is limited for what it can do for HPC. There's a lot more to HPC than raw mathematical power. Memory is often the bottleneck, not the FPUs. The reason we even deal with multiple processors is that the performance increase of single cores has nearly stalled, forcing the use of multiple processors. Communication between multiple cores/processors is a very complicated thing, as well, and getting good performance is a lot more complicated than hooking up a bunch of processors in a grid. For example, the supercomputer I work with has 90,112 2.3GHz cores and 90TB ram; 16 cores per chip in 704 blades, interconnected with a 3d torus network topology. It's the memory/cache size and speed and network topology that makes it a supercomputer. You could get the 800TFLOP/s in a much smaller package using GPUs, but the performance would be drastically less. Even with the 64 cores parallella could have, distributing the workload on a 64 core grid isn't easy. GPGPUs use work groups of smaller numbers of cores to make this sharing a bit more easy to manage. They should have at least made the interconnects a 2d torus rather than a grid, thereby reducing the maximum path length in half. In order to do stuff like quantum mechanics, a 5d torus is optimal. Memory access is the key. This is a bit like comparing apples to oranges, but that's exactly my point: the thing is not a supercomputer.
The main problem is that the variety that is sold in the US is bread for size, not taste. The little ones you get in the tropics are way tastier. Cash crop bananas are about the most altered fruit there is.
Bitcoins != Bitcoin wallets. The breakage wasn't with bitcoin, but people deciding to store PRIVATE KEYS online. There's a big difference between a protocol being broken and people being stupid with private keys.
Of course, the problem is that he's trying to use lack of certainty as an excuse to to avoid taking any action, despite the fact that the science doesn't say anything at all about the best way to fix the issue (or indeed whether it needs fixing...)
This is exactly the problem I have with his response. We went to war with Iraq over the possibility that Saddam had WMDs. Now, the entire world is confirmed to be affected by global warming, and the Republicans want to wait to know exactly how fucked we really are before doing anything about it.
It really depends on if they choose do more with that personhood than the average bar of Soylent Green.
Yeah, nastiest bug I've encountered was from early kdevelop4. I canceled the "make project" dialog, and the project directory hadn't been entered, and was ~ (and not marked invalid due to no project name). Canceling it did an rm -rf ~/*.
KDevelop is starting to get there as well. It's still a bit buggy and needs some usability improvements (such as more automatic ctags stuff), but it's made quite a bit of progress and keeps getting better. I personally quite like the feature that gives each local variable its own tint (saturation can be adjusted in the config, mine's at ~10%); It takes a few hours to get used to and therefore not be a distraction, but once comfortable, the color coding makes visually scanning code easier to track. It's not quite up to Qt Creator level, yet, but it's something to keep an eye on.
As a Gentoo user who likes to cut out bloat, I HATE systemd, and now udev. udev now has networkmanager, systemd, and udev all as part of the same project, now. This huge bloat now makes it so that you can't even boot without an initramdisk (something I don't use) if you mount /usr on a different partition (which I do, because my /usr lives on an ssd). Not to mention, networkmanager uses a binary format for config. It's especially bad for embedded systems. I really don't want anything remotely like a Registry in linux. udev just keeps growing. Things like hotplug were perfectly fine as separate services. There's no reason to make everything monolithic, and is very contrary to Unix design philosophy.
/net. Have a command/syscall mkopengl, which would create /dev/opengl/{context,control,input,monitor,fb,data/}. context and control would have things like the window location and size. Updating those would just mean writing to those files. OpenGL instructions would be written to input. Reading from monitor would stream OpenGL commands written to that context. Reading from fb would grab the fully rendered frames (just cat into a file/encode from to record). Finally, data could have file descriptors to textures in video memory. Not only would a setup like this be powerful in terms of being able to chain programs together, but network transparency would simply be a matter of union mounting the remote /dev/opengl on top of the local /dev/opengl (the way NAT, etc works on plan9), or cat-ing from monitor into a local input "file".
If anything, I'd like to see things moving the direction of plan9. For example: wayland. A great way to make it very unix-y would be to make a virtual filesystem to handle creating windows in the style of plan9's
fdupes is in Cygwin. No need for a livecd or vm. Cygwin can handle ntfs hardlinks well, too (I used a python/sh script in cygwin just a few weeks ago to hard link a bunch of photographs into a series of dvd-sized folders).
Also, in the future, I recommend using rsync to back up files (either remotely or to another directory). Rsync can do incremental backups using deltas (a binary version of diffs), which saves a ton of space. Rsync also works well in cygwin, and there are cron daemons available that will install themselves as windows services for automatic backup.
Indeed, and it's not exactly hard for someone to do the opposite and significantly reduce their own dna output if they were trying to frame someone. The semiconductor industry has been improving techniques to keep the rate of human contamination from technicians/researchers down in their clean rooms for decades, and a couple orders of magnitude difference is considered basic.
Only our lowest courts use juries. Panels of experts (multiple judges) has generally worked quite well in higher courts. Stuff like this and the Apple vs Samsung case make me wonder how great of an idea trial by jury really is. Sure, it's better than trial by fire or combat, but we have a representative democracy precisely because the average person lacks the time and/or training to comprehend the framework they are trying to make decisions from. There are so many things wrong with our criminal justice system.
Plus, if they had any idea as to how science works, they would know that one datum doesn't constitute reliable evidence at all. Even finding the same DNA in multiple locations doesn't rule out systematic contamination. Multiple types of evidence are needed to confirm causation with anything. It scares me that people's entire lives can depend on methods that would get any scientist laughed out of the room.