We're not talking about criminal records or warcrimes here.
Yes we are. The example the EU even uses is someone commits a crime, serves their time, and no longer wants to be associated with that crime. That person can request Google, FB, blogs, etc to have to remove any data pertaining to that person having committed that crime. Also, these sites must actively screen new content that may contain this information.
Essentially you have a black-list that will only grow, plus the same content may be worded many different ways. It will be the cat-and-mouse game the YouTube is playing with copyrighted content; except much easier for YouTube to get punished.
Say I post your address, so you request FB to take it down. Then I repost, the someone else's shares that post. FB has to take both down. Say I post on a different site then post a link on FB, they also has to take that down.
I'm not talking about you making many requests each time I repost that info, I'm talking about you only make the request once and FB has to actively monitor and watch for that information from any and all inputs until the end of time. Every new post must be screened against the black-list of "blocked" data.
Good luck with that because that's what the EU wants.
FB isn't supposed to allow linking to removed data. If someone copies data from FB to MySpace, then FB gets a request to remove said data, FB must also filter any links that could be referencing the copied versions on MySpace's site.
EU wants Google to do the same. If someone requests certain data removed, Google must also remove all references to any/all copies any user has made anywhere on the internet, otherwise get fined.
i++ is atomic if the CPU can handle working at the size of i. On a 32bit machine, if "i" is 64bit, then it won't be atomic, but on a 64bit machine, i++ is atomic. Newer x86 CPUs do support 128bit and soon 256bit atomic memory reads and writes via SIMD registers.
Now, it is not guaranteed to be pushed out to memory immediately. If you're doing some like for(i = 0;... i++), "i" may be stored in a register and never pushed back out to memory. You need to declare i as volatile or use a memory fence to force i to be pushed.
Volatile doesn't fix everything either. All it does is say the update will be pushed out to memory, but because most CPUs use Out-of-Order execution, the read/write may get re-ordered.
eg.
shared volatile int i = 0
shared volatile bool b = false
threaded method()
{
i = 3
b = true
}
Even if the compiler doesn't re-order these two assignments, the CPU might because "i" doesn't have a detectable dependency.
Can you image the amount of heat 64cores of Bulldozers in a 1U case would create?!
I've been looking at custom building a server at home and when looking in the 64GB-256GB of ram with single-dual socket 8-32 threads, Intel wins in price, performance, and power consumption.
Choosing between an Intel Xeon i5 3.3ghz quad+HT with 65watt TDP compared to an Opteron Bulldozer 4module-8core 2.8ghz with a 125watt TDP and lower IPC, for an almost identical price isn't even a choice.
The most common way to acquire your password hash is via SQL injection, the web server is rarely compromised. Short of breaking the password storage, the hacker won't gain access to your password unless it's weak enough to generate.
A strong 12 char password has 612,709,757,329,767,363,772,416 possibilities, which even at quadrillion(10^15) guesses per second, it would take over 7,000 years to exhaust the entire space.
"A 14 character Windows XP password hashed using LM for example, would fall in just six minutes"
Which is nothing impressive. NTLM has a 14 char password max and pads sub-14char passwords with null. It then breaks the password into two 7 byte pieces, hashes both pieces, then concatenates the two hashes together. Using NTLM, a 14 char password at worst 2*96^7 instead of 96^14, which is a factor of 37,572,373,905,408 difference. If NTLM was properly designed, that same 14 char password would have taken 37,572,373,905,408*6min to break or 428,908,378 years.
14 char passwords are still safe assuming there isn't a huge flaw in the password storage.
It's easy to point out a system that is bad, but very hard to design a system that is good. I went the route of pointing out the bad, but never claimed to have a better system.
All I know is the current system has a huge failure and fixing that issue with almost any other system will do better. If we're willing to pay more for people to visit the ER, why not let them get regular check-ups and save some money? Either way, the monetary worse case is the same.
Another fun point. On average it costs the USA economy $250k to raise a child to 18. If that person dies or is not healthy enough to do good work before paying back their debt, we're losing out on money. Health care is an investment as much as basic education.
I am not comparing NHS vs USA's system, just saying any health care system with current day problems should not have a "waiting list" for life threatening issues unless it's because of needing a kidney/liver/etc transplant. In general, if there is a waiting list, hire more doctors or find a way to better weed out people who don't need it.
The NHS example is just another extreme, which is effectively a strawman for the universal healthcare as an idea. NHS is just a poorly implemented version.
I was in the ER recently and while waiting around, I got to talk to a nurse who told me that most of the people they see in the ER are just people who could not afford their regular check-up, so they just come to the ER claiming to have a problem.
The hospital is ethically and legally obliged to take them in and run tests, which cost A LOT more than a regular check-up. Because the price is too high for the patient, the hospital just soaks the loss and raises the prices for everyone else who does have insurance, which causes insurance rates to go up, which causes more people to lose insurance, which causes more people to come into the ER instead of getting regular check ups.
Do you see the problem? It's a positive feedback system that reduces efficiency and increases costs.
A large amount of what makes a modern computer from GPU tech to CPU tech to high performance multi-core kernel designs(Think Linux and FreeBSD) have come from MS R&D in the form of "free research". MS R&D is the largest computer tech R&D in the world. Combine IBM, Intel, and AMD, and you get an idea of their size.
Win7 gets better battery life with composition because it offloads to the GPU instead of using the CPU. Vista had a lot of issues with RTM, not sure how SP3 is doing, but most people I know skipped Vista anyway.
Except the farmers probably can't afford the initial connection, but the state sure thinks it's worth while. The state is expecting to actually increase overall income, but the farmers won't notice much. Kind of funny how the individual won't see much benefit, but the group as a whole does.
Quite a few people in the country side are farmers. Are you saying we should punish farmers? Why not just get rid of public education while we're at it?
Getting farmers connected is important, enough so that Minnesota is spending its own money to run 100/100 fiber out to farms that are 35 miles outside the city.
IBM's senior research engineer thought we'd have batteries with 100x the storage in the next 10 years and he only said this a few years back. I have read about a new battery tech that was in the safety testing phase that could recharge 10x faster than current batteries and could hold about 10x-100x the charge for the same size. It already works functionally, it just needs to be shown to not be a fire hazard and pass a lot of testing.
Too many patents for the private sector to navigate. If the government does the research and gives it out for free, then there's no questions. I don't know, just babbling.
The elderly are the strongest voting population, they would never let that happen. Any politician that gets even a finger pointed at them for tampering with SS will get voted out of office.
Random tale of my friend: He purchased a 450watt cheap PSU and it ran hot for a few months then burned out. He swapped it out for a 250watt namebrand PSU and the system ran another 5 years.
Slashdot Mirror
We're not talking about criminal records or warcrimes here.
Yes we are. The example the EU even uses is someone commits a crime, serves their time, and no longer wants to be associated with that crime. That person can request Google, FB, blogs, etc to have to remove any data pertaining to that person having committed that crime. Also, these sites must actively screen new content that may contain this information.
Essentially you have a black-list that will only grow, plus the same content may be worded many different ways. It will be the cat-and-mouse game the YouTube is playing with copyrighted content; except much easier for YouTube to get punished.
It's much more than that.
Say I post your address, so you request FB to take it down. Then I repost, the someone else's shares that post. FB has to take both down. Say I post on a different site then post a link on FB, they also has to take that down.
I'm not talking about you making many requests each time I repost that info, I'm talking about you only make the request once and FB has to actively monitor and watch for that information from any and all inputs until the end of time. Every new post must be screened against the black-list of "blocked" data.
Good luck with that because that's what the EU wants.
FB isn't supposed to allow linking to removed data. If someone copies data from FB to MySpace, then FB gets a request to remove said data, FB must also filter any links that could be referencing the copied versions on MySpace's site.
EU wants Google to do the same. If someone requests certain data removed, Google must also remove all references to any/all copies any user has made anywhere on the internet, otherwise get fined.
i++ is atomic if the CPU can handle working at the size of i. On a 32bit machine, if "i" is 64bit, then it won't be atomic, but on a 64bit machine, i++ is atomic. Newer x86 CPUs do support 128bit and soon 256bit atomic memory reads and writes via SIMD registers.
... i++), "i" may be stored in a register and never pushed back out to memory. You need to declare i as volatile or use a memory fence to force i to be pushed.
Now, it is not guaranteed to be pushed out to memory immediately. If you're doing some like for(i = 0;
Volatile doesn't fix everything either. All it does is say the update will be pushed out to memory, but because most CPUs use Out-of-Order execution, the read/write may get re-ordered.
eg.
shared volatile int i = 0
shared volatile bool b = false
threaded method()
{
i = 3
b = true
}
Even if the compiler doesn't re-order these two assignments, the CPU might because "i" doesn't have a detectable dependency.
Can you image the amount of heat 64cores of Bulldozers in a 1U case would create?!
I've been looking at custom building a server at home and when looking in the 64GB-256GB of ram with single-dual socket 8-32 threads, Intel wins in price, performance, and power consumption.
Choosing between an Intel Xeon i5 3.3ghz quad+HT with 65watt TDP compared to an Opteron Bulldozer 4module-8core 2.8ghz with a 125watt TDP and lower IPC, for an almost identical price isn't even a choice.
I want support for Linux 1.0. What do you mean I should just upgrade to 2.6+? I thought you said old versions should be supported.
Violent crime has been going down ever since video games came to the home. :-)
The most common way to acquire your password hash is via SQL injection, the web server is rarely compromised. Short of breaking the password storage, the hacker won't gain access to your password unless it's weak enough to generate.
A strong 12 char password has 612,709,757,329,767,363,772,416 possibilities, which even at quadrillion(10^15) guesses per second, it would take over 7,000 years to exhaust the entire space.
"A 14 character Windows XP password hashed using LM for example, would fall in just six minutes"
Which is nothing impressive. NTLM has a 14 char password max and pads sub-14char passwords with null. It then breaks the password into two 7 byte pieces, hashes both pieces, then concatenates the two hashes together. Using NTLM, a 14 char password at worst 2*96^7 instead of 96^14, which is a factor of 37,572,373,905,408 difference. If NTLM was properly designed, that same 14 char password would have taken 37,572,373,905,408*6min to break or 428,908,378 years.
14 char passwords are still safe assuming there isn't a huge flaw in the password storage.
It's easy to point out a system that is bad, but very hard to design a system that is good. I went the route of pointing out the bad, but never claimed to have a better system.
All I know is the current system has a huge failure and fixing that issue with almost any other system will do better. If we're willing to pay more for people to visit the ER, why not let them get regular check-ups and save some money? Either way, the monetary worse case is the same.
Another fun point. On average it costs the USA economy $250k to raise a child to 18. If that person dies or is not healthy enough to do good work before paying back their debt, we're losing out on money. Health care is an investment as much as basic education.
I am not comparing NHS vs USA's system, just saying any health care system with current day problems should not have a "waiting list" for life threatening issues unless it's because of needing a kidney/liver/etc transplant. In general, if there is a waiting list, hire more doctors or find a way to better weed out people who don't need it.
The NHS example is just another extreme, which is effectively a strawman for the universal healthcare as an idea. NHS is just a poorly implemented version.
I was in the ER recently and while waiting around, I got to talk to a nurse who told me that most of the people they see in the ER are just people who could not afford their regular check-up, so they just come to the ER claiming to have a problem.
The hospital is ethically and legally obliged to take them in and run tests, which cost A LOT more than a regular check-up. Because the price is too high for the patient, the hospital just soaks the loss and raises the prices for everyone else who does have insurance, which causes insurance rates to go up, which causes more people to lose insurance, which causes more people to come into the ER instead of getting regular check ups.
Do you see the problem? It's a positive feedback system that reduces efficiency and increases costs.
For whom are we supposed to vote? Rapist 1 or Rapist 2? Either way we get raped.
A large amount of what makes a modern computer from GPU tech to CPU tech to high performance multi-core kernel designs(Think Linux and FreeBSD) have come from MS R&D in the form of "free research". MS R&D is the largest computer tech R&D in the world. Combine IBM, Intel, and AMD, and you get an idea of their size.
My question is are they selling individual data or aggregated anonymous analytic data. Big difference.
Win7 gets better battery life with composition because it offloads to the GPU instead of using the CPU. Vista had a lot of issues with RTM, not sure how SP3 is doing, but most people I know skipped Vista anyway.
Except the farmers probably can't afford the initial connection, but the state sure thinks it's worth while. The state is expecting to actually increase overall income, but the farmers won't notice much. Kind of funny how the individual won't see much benefit, but the group as a whole does.
Quite a few people in the country side are farmers. Are you saying we should punish farmers? Why not just get rid of public education while we're at it?
Getting farmers connected is important, enough so that Minnesota is spending its own money to run 100/100 fiber out to farms that are 35 miles outside the city.
IBM's senior research engineer thought we'd have batteries with 100x the storage in the next 10 years and he only said this a few years back. I have read about a new battery tech that was in the safety testing phase that could recharge 10x faster than current batteries and could hold about 10x-100x the charge for the same size. It already works functionally, it just needs to be shown to not be a fire hazard and pass a lot of testing.
Too many patents for the private sector to navigate. If the government does the research and gives it out for free, then there's no questions. I don't know, just babbling.
21b solar masses http://en.wikipedia.org/wiki/NGC_4889
The elderly are the strongest voting population, they would never let that happen. Any politician that gets even a finger pointed at them for tampering with SS will get voted out of office.
You can submit a bug on their web-site then watch people vote it up/down and add comments. Many of these bug reports get replies from a developer.
The analog version displayed on the monitor, the stored version in the brain.
Random tale of my friend: He purchased a 450watt cheap PSU and it ran hot for a few months then burned out. He swapped it out for a 250watt namebrand PSU and the system ran another 5 years.
Goto http://www.jonnyguru.com/ or http://www.hardwaresecrets.com/ and read some PSU reviews. HWSecrets has a really good article somewhere on the different ways to measure a PSU.
I now look for low-heat and jitter under above-rated load while maintaining excellent efficiency.