The goal is to make the consequences of a first-strike so terrible that no one would embark upon such a course.
Well, that may be your goal. I believe that conventional weapons could make first-strike sufficiently terrible that no one would embark on such a course. I also believe that such a policy would be safer in case of an accidental launch.
I find it odd that people have a problem with the United States having weapons of mass destruction. Given that we have been in the Korean war, the Vietnam War, two Gulf wars and countless military operations without using them.
Korea and Vietnam would have prompted nuclear responses from at least Russia. In the Gulf wars, the use would have been fairly pointless.
They don't think about the need to defend against American nukes because they know that the US refrains from using nukes except when attacked by nukes.
The US has a clear nuclear first-strike policy. Nuclear weapons (specifically bunker-busters) were definitely considered for both Afghanistan and Iraq, but they were (fortunately, IMHO) ultimately not used.
If you knew that everyone in your country would be dead, there wouldn't be any point in doing the retributive strike. For something less than that (which is almost certain to be the case; it's pretty hard to kill everyone), it could certainly be moral to launch a retributive strike -- unless your morals forbid killing people in war in general.
So, you do agree then, that our conventional military force is incapable of carrying out the war aims of any war we've fought in the last 60 years (save the first Persian Gulf)?
It isn't like the non-conventional military force is doing any better.
The fact still remains that it is reasonably easy for a modern military force to outdo Nagasaki with conventional weapons.
What's the problem? If you need to store 1TB and you can live with the slowness of hard drives, you use hard drives. If you can't, you use SSD's -- or possibly an array of small fast hard drives, but SSD's will most often be cheaper.
What you don't do is build a hybrid system with automatic page migration between SSD and hard drives -- and that is what the article assumes will be commonly used. Hierarchial storage has a very small niche, and SSD's won't make it more popular.
The article I read spent a good deal of time talking about flash memory. What article are YOU referring to?
The article treats flash as something you place in between hard drives and memory. This turned out not to happen (with a few exceptions). SSD's simply replace hard drives. Hybrid systems are rare, and it doesn't look like they will become more common -- either you can live with the slowness of hard drives, or you can't. The mainstream will switch to SSD's for everything except backup applications.
There are some hybrid SAN's, but they're damn expensive. At that price they have a hard time competing with simpler pure-flash SAN's.
In the investigation of a recent Airbus airliner crash, it was decided [...]
and then
It wasn't from the investigation, that's how the airbus is built
So it WASN'T decided in the investigation that the onboard computers stopped the pilot from controlling the plane. You can guess as much as you want, but if it was so simple, the investigators would probably have figured it out.
Anyway, what the telemetry shows is that the onboard computers detected that their instruments provided inconsistent input, and therefore they switched to the lowest so-called "law". At the lowest "law", the pilots have almost complete freedom to do stupid things to the plane.
We will likely never know precisely what went wrong in this case, since the black boxes weren't found.
In the investigation of a recent Airbus airliner crash, it was decided that the on-board computer systems did not allow the pilots to override the controls and the plane crashed.
Citation needed. As far as I know, this is not true.
Good luck with fitting your jobs in the tiny amount of memory and flash. If your chips are 1/10th as fast, you need to run 10 copies of your jobs for the same performance, and since you can't do SMP/NUMA, that means 10 times the memory.
And then you need the 64-port ethernet switch to link it all.
They would need to make their own chipset to link the CPU's. Probably beyond Facebook's resources.
You might have to implement better synchronisation primitives for the ARM architecture as well. It has taken x86 a long time to reach the point where 64-core servers actually deliver decent performance, and it will likely take a while for ARM too.
Mainframes aren't known for good CPU performance. Brilliant at I/O, but if Facebook is I/O-limited they're either doing something very wrong or something very right.
You can't get massively-SMP (or NUMA) ARM machines. That means you're stuck with using lots and lots of machines, which again means you'll use too much space and spend too much money on switches etc.
No, he just found that RAID controllers suck. Which they do, universally, all the time. The only ones that actually perform decently are the ones in external SAN boxes, and inside they are typically servers with software RAID...
Following this logic, ancient computer makers should have been sued years ago for bundling their OS on their mainframes. There could have been an independent market for OSes.
Err, that's exactly what happened.
"Then in January 1969 the US Justice Department brought an antitrust action against IBM for monopolizing the computer market. At the time IBM sold its hardware, software, training, and all services as a bundled product. That is, if someone wanted the mainframe software they also had to purchase hardware, training, and everything else from IBM. So in the July 1969 IBM signed another consent decree to unbundle which led to the development of hundreds of companies for supplying software (like University Computing and Computer Associates), hardware (disk drives, memory, and the like)."
Their FAQ claims the car is a great lasting investment due to lack of complexity and moving parts, but having to drop $25k every 6 years for a new battery would be a deal breaker.
There should be good chances that the old batteries will be worth a decent amount of money, because the raw materials are so expensive. There are also batteries on the way using much cheaper materials, so it is possible that the second battery will only be a fraction of the price (but in that case you probably won't get much for your now-obsolete battery).
The "if it's statically linked then the entire blob is now GPLed" gets in the way of perfectly legitimate software designs where dynamic linking is not possible or desirable.
You seem to be under the misconception that dynamic linking gets around the GPL. It doesn't. It can, in certain cases, make a difference for the LGPL.
You can't subnet it. Well you can, if you assign everything statically and you avoid certain implementations of IPv6, but it's a bad idea if you want things to "just work" -- and one of the advantages of IPv6 is the way that things "just work".
So with just a/64 you can't make a separate network for guest Wifi, or a DMZ, or for separate departments. All that is easy to do with IPv4 and NAT, so if ISP's only give out a/64, NAT will quickly get popular in IPv6 too.
Slashdot Mirror
The goal is to make the consequences of a first-strike so terrible that no one would embark upon such a course.
Well, that may be your goal. I believe that conventional weapons could make first-strike sufficiently terrible that no one would embark on such a course. I also believe that such a policy would be safer in case of an accidental launch.
I find it odd that people have a problem with the United States having weapons of mass destruction. Given that we have been in the Korean war, the Vietnam War, two Gulf wars and countless military operations without using them.
Korea and Vietnam would have prompted nuclear responses from at least Russia. In the Gulf wars, the use would have been fairly pointless.
They don't think about the need to defend against American nukes because they know that the US refrains from using nukes except when attacked by nukes.
The US has a clear nuclear first-strike policy. Nuclear weapons (specifically bunker-busters) were definitely considered for both Afghanistan and Iraq, but they were (fortunately, IMHO) ultimately not used.
If you knew that everyone in your country would be dead, there wouldn't be any point in doing the retributive strike. For something less than that (which is almost certain to be the case; it's pretty hard to kill everyone), it could certainly be moral to launch a retributive strike -- unless your morals forbid killing people in war in general.
So, you do agree then, that our conventional military force is incapable of carrying out the war aims of any war we've fought in the last 60 years (save the first Persian Gulf)?
It isn't like the non-conventional military force is doing any better.
The fact still remains that it is reasonably easy for a modern military force to outdo Nagasaki with conventional weapons.
Intel SSD's are inexpensive per IOPS, compared to just about anything else. They aren't even particularly expensive per GB compared to 15K SAS drives.
What's the problem? If you need to store 1TB and you can live with the slowness of hard drives, you use hard drives. If you can't, you use SSD's -- or possibly an array of small fast hard drives, but SSD's will most often be cheaper.
What you don't do is build a hybrid system with automatic page migration between SSD and hard drives -- and that is what the article assumes will be commonly used. Hierarchial storage has a very small niche, and SSD's won't make it more popular.
Hence the article is useless.
The article I read spent a good deal of time talking about flash memory. What article are YOU referring to?
The article treats flash as something you place in between hard drives and memory. This turned out not to happen (with a few exceptions). SSD's simply replace hard drives. Hybrid systems are rare, and it doesn't look like they will become more common -- either you can live with the slowness of hard drives, or you can't. The mainstream will switch to SSD's for everything except backup applications.
There are some hybrid SAN's, but they're damn expensive. At that price they have a hard time competing with simpler pure-flash SAN's.
Yes, lots of blogger speculation. Great.
In the investigation of a recent Airbus airliner crash, it was decided [...]
and then
It wasn't from the investigation, that's how the airbus is built
So it WASN'T decided in the investigation that the onboard computers stopped the pilot from controlling the plane. You can guess as much as you want, but if it was so simple, the investigators would probably have figured it out.
Anyway, what the telemetry shows is that the onboard computers detected that their instruments provided inconsistent input, and therefore they switched to the lowest so-called "law". At the lowest "law", the pilots have almost complete freedom to do stupid things to the plane.
We will likely never know precisely what went wrong in this case, since the black boxes weren't found.
In the investigation of a recent Airbus airliner crash, it was decided that the on-board computer systems did not allow the pilots to override the controls and the plane crashed.
Citation needed. As far as I know, this is not true.
TCP only guarantees that it will tell you if it fails.
It isn't much of a guarantee either. If your physical layer sucks, a lousy 16-bit checksum won't save you.
Good luck with fitting your jobs in the tiny amount of memory and flash. If your chips are 1/10th as fast, you need to run 10 copies of your jobs for the same performance, and since you can't do SMP/NUMA, that means 10 times the memory.
And then you need the 64-port ethernet switch to link it all.
They would need to make their own chipset to link the CPU's. Probably beyond Facebook's resources.
You might have to implement better synchronisation primitives for the ARM architecture as well. It has taken x86 a long time to reach the point where 64-core servers actually deliver decent performance, and it will likely take a while for ARM too.
Mainframes aren't known for good CPU performance. Brilliant at I/O, but if Facebook is I/O-limited they're either doing something very wrong or something very right.
You can't get massively-SMP (or NUMA) ARM machines. That means you're stuck with using lots and lots of machines, which again means you'll use too much space and spend too much money on switches etc.
No, he just found that RAID controllers suck. Which they do, universally, all the time. The only ones that actually perform decently are the ones in external SAN boxes, and inside they are typically servers with software RAID...
beaming back terrabytes of data every second.
Aresbytes, not Terrabytes.
Following this logic, ancient computer makers should have been sued years ago for bundling their OS on their mainframes. There could have been an independent market for OSes.
Err, that's exactly what happened.
"Then in January 1969 the US Justice Department brought an antitrust action against IBM for monopolizing the computer market. At the time IBM sold its hardware, software, training, and all services as a bundled product. That is, if someone wanted the mainframe software they also had to purchase hardware, training, and everything else from IBM. So in the July 1969 IBM signed another consent decree to unbundle which led to the development of hundreds of companies for supplying software (like University Computing and Computer Associates), hardware (disk drives, memory, and the like)."
cited from Peter Vogel's blog.
Their FAQ claims the car is a great lasting investment due to lack of complexity and moving parts, but having to drop $25k every 6 years for a new battery would be a deal breaker.
There should be good chances that the old batteries will be worth a decent amount of money, because the raw materials are so expensive. There are also batteries on the way using much cheaper materials, so it is possible that the second battery will only be a fraction of the price (but in that case you probably won't get much for your now-obsolete battery).
I believe this requirement is only present in GPL v3.
It's only spelled out in GPL v3. Some people believe that other language in the GPL v2 imply the same thing.
GPL absolutely relies on the copyright laws. If there were no copyright laws there isn't any way the GPL could exist, legally speaking.
A lot of people who like the GPL would be perfectly happy to see it go away because copyright went away. Me included.
The "if it's statically linked then the entire blob is now GPLed" gets in the way of perfectly legitimate software designs where dynamic linking is not possible or desirable.
You seem to be under the misconception that dynamic linking gets around the GPL. It doesn't. It can, in certain cases, make a difference for the LGPL.
Running under a different user account with a clean profile and no extensions makes no difference.
I bet there is a setting for what cursor-down is supposed to do, which Fedora sets differently from Debian. Try a diff on about:config.
You can't subnet it. Well you can, if you assign everything statically and you avoid certain implementations of IPv6, but it's a bad idea if you want things to "just work" -- and one of the advantages of IPv6 is the way that things "just work".
So with just a /64 you can't make a separate network for guest Wifi, or a DMZ, or for separate departments. All that is easy to do with IPv4 and NAT, so if ISP's only give out a /64, NAT will quickly get popular in IPv6 too.