On a related matter: saving state doesn't help if your hard disk crashes, or something else goes seriously wrong with your hardware (as opposed to a power outage). How about saving the application state on a different machine, i.e., uploading it to some remote server or whatever? Disaster recovery is then possible by downloading the state to a different machine (of course you'll need a backup of your application as well) and resuming the calculation.
Of course, you could write your application so that it saves state at regular intervals (aka checkpointing). Especially with calculations you should be able to store intermediate results.
I'm not disagreeing with anything you say, however, a lot of companies do not pay dividends at all.
I completely agree that nobody should gamble with their retirement fund (or college fund, or any money they can't afford to lose), since even a "solid company" can be a bad investment (Enron).
Btw, "speculator"? I don't even have money in stocks. I was merely pointing out that a lot of stock prices are based on nothing substantial, and therefore stocks can be a risky investment.
Just because Amazon turns a slight profit doesn't make your stocks worth anything
Right! Perception is reality in the stock market. Amazon's stock is worth what people think it is worth, mundane things like profits etc. only play a role because they may change the perception of the value, hence the value. It's the principle that Wall Street is based on.
Now, this is where it gets fun. (Yes, I've been reading up a bit on this.:) The processor uses a clock multiplier, to translate the clock rate of the system into the clock rate for the processor. You -can- fiddle with this multiplier, but it occured to me that you don't actually need to. You can simply replace the system's clock, instead.
I don't know what the "default" multiplier is, but I'd guess it's probably x16, or thereabouts, given the speed of the rest of the system. In which case, if you threw in a 1GHz clock (I suggest oven-baked, as they taste better), you'd get 16 GHz on the processor.
Now, THAT kind of ramp-up is not going to be easy. Chips run hot as it its, and if you plan on overclocking an Athlon by a factor of 8, you can expect to fry eggs on the surface. As for the poor RAM chips.... those are REALLY going to suffer!
If things were that simple, people would have done it already.
I'm no expert on overclocking, but there are a lot of problems with this approach.
- You can't just apply a 1GHz clock to a system. There's the matter of peripheral chips that simply will not work. As you mention, the RAM is going to have a huge problem. Then, there will be all sorts of problems with the motherboard itself (coupling between PCB traces, electromagnetic interference). There is a reason why overclockers don't make motherboards themselves: it's just too hard to get it right, never mind the difficulties in making multi-layer PCBs.
- Heat production by chips is a limiting factor, so cooling will help. But: when you try to overclock an Athlon 8 times, you're going to run into different problems. There's the matter of electron mobility inside the chip. No matter how cold it is, there is a limit to the speed they move, given certain dimensions of the transistors. Same goes for the RAM: you can cool it all you want, at a certain point it just won't go any faster any more (and that point is reached long before your 8X speed increase).
Well, correct me if I'm wrong, but there are two types of players that could solve the region problem (I'm not talking about DivX, or ripping or whatever, I don't know anything about that).
First, you can get a "codeless" DVD player. These things ignore the region coding. They're a little expensive , though. But (!), the movie industry doesn't like these things, so they started releasing DVD's with a slightly modified region 1 code. Some of these DVD's will not play on a codeless player!
I'm not endorsing any vendors, but doing a search on Google with "codeless DVD" gives several links to places where you can buy these players. Usually they sell solutions for the PAL/NTSC (often built into the codeless players) and 110V/220V issue too.
Second, you can get a "multi-region" player. It just recognizes the region code, and adjusts itself as needed. A friend of mine has one of these. I think they can play any region DVD. Alas, they are also substantially more expensive than regular players.
A Google search for "multi-region DVD" also results in several links, however, the companies selling these seem to be mainly in the UK or Australia.
Hah! Finally the real reason behind region coding becomes clear! All the time we thought it was a conspiracy by the movie industry, but now it's clear that the US government is behind this! Can't have American minds poisoned with this Euro-commie garbage ("Hitchhiking" is bad, and "Red Dwarf" must be a Russian guy)! The people who understand anything about region coding are beyond salvation, they have been exposed to too much information already.
Does it occur to anyone that the lawyers seem to be running companies (and implicitly, the country) nowadays? I mean, did George Lucas look at this and decide that this evil stuff must be stopped? Seems more likely that some peon in Lucasfilms alerted the legal department, who then decided that they should create the appearance that they are actually doing something for their money.
Everyone seems to have this attitude that 'you don't need this' or that is fast 'enough'... but why?
Because there is a cost issue. You could probably get a telco-grade line at 100Mb/s today. But would you be willing to pay thousands of dollars a month for it? Granted, some companies or institutions may need this, and are willing to fork over the money for it. I'm just looking at it from a personal point of view.
If someone would offer 100 Mb/s service for $40 dollars a month, I'd order it today...
I once ran an X-client over 28.8K dial-up. That's not workable, I agree:-)
I agree that it would be cool and convenient and nice to have high bandwidth. I have a cable connection, and it sure beats dial-up. It's worth the $40 a month I'm paying.
But: is it absolutely needed? Not for me. I could read slashdot with my dial-up, and read/write e-mail, and browse the Web. Sure, I would think twice about downloading an ISO image, so in that sense things have improved.
I don't doubt that things will be moving in the high-bandwidth direction, I just still don't understand how it would be a national priority. Surely this plan must cost money (if it were free, I'd say "go for it today").
Maybe offering affordable broadband (~1.5 Mb/s) for more people could be a higher priority. A bandwidth like that allows you to do pretty much anything to improve productivity, exchange information, etc.
Then again, maybe in 5 years I'll be going "what was I thinking???".:-)
Excuse my stupidity, but how would a 100Mb/s connection for everyone help the economy? What in heavens name do you need that for? Watching movies? How does that help the economy (other than the MPAA, RIAA and the rest of them, since we're not going to get all that content for free)? Watching movies is nice, but I already have something for that (TV, VCR, DVD).
Don't tell me you need 100Mb/s for browsing. Downloading big files, sure, but how often do you do that? And why isn't 1.5 Mb/s fast enough for that? Again, watching movies in real-time over an ultrafast connection would be cool, but why is that a national priority?
1: The American spy plane was in international waters (as recognized by the rest of the world save China)
No, it was a plane, not a boat. But, yes, it was in international airspace, until it had to make an emergency landing on a Chinese airfield. There's nothing wrong with that in my opinion (an emergency is an emergency), but formally it violated Chinese airspace.
2: The pilot of the Chinese plan was killed because he flew too close to a larger plane and ran into it
It seems likely, but we don't know that for a fact. The Chinese would argue (have argued?) that the American plane flew too close to the Chinese plane, causing it to crash. Most likely, both pilots where playing "chicken", and that's a dangerous game with airplanes (duh).
Lost what, the Cold War? Maybe they (Gorbachev) saved themselves and the rest of us from World War III? We'll never know. The US "victory" in the Cold War cost this country dearly: defense spending that took the economy 20 years to recover from.
Don't confuse Karl Marx with communism. They don't have the word "marxism" for nothing.
Marxism evolved into communism on the one hand, and socialism on the other hand. Communism has not been a great success, partly because it was first implemented in Russia (which had a long history of oppression), and partly because it denies a few basic human impulses, such as greed and the need for religion (I know, those are part of Marxism as well).
Socialism on the other hand has been rather successful in quite a number of cases, especially when it is merged with democracy ("social democracy").
In any case I don't imagine that a "killswitch" would be a big dashboard LED, but rather it'd be subtle -> Gosh darnit the amraam just refuses to get a lock for some reason! Why is the radar cluttered and incapable of separating targets?
I don't think it's as simple as you think. First of all, when the US delivers (say) jet fighters to an allied country, they deliver all the technical information as well (note that if this stuff is too advanced, the plane does not get exported, period. See for example the F-22). The technical information is needed so that maintenance personnel can be trained. There may be exceptions, but usually maintenance and repair are done by local mechanics and engineers. Despite what some people may think, these guys are usually not morons. There is a substantial risk that a trojan horse device will eventually be found. If that happens, all hell breaks loose, and US exports will suffer, to say the least.
Also, consider that a lot of the equipment like radars, etc. are supplied by sub-contractors. With so many (often hundreds) of companies contributing, the risk of leaks becomes rather big.
Last, consider that an ally turning into an enemy overnight doesn't happen all that often. The only case I can think of is Iran.
As it is doesn't Microsoft do all development in Redmond, and satellite offices are merely sales and consulting?),...
No, I don't think so. I think the scandal about anti-communist messages in Chinese Windows versions was caused by the fact that the software was developed in Taiwan. Makes sense, easier to find Chinese speaking programmers there...
Furthermore, I know that Microsoft has (or had, haven't checked) a big operation in Ireland. They did all the internalization (translating strings, currency formats, date formats, etc.) for the European market there.
I'm wondering: were this bugs marked "made in USA"?
Seems silly, but (slightly offtopic): years ago, the US tapped a Soviet military communications cable that was running underwater in some bay somewhere near the Asian Soviet coast. Worked well for a while, but when the Soviets finally discovered the tapping device/recorder, it turned out it was marked "Property of the United States government". Somewhat amusing.
You can read all about it in "Blind Man's Bluff" by Sherry Sontag et al.
Yeah, duh. Look around your house and see how much stuff is actually marked "Made in China". You will be surprised how many items you can find. Now, suppose they all had explosives in them, or were designed to short circuit and set your house on fire, or whatever. Would you go "duh, what do you expect when you buy stuff from an enemy nation?".
You can call China an enemy or not, but fact is that trade relations are normal and there's a whole lot of trade going on between the US and China. The Chinese have the right to expect that (in this commercial transaction) they will get what they paid for, without silly surprises.
If I were the Chinese, I'd ask for my money back...
Maybe. Or maybe they want to protect the sales of their CD-drives/players/copiers? Or maybe they're ticked off because somebody is messing with their invention (i.e. compact discs) without asking them first? Or maybe they see all this as good PR for them (lots of posters indicated they are going to buy Philips next time around)? Or maybe they are the "good guys" in the Corporate World? Or have we all gotten too cynical to believe in such a fairytale?
There's a lot of talk going on here about how stupid the terrorists were to use only 40-bit encryption.
Consider that it took several weeks of armed combat to compromise the physical security of these machines. What I mean is that these machines were in Kabul. As I understand, the 40-bit encryption applied to the disk, not to any data sent out over networks. So, their first line of defense was the fact that the machines were in Afghanistan. Given that, the 40-bit encryption almost seems paranoid, not sloppy.
I have no factual information about it, but I would guess that a lot of institutions and companies that have physical security (i.e., you can't just walk into the building), have a somewhat relaxed policy when it comes to the size of the encryption key used for their hard disks (PGP-encrypted e-mail, sure, but 128-bit encryption for every local disk? Me thinks not).
Matter and energy are governed by the same conservation law and are different forms of the same thing. E=mc^2, right?
So when you "annihilate" matter using antimatter, you're not removing it from the universe, you're turning into pure energy, which is just another of its forms. The system is closed; nothing is lost.
Correct.
As for spontaneous particle-anti-particle pairs, that's more to do with zero-point fluctuation and things like Hawking radiation.
I was not referring to spontaneous pairs (I think you're right about that), but rather to controlled lab experiments, where the idea is to create matter out of energy. The point was that it's not a one-way process, you can also transform energy into matter.
No, I wasn't referring to the Church-Turing hypothesis. You'll have to forgive me for not having a reference immediately, but I distinctly recall a mathematical proof that you cannot prove program correctness. It's a fundamental problem, it doesn't have anything to do with the size of the processor or the number of registers. I'll have to look it up, I'm afraid. If I can find a reference, I'll post it.
Slashdot Mirror
Can anyone name applications that would require larger than 64-bit numbers?
Public-key cryptography.
Mind you, there is more to it than just the range of integers, there's also the matter of the amount of data that can be processed at the same time.
Yes, obviously you need the source.
On a related matter: saving state doesn't help if your hard disk crashes, or something else goes seriously wrong with your hardware (as opposed to a power outage). How about saving the application state on a different machine, i.e., uploading it to some remote server or whatever? Disaster recovery is then possible by downloading the state to a different machine (of course you'll need a backup of your application as well) and resuming the calculation.
Of course, you could write your application so that it saves state at regular intervals (aka checkpointing). Especially with calculations you should be able to store intermediate results.
I'm not disagreeing with anything you say, however, a lot of companies do not pay dividends at all.
I completely agree that nobody should gamble with their retirement fund (or college fund, or any money they can't afford to lose), since even a "solid company" can be a bad investment (Enron).
Btw, "speculator"? I don't even have money in stocks. I was merely pointing out that a lot of stock prices are based on nothing substantial, and therefore stocks can be a risky investment.
Just because Amazon turns a slight profit doesn't make your stocks worth anything
Right! Perception is reality in the stock market. Amazon's stock is worth what people think it is worth, mundane things like profits etc. only play a role because they may change the perception of the value, hence the value. It's the principle that Wall Street is based on.
Now, this is where it gets fun. (Yes, I've been reading up a bit on this. :) The processor uses a clock multiplier, to translate the clock rate of the system into the clock rate for the processor. You -can- fiddle with this multiplier, but it occured to me that you don't actually need to. You can simply replace the system's clock, instead.
I don't know what the "default" multiplier is, but I'd guess it's probably x16, or thereabouts, given the speed of the rest of the system. In which case, if you threw in a 1GHz clock (I suggest oven-baked, as they taste better), you'd get 16 GHz on the processor.
Now, THAT kind of ramp-up is not going to be easy. Chips run hot as it its, and if you plan on overclocking an Athlon by a factor of 8, you can expect to fry eggs on the surface. As for the poor RAM chips.... those are REALLY going to suffer!
If things were that simple, people would have done it already.
I'm no expert on overclocking, but there are a lot of problems with this approach.
- You can't just apply a 1GHz clock to a system. There's the matter of peripheral chips that simply will not work. As you mention, the RAM is going to have a huge problem. Then, there will be all sorts of problems with the motherboard itself (coupling between PCB traces, electromagnetic interference). There is a reason why overclockers don't make motherboards themselves: it's just too hard to get it right, never mind the difficulties in making multi-layer PCBs.
- Heat production by chips is a limiting factor, so cooling will help. But: when you try to overclock an Athlon 8 times, you're going to run into different problems. There's the matter of electron mobility inside the chip. No matter how cold it is, there is a limit to the speed they move, given certain dimensions of the transistors. Same goes for the RAM: you can cool it all you want, at a certain point it just won't go any faster any more (and that point is reached long before your 8X speed increase).
Well, correct me if I'm wrong, but there are two types of players that could solve the region problem (I'm not talking about DivX, or ripping or whatever, I don't know anything about that).
First, you can get a "codeless" DVD player. These things ignore the region coding. They're a little expensive , though. But (!), the movie industry doesn't like these things, so they started releasing DVD's with a slightly modified region 1 code. Some of these DVD's will not play on a codeless player!
I'm not endorsing any vendors, but doing a search on Google with "codeless DVD" gives several links to places where you can buy these players. Usually they sell solutions for the PAL/NTSC (often built into the codeless players) and 110V/220V issue too.
Second, you can get a "multi-region" player. It just recognizes the region code, and adjusts itself as needed. A friend of mine has one of these. I think they can play any region DVD. Alas, they are also substantially more expensive than regular players.
A Google search for "multi-region DVD" also results in several links, however, the companies selling these seem to be mainly in the UK or Australia.
Hah! Finally the real reason behind region coding becomes clear! All the time we thought it was a conspiracy by the movie industry, but now it's clear that the US government is behind this! Can't have American minds poisoned with this Euro-commie garbage ("Hitchhiking" is bad, and "Red Dwarf" must be a Russian guy)! The people who understand anything about region coding are beyond salvation, they have been exposed to too much information already.
Does it occur to anyone that the lawyers seem to be running companies (and implicitly, the country) nowadays? I mean, did George Lucas look at this and decide that this evil stuff must be stopped? Seems more likely that some peon in Lucasfilms alerted the legal department, who then decided that they should create the appearance that they are actually doing something for their money.
Everyone seems to have this attitude that 'you don't need this' or that is fast 'enough'... but why?
Because there is a cost issue. You could probably get a telco-grade line at 100Mb/s today. But would you be willing to pay thousands of dollars a month for it? Granted, some companies or institutions may need this, and are willing to fork over the money for it. I'm just looking at it from a personal point of view.
If someone would offer 100 Mb/s service for $40 dollars a month, I'd order it today...
I once ran an X-client over 28.8K dial-up. That's not workable, I agree :-)
:-)
I agree that it would be cool and convenient and nice to have high bandwidth. I have a cable connection, and it sure beats dial-up. It's worth the $40 a month I'm paying.
But: is it absolutely needed? Not for me. I could read slashdot with my dial-up, and read/write e-mail, and browse the Web. Sure, I would think twice about downloading an ISO image, so in that sense things have improved.
I don't doubt that things will be moving in the high-bandwidth direction, I just still don't understand how it would be a national priority. Surely this plan must cost money (if it were free, I'd say "go for it today").
Maybe offering affordable broadband (~1.5 Mb/s) for more people could be a higher priority. A bandwidth like that allows you to do pretty much anything to improve productivity, exchange information, etc.
Then again, maybe in 5 years I'll be going "what was I thinking???".
Excuse my stupidity, but how would a 100Mb/s connection for everyone help the economy? What in heavens name do you need that for? Watching movies? How does that help the economy (other than the MPAA, RIAA and the rest of them, since we're not going to get all that content for free)? Watching movies is nice, but I already have something for that (TV, VCR, DVD).
Don't tell me you need 100Mb/s for browsing. Downloading big files, sure, but how often do you do that? And why isn't 1.5 Mb/s fast enough for that? Again, watching movies in real-time over an ultrafast connection would be cool, but why is that a national priority?
1: The American spy plane was in international waters (as recognized by the rest of the world save China)
No, it was a plane, not a boat. But, yes, it was in international airspace, until it had to make an emergency landing on a Chinese airfield. There's nothing wrong with that in my opinion (an emergency is an emergency), but formally it violated Chinese airspace.
2: The pilot of the Chinese plan was killed because he flew too close to a larger plane and ran into it
It seems likely, but we don't know that for a fact. The Chinese would argue (have argued?) that the American plane flew too close to the Chinese plane, causing it to crash. Most likely, both pilots where playing "chicken", and that's a dangerous game with airplanes (duh).
Don't forget, the USSR lost...
Lost what, the Cold War? Maybe they (Gorbachev) saved themselves and the rest of us from World War III? We'll never know. The US "victory" in the Cold War cost this country dearly: defense spending that took the economy 20 years to recover from.
Don't confuse Karl Marx with communism. They don't have the word "marxism" for nothing.
Marxism evolved into communism on the one hand, and socialism on the other hand. Communism has not been a great success, partly because it was first implemented in Russia (which had a long history of oppression), and partly because it denies a few basic human impulses, such as greed and the need for religion (I know, those are part of Marxism as well).
Socialism on the other hand has been rather successful in quite a number of cases, especially when it is merged with democracy ("social democracy").
Which country gives communism a good name then?
Cuba! Great cigars and nice weather! And everybody speaks Spanish!
If you're a big nuclear superpower, just make an international incident out of it!
That's ridiculous. $30 million is not worth an international incident. How stupid do you think the Chinese are?
In any case I don't imagine that a "killswitch" would be a big dashboard LED, but rather it'd be subtle -> Gosh darnit the amraam just refuses to get a lock for some reason! Why is the radar cluttered and incapable of separating targets?
I don't think it's as simple as you think. First of all, when the US delivers (say) jet fighters to an allied country, they deliver all the technical information as well (note that if this stuff is too advanced, the plane does not get exported, period. See for example the F-22). The technical information is needed so that maintenance personnel can be trained. There may be exceptions, but usually maintenance and repair are done by local mechanics and engineers. Despite what some people may think, these guys are usually not morons. There is a substantial risk that a trojan horse device will eventually be found. If that happens, all hell breaks loose, and US exports will suffer, to say the least.
Also, consider that a lot of the equipment like radars, etc. are supplied by sub-contractors. With so many (often hundreds) of companies contributing, the risk of leaks becomes rather big.
Last, consider that an ally turning into an enemy overnight doesn't happen all that often. The only case I can think of is Iran.
As it is doesn't Microsoft do all development in Redmond, and satellite offices are merely sales and consulting?),...
No, I don't think so. I think the scandal about anti-communist messages in Chinese Windows versions was caused by the fact that the software was developed in Taiwan. Makes sense, easier to find Chinese speaking programmers there...
Furthermore, I know that Microsoft has (or had, haven't checked) a big operation in Ireland. They did all the internalization (translating strings, currency formats, date formats, etc.) for the European market there.
I'm wondering: were this bugs marked "made in USA"?
Seems silly, but (slightly offtopic): years ago, the US tapped a Soviet military communications cable that was running underwater in some bay somewhere near the Asian Soviet coast. Worked well for a while, but when the Soviets finally discovered the tapping device/recorder, it turned out it was marked "Property of the United States government". Somewhat amusing.
You can read all about it in "Blind Man's Bluff" by Sherry Sontag et al.
Yeah, duh. Look around your house and see how much stuff is actually marked "Made in China". You will be surprised how many items you can find. Now, suppose they all had explosives in them, or were designed to short circuit and set your house on fire, or whatever. Would you go "duh, what do you expect when you buy stuff from an enemy nation?".
You can call China an enemy or not, but fact is that trade relations are normal and there's a whole lot of trade going on between the US and China. The Chinese have the right to expect that (in this commercial transaction) they will get what they paid for, without silly surprises.
If I were the Chinese, I'd ask for my money back...
Maybe. Or maybe they want to protect the sales of their CD-drives/players/copiers? Or maybe they're ticked off because somebody is messing with their invention (i.e. compact discs) without asking them first? Or maybe they see all this as good PR for them (lots of posters indicated they are going to buy Philips next time around)? Or maybe they are the "good guys" in the Corporate World? Or have we all gotten too cynical to believe in such a fairytale?
There's a lot of talk going on here about how stupid the terrorists were to use only 40-bit encryption.
Consider that it took several weeks of armed combat to compromise the physical security of these machines. What I mean is that these machines were in Kabul. As I understand, the 40-bit encryption applied to the disk, not to any data sent out over networks. So, their first line of defense was the fact that the machines were in Afghanistan. Given that, the 40-bit encryption almost seems paranoid, not sloppy.
I have no factual information about it, but I would guess that a lot of institutions and companies that have physical security (i.e., you can't just walk into the building), have a somewhat relaxed policy when it comes to the size of the encryption key used for their hard disks (PGP-encrypted e-mail, sure, but 128-bit encryption for every local disk? Me thinks not).
Anyway, just my 2 cents.
Matter and energy are governed by the same conservation law and are different forms of the same thing. E=mc^2, right?
So when you "annihilate" matter using antimatter, you're not removing it from the universe, you're turning into pure energy, which is just another of its forms. The system is closed; nothing is lost.
Correct.
As for spontaneous particle-anti-particle pairs, that's more to do with zero-point fluctuation and things like Hawking radiation.
I was not referring to spontaneous pairs (I think you're right about that), but rather to controlled lab experiments, where the idea is to create matter out of energy. The point was that it's not a one-way process, you can also transform energy into matter.
No, I wasn't referring to the Church-Turing hypothesis. You'll have to forgive me for not having a reference immediately, but I distinctly recall a mathematical proof that you cannot prove program correctness. It's a fundamental problem, it doesn't have anything to do with the size of the processor or the number of registers. I'll have to look it up, I'm afraid. If I can find a reference, I'll post it.