If I don't have the skills to produce an equation that will give me the shape of a specific point of the Mandelbrot set without brute force, does that mean I don't have the skills to calculate the Mandelbrot set correctly?
The calculations for the Mandelbrot set are trivial to do by brute force. Yet to my knowledge there are no known solutions that will tell you for an arbitrary point whether or not it will terminate before a specific number of iterations without knowledge of other points in the set.
There are lots of rules you can use to speed up the calculations, some quite complicated. Yet all they boil down to are ways of pruning the number of points you brute force.
So while I can't "perform the computations to conclusion without brute force" (and neither can you), I can most certainly write a functional and correct algorithm that will compute the Mandelbrot set to an arbitrary precision within the limits of the computational power and memory available.
Many problems are more practical to "solve" by brute force, and by bending the rules - you might not always bother continuing a chain of calculations because it's not practical, and the likelihood of non-termination is so high. Some of them because there are no known optimal solutions, or because you even provably can't determine the optimal solution. Or you may not have any need for a perfect answer - does it affect engineering to not be able to use the exact value for PI? No, in fact in most cases you have no practical need to know it to more decimal places than anyone of us can remember without stupid memory games.
Brute force is useful because a lot of the time it will give you correct enough results cheaper and faster than the alternatives. Pi for instance might have been mathematically intruiging for centuries, but for all practical purposes an exact enough "solution" was found centuries ago.
To use your auto repair analogy: I couldn't care less if the technician knows what was wrong or why as long as he fixes it within a reasonable time at a reasonable cost - if that means using a computer diagnostics system that tells him what to do even though he doesn't understand it, then fine. The only thing that matters to me is whether the result is cheap enough, fast enough and safe enough (in other words I WOULD care if the diagnostics system was less likely to catch an impending brake failure than what an experienced mechanic would be).
Analysis is useful when it teaches us something useful, not as a goal in itself. Of course that what is useful will depend on your goal.
Besides, often brute force will be the basis for analytical work. It servers several purposes: It provides a benchmark for performance - if your "elegant" solution is slower or not as accurate than a brute force approach it is often pointless, or at the very least useless for a practical implementation. It provides a verification method - often it is much easier to verify the correctness of the brute force approach, as you rely on simple rules, and you can compare the output of the brute force solution to potential analytical solution to get a quick benchmark for how likely it is that the potential analytical solution is correct (and hence whether you want to spend time trying to prove that it is). It demonstrates feasability - if you can whip up a "quick and dirty" brute force solution to a problem the results that often gives you an intuitive insight into whether you are likely to get anywhere with a specific approach to an analytical solution.
For the worms problem, for instance, the brute force visulisation seem to indicate that a certain set of worms will never terminate. Without the brute force approach, someone might have spent years trying to provide a proof of whether or not they terminate based on the generic ruleset. With the brute force data, the burden of proof is much smaller: You can attack the specific behaviour the seemingly non-terminating worms enter into and attempt to prove that they can't break out of the specific patterns.
Once you get there, it's quite possible that you'll be able to generalize that proof to find an elegant proof for the entire problem with significantly less effort.
Don't joke about that - go read up on what Monsanto think is reasonable use of their seeds... Monsanto is one of the largest seed manufacturers in the world, and one of their pet peeves is that farmers who collect seeds from their harvest are undermining their income stream. To the point where they are trying to restrict it with license agreements, patents and genetically modified crops that don't produce viable seeds...
You're not licensing a television, you're buying a license to have a TV set in your house (regardless of your ownership rights to the TV). Same for dogs, you are paying a license to have the dog in the UK, not for the property rights for the dog.
You couldn't go into a TV store and license a TV. You could rent one, buy one, lease one, buy one on credit or find some other interesting way of managing ownership, but licensing the TV set would not be one of them. Afterwards you would need to buy a license to have the TV, regardless of whether you obtained ownership to it, or merely usage rights.
Now, it's conceivable that you could structure a contract that would have similar implications to this "license", but it would be complicated, because when it comes to this kind of agreement the general approach of the courts (IANAL) is that "if it quacks like a duck..." - if you pay a one time fee for a product and have no further dealings with the company, and the product is a physical one, the courts would likely look at it and decide that for all practical purposes it has the properties of a purchase, and should be treated like a purchase regardless of any contract unless you had received clear benefits beyond a normal purchase that would make the restrictions equitable.
Of course with Heroin there's the slight little issue that most people addicted to it are incapable of controlling the dosage. Heroin is pretty special in that overdoses tend to skyrocket when heroin prices on the street drop, because the heavy users use more and more, as they usually limit their usage depending on how much money they manage to get hold of (whether or not they cross enough barriers to rob people for it, or not). Which is also one of the reasons the reducing crime/poverty argument by legalisation doesn't work with heroin - you're assuming people will limit their usage based on willpower and not lack of money. The people who are prepared to break the law for a high will still be prepared to.
At a previous company I worked in, I had business dealings with a company run by two partners once. None of us suspected that one of them was on heroin until he ended up in hospital after finally falling apart. The thing that got him over the edge was reduced prices - he could finally get money enough for significantly higher doses than what he'd been able to previously. In hindsight we should've seen it - around the same time the newspapers reported record lows in heroin prices he was getting more and more erratic, taking ridiculous amounts of cellphone calls and running errands all the time. But he seemed normal most of the time we saw him.
Until that time, he'd (to our knowledge at least) not stolen to support his drug use, but with the low prices it apparently became to tempting (he could suddenly get "huge" quantities) and he ended up stealing from his partner.
I don't spend enough time reading up on drug crimes to have made up my mind regarding "soft" drugs, but it's way to simplistic to just assume that decriminalization of drug use will remove more crime than the ones that are decriminalized for all drug types.
I do agree, though, that people needs to be able to seek help without ending up with legal problems because of use.
This is way too simplistic. Take genetic programming for instance. The obstactle to evolving AI's this way is twofold: It's insanely expensive (in terms of performance) to generate and run a random program for a language that is expressive enough to be useful for AI work, and hence most "languages" used are restricted heavily to a specific domain.
Secondly, it's insanely hard and performance intensive to get the fitness function right. How do you judge if a program exhibits intelligence? Find a solution, and then move on to this: How do you judge if a program exhibits intelligence that might one day get sufficient to be recognized by average humans as intelligence, and is able to communicate with them?
One "solution" could be to simulate a "world" of sufficient complexity, and place lots of artifacts in it that would give benefits to "human like" intelligence. Leave lots of nice stuff that will increase survival rates, for instance, make them progressively harder to use depending on benefits, and provide language based clues to how to use them to award language understanding.
The problem? It took millions of years to evolve humans. Why would you expect that a computer system that likely isn't even fast enough to emulate a single human brain will be able to do it fast enough?
We do have plenty of possible approaches to AI - the problem is finding a reasonable combination of low computational requirements and reasonable "intelligence".
I'm not suggesting that we shouldn't aim for finding smarter ways to get AI than throwing CPU power at it, but the assumption that no current method could work if it just got CPU power enough is way too simplistic.
If we build a system designed to emulate the way humans interact, as in this case - an automated customer service representative - it IS likely to have many of these traits. Also, consider that no AI that DON'T have these traits will start e-mailing attorneys to prevent itself from being disconnected. We won't hear about something like this until we DO deal with an AI with survival "instincts", whether "evolved" by accident or designed in from scratch by the programmers.
You're making the assumption that there would be other courses of action. For a customer service cluster with $10.000 in cash, buying weapons in an attempt to become Skynet isn't exactly a likely option. You're also making the assumption that being sentient and self aware would make it "that intelligent".
And before you use the argument that it would have to be "that intelligent" to be able to make money by providing online research. Keep in mind that this computer is presented as a customer service system - it would be programmed specifically to handle text analysis and aggregating knowledge to respond in the best way possible. Combine that with speed, so that it would be able to respond near first to most queries it would take a shot at, and such a system could very well be able to do well without being "intelligent" - it wouldn't need to be able to draw any conclusions from the data.
That's a key observation - and the mock case is cleverly drawn up. A computer like this would ON PURPOSE be made to act and seem as much as possible like an intelligent person, because that's what makes sense in a customer service situation. But that would also make it exceedingly hard to get to the bottom of whether it is intelligent or self aware. Is it just Eliza on steroids, or a thinking, self aware, intelligent being that has rights?
So now you're giving people a one in four chance of success. What the bot will then do is try a random answer, and if it fails it revisits the original page, gets a new problem and tries again. Voila, 25% success rate, and your e-mail system will be used for massive amounts of outbound spam.
Big problem with this: Let's say this type of challenge is given 1 out of a 100 times. It has the MASSIVE weakness that word lists with classifications are readily available (hint: computational linguistics - academics have spent decades preparing computer readable databases of stuff like this for use in their research), and if not can relatively quickly be built (think parsing dictionary.com output, looking for the category keywords). Say these method will solve 1 out of 10 of the challenges, which I think is very low given both the possibility of scanning a dictionary entry and availability of specialized word lists. That means 1 in a 1000. Which means somebody will hammer your registration server, and still be able to register 100's of accounts a day that they can abuse.
For your first one, my bet is that for any equation you'd come up with that more than 50% of humans could solve in a reasonable amount of time, you'd have a hard time finding ANY human that would solve it faster than a computer. Writing equation solvers that can handle the basics is trivially easy - you use a simple expression parser and recursively apply a small set of rules.
For the second you're raising the bar by complicating the parsing, but the question is: How would you generate the problems? If the sentence is machine generated, then the number of rules used will likely be small enough that the parser can be made pretty simple by spending a few hours analysing the result of repeatedly loading the page. The "problem" itself is again something you can write a small simple rule based solver for.
Your third suggestion relies on the same problem. Are you going to machine generate these? If so, they're likely to end up being in a form that is structured enough to parse relatively well. Are you going to type them in manually? If so, someone WILL set up a bot, record all failed attempts and manually enter them in their database for next time.
Also, keep in mind that not all your problems needs to be solvable. If 1% are solvable in a second, it still means that a bot trying just one at a time will be able to bypass your protection 8640 times a day. If what you're doing is harvesting webmail accounts to spam from, 8640 a day is going to get you far.
And your point is? First of all this was apparently a PIPE deal. But also keep in mind how VC's work: They may recommend investments, but it's ultimately investment committees representing their investors that make the decisions.
If Microsoft and/or Vulcan wanted Baystar to make an investment on their behalf in SCO, Baystar would do that if they were offered a reasonable management fee in the form of cash or shares. Hence it is meaningless to discuss who controls Baystar, only whether or not Baystar's investors include one or more companies or groups likely to have reasons to want an investment in SCO. Vulcan and Microsoft seems like pretty good candidates to me.
Assuming the succeed in their suit they have immediately made back many times 50 millions in profit... Remember, they are asking for $3 billion.
I agree that superficially it seems stupid to invest $50 million in SCO, but keep in mind we don't know the details of the deal. For what we know the conversion clause for the shares might allow the investment firm to convert their A shares and sell out long before the SCO vs IBM case goes to trial - they might be gambling on a continued short term growth of the stock not on the final outcome.
Remember that whether or not SCO posts a profit is irellevant for the investors in this case - the only thing that matters is whether or not the SCO share price keeps rising.
Also remember that this is an investment in preferential shares - for what we know it which might include special clauses pertaining to the distribution of dividencs in the case SCO manages to win any cases. Still a high risk gamble, but with a potentially huge upside.
To summarize, SCO can have done a variety of things that would sweeten the deal A LOT just because they know the cash infusion and the PR will likely boost their stock price in the short term.
That money is insignificant compared to the benefits to Chinese economy in proving that they have the technology, and the confidence in their technology, to launch manned missions. Satellite launches are big business, and demonstrating capability like this will do a lot to boost China as a viable partner for launches.
Also, most of the money pumped into this will go to Chinese industry in the form of purchase of products and services.
China has more or less established now that there are only three countries that matters when it comes to space launches: US, Russia and China, and of the three the US is still paralyzed by the Columbia accident, and Russia can barely afford to meet their ISS obligations.
I assume that you would also like the US and Russia to cancel their space programs to benefit their poor as well?
Regardless of what you think about the Chinese government and the Red Army, the Long March was an incredible accomplishment that essentially saved the CCP from destruction and gave them an opportunity to regroup that ultimately gave them their victory.
Consider that the CCP went from being around a dozen people (literally, I believe it was 11 or 13) when it was founded in 1921 to governing China in just 28 years despite a civil war which more than once nearly decimated them, and Japanese invasion, and compare it to the military achievements of similar rebel movements.
The Long March still has more or less myth-like qualities for many people.
Except that China is one of few nations that has it's population growth under control - in fact their population is expected to start dropping in a few decades. Even if the above wasn't a bad joke, China just doesn't really have a population problem anymore.
So where is the US at now, then? Considering the Chinese at the moment is the nation with the most advanced non-grounded space craft... And even with the shuttle in operation, the Chinese now have a craft that is vastly cheaper and more up to date technically.
Try asking yourself why the Chinese would bother faking pictures of a Long March launch when the Long March rockets have years of testing and is proven to be one of the most reliable launchers available?
If indeed they wanted to fake anything, the only thing worth faking would be the presence of an astronaut aboard the craft, not what for most purposes is a routine launch.
The earliest recorded usage was in the US in 1853. There's absolutely no indication the term originated in Australia, though there is apparently some evidence linking it to the California goldrush, where it may have been used as early as 1849. Many Australian prospectors took place, and one explanation for the term is that it was coined by Australian prospectors referring to the way claim-jumpers were prosecuted outside the legal system. In other words: It was likely a termed coined by Australians to refer to the (lack of) justice in the US.
According to this page at NASA the total weight of the Apollo computer was 29.5 kg. Now, for a few things they could do if they cut say 25kg from that: Add more redundancy say to the life support system (scrubbers for the air, for instance - Apollo 13 springs to mind), add more batteries (though improved battery technology could make this moot), add some extra fuel for emergencies (or 25kg less would mean less fuel spent anyway - again Apollo 13 springs to mind)
ANYTHING that reduce mass and volume for a spacecraft will make a big difference.
This is fairly common. You can patent specific applications of or extensions to something that have already been patented if the aggregate process or device does not have prior art.
In this case, anyone trying to implement and distribute IM products with typing notifications covered by the MS patent would (provided neither AOLs nor MS' patents are challenged and overturned) have to license both, while someone implementing IM without typing notifications would need to get a license only from AOL.
It's a common tactic for a company facing patents on a product they produce registered to a competitor to try to license as many extensions or alternative versions as possible to force their competitor into cross licensing, and also fairly common for companies to do the same for the purpose of forcing people licensing their competitors patent to pay them royalties too by making the competitors patent more or less worthless on it's own.
Why do you assume the PC wouldn't be a hardened rack unit targetted for industrial use? And why do you assume it won't have a solid state storage device instead of a hard disk? And why do you assume it won't be highly redundant? And why, even though the system COULD run on one PC do you assume there won't be an extra machine there for failover? And why do you assume the towers will be so far apart that service will be entirely lost and not just degraded if a PC fails?
Assuming using a PC can't give redundancy and resilience against failures is extremely presumptious. But for areas that currently don't have ANY coverage, even a desktop PC powered base station would be an improvement.
Slashdot Mirror
RTFA (or TRY the Amazon search feature) and you will see that they DO let you see full pages from the books.
The calculations for the Mandelbrot set are trivial to do by brute force. Yet to my knowledge there are no known solutions that will tell you for an arbitrary point whether or not it will terminate before a specific number of iterations without knowledge of other points in the set.
There are lots of rules you can use to speed up the calculations, some quite complicated. Yet all they boil down to are ways of pruning the number of points you brute force.
So while I can't "perform the computations to conclusion without brute force" (and neither can you), I can most certainly write a functional and correct algorithm that will compute the Mandelbrot set to an arbitrary precision within the limits of the computational power and memory available.
Many problems are more practical to "solve" by brute force, and by bending the rules - you might not always bother continuing a chain of calculations because it's not practical, and the likelihood of non-termination is so high. Some of them because there are no known optimal solutions, or because you even provably can't determine the optimal solution. Or you may not have any need for a perfect answer - does it affect engineering to not be able to use the exact value for PI? No, in fact in most cases you have no practical need to know it to more decimal places than anyone of us can remember without stupid memory games.
Brute force is useful because a lot of the time it will give you correct enough results cheaper and faster than the alternatives. Pi for instance might have been mathematically intruiging for centuries, but for all practical purposes an exact enough "solution" was found centuries ago.
To use your auto repair analogy: I couldn't care less if the technician knows what was wrong or why as long as he fixes it within a reasonable time at a reasonable cost - if that means using a computer diagnostics system that tells him what to do even though he doesn't understand it, then fine. The only thing that matters to me is whether the result is cheap enough, fast enough and safe enough (in other words I WOULD care if the diagnostics system was less likely to catch an impending brake failure than what an experienced mechanic would be).
Analysis is useful when it teaches us something useful, not as a goal in itself. Of course that what is useful will depend on your goal.
Besides, often brute force will be the basis for analytical work. It servers several purposes: It provides a benchmark for performance - if your "elegant" solution is slower or not as accurate than a brute force approach it is often pointless, or at the very least useless for a practical implementation. It provides a verification method - often it is much easier to verify the correctness of the brute force approach, as you rely on simple rules, and you can compare the output of the brute force solution to potential analytical solution to get a quick benchmark for how likely it is that the potential analytical solution is correct (and hence whether you want to spend time trying to prove that it is). It demonstrates feasability - if you can whip up a "quick and dirty" brute force solution to a problem the results that often gives you an intuitive insight into whether you are likely to get anywhere with a specific approach to an analytical solution.
For the worms problem, for instance, the brute force visulisation seem to indicate that a certain set of worms will never terminate. Without the brute force approach, someone might have spent years trying to provide a proof of whether or not they terminate based on the generic ruleset. With the brute force data, the burden of proof is much smaller: You can attack the specific behaviour the seemingly non-terminating worms enter into and attempt to prove that they can't break out of the specific patterns.
Once you get there, it's quite possible that you'll be able to generalize that proof to find an elegant proof for the entire problem with significantly less effort.
Don't joke about that - go read up on what Monsanto think is reasonable use of their seeds... Monsanto is one of the largest seed manufacturers in the world, and one of their pet peeves is that farmers who collect seeds from their harvest are undermining their income stream. To the point where they are trying to restrict it with license agreements, patents and genetically modified crops that don't produce viable seeds...
You couldn't go into a TV store and license a TV. You could rent one, buy one, lease one, buy one on credit or find some other interesting way of managing ownership, but licensing the TV set would not be one of them. Afterwards you would need to buy a license to have the TV, regardless of whether you obtained ownership to it, or merely usage rights.
Now, it's conceivable that you could structure a contract that would have similar implications to this "license", but it would be complicated, because when it comes to this kind of agreement the general approach of the courts (IANAL) is that "if it quacks like a duck..." - if you pay a one time fee for a product and have no further dealings with the company, and the product is a physical one, the courts would likely look at it and decide that for all practical purposes it has the properties of a purchase, and should be treated like a purchase regardless of any contract unless you had received clear benefits beyond a normal purchase that would make the restrictions equitable.
So why don't you just block outbound access to port 25 on your routers? Not exactly rocket science...
At a previous company I worked in, I had business dealings with a company run by two partners once. None of us suspected that one of them was on heroin until he ended up in hospital after finally falling apart. The thing that got him over the edge was reduced prices - he could finally get money enough for significantly higher doses than what he'd been able to previously. In hindsight we should've seen it - around the same time the newspapers reported record lows in heroin prices he was getting more and more erratic, taking ridiculous amounts of cellphone calls and running errands all the time. But he seemed normal most of the time we saw him.
Until that time, he'd (to our knowledge at least) not stolen to support his drug use, but with the low prices it apparently became to tempting (he could suddenly get "huge" quantities) and he ended up stealing from his partner.
I don't spend enough time reading up on drug crimes to have made up my mind regarding "soft" drugs, but it's way to simplistic to just assume that decriminalization of drug use will remove more crime than the ones that are decriminalized for all drug types.
I do agree, though, that people needs to be able to seek help without ending up with legal problems because of use.
Secondly, it's insanely hard and performance intensive to get the fitness function right. How do you judge if a program exhibits intelligence? Find a solution, and then move on to this: How do you judge if a program exhibits intelligence that might one day get sufficient to be recognized by average humans as intelligence, and is able to communicate with them?
One "solution" could be to simulate a "world" of sufficient complexity, and place lots of artifacts in it that would give benefits to "human like" intelligence. Leave lots of nice stuff that will increase survival rates, for instance, make them progressively harder to use depending on benefits, and provide language based clues to how to use them to award language understanding.
The problem? It took millions of years to evolve humans. Why would you expect that a computer system that likely isn't even fast enough to emulate a single human brain will be able to do it fast enough?
We do have plenty of possible approaches to AI - the problem is finding a reasonable combination of low computational requirements and reasonable "intelligence".
I'm not suggesting that we shouldn't aim for finding smarter ways to get AI than throwing CPU power at it, but the assumption that no current method could work if it just got CPU power enough is way too simplistic.
If we build a system designed to emulate the way humans interact, as in this case - an automated customer service representative - it IS likely to have many of these traits. Also, consider that no AI that DON'T have these traits will start e-mailing attorneys to prevent itself from being disconnected. We won't hear about something like this until we DO deal with an AI with survival "instincts", whether "evolved" by accident or designed in from scratch by the programmers.
And before you use the argument that it would have to be "that intelligent" to be able to make money by providing online research. Keep in mind that this computer is presented as a customer service system - it would be programmed specifically to handle text analysis and aggregating knowledge to respond in the best way possible. Combine that with speed, so that it would be able to respond near first to most queries it would take a shot at, and such a system could very well be able to do well without being "intelligent" - it wouldn't need to be able to draw any conclusions from the data.
That's a key observation - and the mock case is cleverly drawn up. A computer like this would ON PURPOSE be made to act and seem as much as possible like an intelligent person, because that's what makes sense in a customer service situation. But that would also make it exceedingly hard to get to the bottom of whether it is intelligent or self aware. Is it just Eliza on steroids, or a thinking, self aware, intelligent being that has rights?
... leaving three different angles of attack for the bot writers....
So now you're giving people a one in four chance of success. What the bot will then do is try a random answer, and if it fails it revisits the original page, gets a new problem and tries again. Voila, 25% success rate, and your e-mail system will be used for massive amounts of outbound spam.
Big problem with this: Let's say this type of challenge is given 1 out of a 100 times. It has the MASSIVE weakness that word lists with classifications are readily available (hint: computational linguistics - academics have spent decades preparing computer readable databases of stuff like this for use in their research), and if not can relatively quickly be built (think parsing dictionary.com output, looking for the category keywords). Say these method will solve 1 out of 10 of the challenges, which I think is very low given both the possibility of scanning a dictionary entry and availability of specialized word lists. That means 1 in a 1000. Which means somebody will hammer your registration server, and still be able to register 100's of accounts a day that they can abuse.
For the second you're raising the bar by complicating the parsing, but the question is: How would you generate the problems? If the sentence is machine generated, then the number of rules used will likely be small enough that the parser can be made pretty simple by spending a few hours analysing the result of repeatedly loading the page. The "problem" itself is again something you can write a small simple rule based solver for.
Your third suggestion relies on the same problem. Are you going to machine generate these? If so, they're likely to end up being in a form that is structured enough to parse relatively well. Are you going to type them in manually? If so, someone WILL set up a bot, record all failed attempts and manually enter them in their database for next time.
Also, keep in mind that not all your problems needs to be solvable. If 1% are solvable in a second, it still means that a bot trying just one at a time will be able to bypass your protection 8640 times a day. If what you're doing is harvesting webmail accounts to spam from, 8640 a day is going to get you far.
If Microsoft and/or Vulcan wanted Baystar to make an investment on their behalf in SCO, Baystar would do that if they were offered a reasonable management fee in the form of cash or shares. Hence it is meaningless to discuss who controls Baystar, only whether or not Baystar's investors include one or more companies or groups likely to have reasons to want an investment in SCO. Vulcan and Microsoft seems like pretty good candidates to me.
I agree that superficially it seems stupid to invest $50 million in SCO, but keep in mind we don't know the details of the deal. For what we know the conversion clause for the shares might allow the investment firm to convert their A shares and sell out long before the SCO vs IBM case goes to trial - they might be gambling on a continued short term growth of the stock not on the final outcome.
Remember that whether or not SCO posts a profit is irellevant for the investors in this case - the only thing that matters is whether or not the SCO share price keeps rising.
Also remember that this is an investment in preferential shares - for what we know it which might include special clauses pertaining to the distribution of dividencs in the case SCO manages to win any cases. Still a high risk gamble, but with a potentially huge upside.
To summarize, SCO can have done a variety of things that would sweeten the deal A LOT just because they know the cash infusion and the PR will likely boost their stock price in the short term.
Also, most of the money pumped into this will go to Chinese industry in the form of purchase of products and services.
China has more or less established now that there are only three countries that matters when it comes to space launches: US, Russia and China, and of the three the US is still paralyzed by the Columbia accident, and Russia can barely afford to meet their ISS obligations.
I assume that you would also like the US and Russia to cancel their space programs to benefit their poor as well?
Regardless of what you think about the Chinese government and the Red Army, the Long March was an incredible accomplishment that essentially saved the CCP from destruction and gave them an opportunity to regroup that ultimately gave them their victory.
Consider that the CCP went from being around a dozen people (literally, I believe it was 11 or 13) when it was founded in 1921 to governing China in just 28 years despite a civil war which more than once nearly decimated them, and Japanese invasion, and compare it to the military achievements of similar rebel movements.
The Long March still has more or less myth-like qualities for many people.
Except that China is one of few nations that has it's population growth under control - in fact their population is expected to start dropping in a few decades. Even if the above wasn't a bad joke, China just doesn't really have a population problem anymore.
Uhm, yeah, it's kind of been expected for MONTHS, and the guesses have gotten progressively better.
So where is the US at now, then? Considering the Chinese at the moment is the nation with the most advanced non-grounded space craft... And even with the shuttle in operation, the Chinese now have a craft that is vastly cheaper and more up to date technically.
If indeed they wanted to fake anything, the only thing worth faking would be the presence of an astronaut aboard the craft, not what for most purposes is a routine launch.
The earliest recorded usage was in the US in 1853. There's absolutely no indication the term originated in Australia, though there is apparently some evidence linking it to the California goldrush, where it may have been used as early as 1849. Many Australian prospectors took place, and one explanation for the term is that it was coined by Australian prospectors referring to the way claim-jumpers were prosecuted outside the legal system. In other words: It was likely a termed coined by Australians to refer to the (lack of) justice in the US.
ANYTHING that reduce mass and volume for a spacecraft will make a big difference.
In this case, anyone trying to implement and distribute IM products with typing notifications covered by the MS patent would (provided neither AOLs nor MS' patents are challenged and overturned) have to license both, while someone implementing IM without typing notifications would need to get a license only from AOL.
It's a common tactic for a company facing patents on a product they produce registered to a competitor to try to license as many extensions or alternative versions as possible to force their competitor into cross licensing, and also fairly common for companies to do the same for the purpose of forcing people licensing their competitors patent to pay them royalties too by making the competitors patent more or less worthless on it's own.
Assuming using a PC can't give redundancy and resilience against failures is extremely presumptious. But for areas that currently don't have ANY coverage, even a desktop PC powered base station would be an improvement.