No, there's no way to prove it, but, do you have any indication that Symbian has up to 85,000 apps available for the platform, or even 8,500? My guess would be that 850 is of the right order. Why do you think it's more than 85,000?
I know this one differently, and with a different pun:
A man was flying in a balloon, and lost his way. Luckily, on a hill, he spotted a guy that he could steer to. He asked the guy: "Where am I>". The guy answered: "You are in a balloon". "Oh", said the balloon guy, "you must be a software developer". "Why", asked the guy on the hill. "Well", said the balloon guy, "your answer is factually correct, but completely useless". "Oh", said the guy on the hill, "then you must be a manager". "Why" asked the guy in the balloon. "Simple", said the guy on the hill: "you don't know where you are, you don't know where you are going, and suddenly it is all my fault!"
When you have your head in an oven, and your feet in a freezer, on average, you body temperature is a-ok. Same goes for life-span. Historically, the expected life-span of people surviving childhood is much larger than the expected life-span at birth. I.e., children used to die as flies. So much, that your entire statistic can be explained by more people not dying in the cradle (with elderly people dying as quickly as in the 19th century). I'm not saying that this is the case, but your statistic for showing this is useless.
Cue back to the eighties. Hardware vendors were selling hardware, software was there to sell hardware. Then Bill came along and proved that software could be sold on its own right. An industry was born.
Same time-frame. Hardware vendors were selling hardware, software was there to sell hardware. Industry decided that software was company secret. Richard Random Hacker wanted to tune his printer driver to be able to print. HP didn't want to open up the source code to that driver. This annoyed Richard, and a counter-movement was born.
Now Nokia wants to hire Richard to beat Bill. Small world.
No serious. Part of the reason we're in such a mess is because of over-efficient business. Business doesn't keep items in stock, they don't have spare capacity, everything is optimized so that every penny is squeezed from the business as long as everything runs as normal. The slightest hiccup will make the whole house of cards come tumbling down.
Contrast this with nature. Overefficient predators will die out (as they deplete their source of food), and everything is set up to be adaptive, not optimized. So yes, efficiency is overrated, adaptability is better. In the long run we're dead, but hopefully not extinct.
I don't think it's Obama who was shocked to discover that closing Gitmo wasn't something that could be done in a day. That would be the ones crying betrayal because he signed the order to close it, but because a lot of those prisoners simply can't be let go, they have to be moved to the mainland and integrated into the justice system, there's a lot of deal-making and planning required. "Close Gitmo now!" is the soundbite that people are unrealistically clinging to.
WTF? 'Move to the mainland': put them in a plane and bring them to the mainland. Where was the difficulty in moving them to Gitmo?? 'Integrate them in the justice system': allocate a block of cells and put them in there. Sort out the rest later. Deal-making and planning? Use federal prisons and don't accept any argument from the idiots that think that terrists can spontaneously create a nuclear explosion.
That Gitmo still exists is inexcusable and the protests from the population of the US against having them in their prisons says a lot about the cowardice of the US as a whole.
On the other hand, you could look at the stuff he did achieve in the last 8 months. He has changed the role of the US from the biggest threat to peace in the world to a strong but neutral party.
His address in Cairo has made a deep impact in the muslim world, deep enough so that the 700 million in the West are now in a sort of cease-fire situation rather than in an escalating conflict with the 1.5 billion muslims in the world.
He has killed the rocket shield, thereby deescalating the conflict with Russia about this, resulting in an agreement around nuclear disarmament between the US and Russia. He has opened room for negotiation where the door was previously shut by US arrogance.
In general he has provided a message that the strongest power in the world is now open for cooperation rather than being a dumb bully.
All in all, not bad, I'm not sure anyone has ever achieved so much in so little time. I think it's a bit early for a Nobel prize, but Obama did change the world past 8 months... but unfortunately not America.
Every year, the Norwegian Nobel Committee sends out thousands of letters inviting qualified people to submit their nominations for the Nobel Peace Prize. The names of the nominees and other information about the nominations cannot be revealed until 50 years later.
As members of governments are allowed to send in nominees, I am pretty sure that G. W. Bush and any POTUS before him has been nominated for this particular prize every year they were in office. No this doesn't imply that they decide on the nomination date. Why do you think that? Are you stupid?
Yes, everything has some form of bias. That's obvious and irrelevant. It is the degree of bias that matters, and and comparing to American sources, the BBC can be considered largely objective and unbiased, while representing the Western point of view (chew on that for a while to see if your mind can understand that the apparent contradiction is irrelevant)
If not, then couldn't the same mechanism that made the Earth warmer back then, also be responsible now?
Of course it could be that mechanism, and it could also be fairies. What we have is a situation where we have a very noticable increase in greenhouse gasses in the atmosphere, and increasingly stronger evidence that the earth is warming up. Note the order. We first saw the greenhouse gasses, and global warming has been predicted for a long time. Personally I am aware of the global warming threat since the beginning of the eighties. . This was based on our limited understanding of the effect of Co2 on the atmosphere. And now we are getting some very strong evidence that corroborate these predictions. Yes, it could still be fairies, but are you willing to bet your children's life on that?
At the very least, Obama should have ditched GM. There was only one thing wrong with GM: they couldn't make a profit building and selling cars. Their marketing was okay, the US population liked the SUVs and wanted to buy them, even when oil went up. There was time enough to move to somewhat smaller cars if the market wanted that. But the problem was: they couldn't build cars and do this profitably. Their way of producing cars was bad. Now they are bailed out for whatever reason, and they will still not be able to build cars and make a profit. Ford can, Toyota can, GM cannot. Free market capitalism 101: no profit, you die. There were tons of ways Obama could have softened the blow for the employees of GM, and even for saving car manufacturing ability. But keeping GM alive is not is. It's a fucked company, and should die.
This is totally different than an imploding financial sector due to cowboy 'investors' legally run up tabs of a couple of trillion. The cowboys need to be leashed in and kicked to a hedge fund where they can only kill off investor money, not wreck the economy, and banking should become boring again. But also that is not what's being done.
Yep, I think the innate American desire to be fucked over by their corporate overlords will prevent the US from sliding into a European style society. Americans just like to be indentured servants.
I kind of think that my dog by not pooping in my neighbour's garden has done more for world peace than the new laureate. It's interesting that 'anybody but Bush (as long as he is a Democrat)' gets an automatic Nobel prize. This hints to an interesting new strategy of 'good cop/bad cop' that should land the US a Nobel peace prize every other decade.
One thing you forget to cater in is that airplanes are typically much further from anyplace interesting than train stations. In Europe, train stations are usually down-town, and you can walk to a hotel from there. I understand that in a typical American city does not have such a down-town, but still, train stations can be much closer to whatever is the actual middle of nowhere in the typical American town. Contrast this with an airport which typically is 40 mins or more away from where people actually are.
He's been building AMD boxes for his customers. Now he built an AMD box for himself. Why is this not eating his own dogfood? Following your logic, only farmers can eat their own dogfood, not dogfood-makers.
I think this happened when we as a society decided that we should live by a system of law, rather than by force of arms alone. This happened a few centuries/millenia ago, same time when we decided to lock up Neanderthals that think killing people for trespassing alone is a good way to run a society. Although many people (most often from the South or Colorado) often claim that they will shoot on sight whenever somebody even but touches their lawn, remarkably few actually do.
You Germans are wimps in obfuscating your language. We Dutch have wonderful words with at least 8 in a row: "angstschreeuw" (cry of fear). For extra points, 7 vowels in a row: "papagaaieeieren" (parrot eggs), though not sure if the latter is valid anymore given the last spelling reform.
Where are the Welch? They should be able to win this.
Honestly, thanks for the tip. Never realized that F5 would reload the page. In hindsight it makes sense, but I always used the mouse for reloading, not the keyboard.
Joining forces with the grandparent: how dense are you? The EU is not interfering in the US, they are interfering in the EU. If it were possible for Oracle to organize itself in such a way that it merges with Sun in the US, but be separate, competing entities in the EU, the EU would have no beef with Oracle. Why do you feel that US law should apply outside of the US?
It would be an interesting world in which multinational companies only have to abide the law of the country their headquarters is in: I expect that each Fortune-500 CEO would soon become el Presidente of some Pacific island-country.
And for each Google, how many other companies are out there that need a relational system to do their payroll, their management information system, their data warehouse, their inventory control, their... Come to think of it, do you actually think Google runs their payroll from bigtable?
There's a lot of argument over why ensemble techniques work well in general, when using them on well-posed statistical problems.
Not sure if there's a lot of argument, as the bias-variance decomposition does seem to give some critical insight in why ensemble techniques work. Couldn't find a good link, but loosely every statistical prediction method will make errors due to bias (being systematicaly wrong, always in the same direction), and errors due to variance (being highly dependent on the fitness data, overfitted). Error methods such as sum-of-squares can be readily decomposed into a term for bias and a term for variance.
Some methods will have errors mainly due to bias (linear methods). Others have error mainly due to variance (neural nets for instance). Whichever method has the lowest sum of the errors - bias + variance - wins. This holds for any prediction method, so also for an ensemble method. If you take for instance the ensemble method of bagging, you will use the same method on differently sampled data (Bootstrap AGGregated). This has as an effect that you average out all the error due to variance, and are left with the error due to bias. If you happen to use a low bias method (such as a neural network or a decision tree), this works fine.
For the recommender challenges, something even niftier is happening. Here many different methods are aggregated, each with different biases and different variances. So, in theory, when doing this, one should be able to average out both bias and variance error, and converge on the (Bayesian) optimal predictor.
Note however that averaging is in itself a prediction method, so ensemble methods have their own bias and variance tradeoff. Fortunately, computing the mean is an unbiased and low variance method in its own right.
What I always found very interesting about ensemble methods is that they effectively contradict Occam's razor, in that the end result of an ensemble is not a single theory that predicts the data well, but a whole set of mutually contradicting theories that each hold some of the truth. The ensemble result might actually be huge, even when the system is simple.
Slashdot Mirror
No, there's no way to prove it, but, do you have any indication that Symbian has up to 85,000 apps available for the platform, or even 8,500? My guess would be that 850 is of the right order. Why do you think it's more than 85,000?
A man was flying in a balloon, and lost his way. Luckily, on a hill, he spotted a guy that he could steer to. He asked the guy: "Where am I>". The guy answered: "You are in a balloon". "Oh", said the balloon guy, "you must be a software developer". "Why", asked the guy on the hill. "Well", said the balloon guy, "your answer is factually correct, but completely useless". "Oh", said the guy on the hill, "then you must be a manager". "Why" asked the guy in the balloon. "Simple", said the guy on the hill: "you don't know where you are, you don't know where you are going, and suddenly it is all my fault!"
When you have your head in an oven, and your feet in a freezer, on average, you body temperature is a-ok. Same goes for life-span. Historically, the expected life-span of people surviving childhood is much larger than the expected life-span at birth. I.e., children used to die as flies. So much, that your entire statistic can be explained by more people not dying in the cradle (with elderly people dying as quickly as in the 19th century). I'm not saying that this is the case, but your statistic for showing this is useless.
Can someone please translate the summary into English?
Same time-frame. Hardware vendors were selling hardware, software was there to sell hardware. Industry decided that software was company secret. Richard Random Hacker wanted to tune his printer driver to be able to print. HP didn't want to open up the source code to that driver. This annoyed Richard, and a counter-movement was born.
Now Nokia wants to hire Richard to beat Bill. Small world.
No serious. Part of the reason we're in such a mess is because of over-efficient business. Business doesn't keep items in stock, they don't have spare capacity, everything is optimized so that every penny is squeezed from the business as long as everything runs as normal. The slightest hiccup will make the whole house of cards come tumbling down.
Contrast this with nature. Overefficient predators will die out (as they deplete their source of food), and everything is set up to be adaptive, not optimized. So yes, efficiency is overrated, adaptability is better. In the long run we're dead, but hopefully not extinct.
Can this in any way be related back to the fifth horseman: the EU competition regulators that demanded interoperability from Microsoft?
WTF? 'Move to the mainland': put them in a plane and bring them to the mainland. Where was the difficulty in moving them to Gitmo?? 'Integrate them in the justice system': allocate a block of cells and put them in there. Sort out the rest later. Deal-making and planning? Use federal prisons and don't accept any argument from the idiots that think that terrists can spontaneously create a nuclear explosion.
That Gitmo still exists is inexcusable and the protests from the population of the US against having them in their prisons says a lot about the cowardice of the US as a whole.
If not being a dickhead is all you need to win a Nobel Peace Prize, you won't get one.
His address in Cairo has made a deep impact in the muslim world, deep enough so that the 700 million in the West are now in a sort of cease-fire situation rather than in an escalating conflict with the 1.5 billion muslims in the world. He has killed the rocket shield, thereby deescalating the conflict with Russia about this, resulting in an agreement around nuclear disarmament between the US and Russia. He has opened room for negotiation where the door was previously shut by US arrogance. In general he has provided a message that the strongest power in the world is now open for cooperation rather than being a dumb bully.
All in all, not bad, I'm not sure anyone has ever achieved so much in so little time. I think it's a bit early for a Nobel prize, but Obama did change the world past 8 months... but unfortunately not America.
People from Slashdot. From the actual site:
As members of governments are allowed to send in nominees, I am pretty sure that G. W. Bush and any POTUS before him has been nominated for this particular prize every year they were in office. No this doesn't imply that they decide on the nomination date. Why do you think that? Are you stupid?
Yes, everything has some form of bias. That's obvious and irrelevant. It is the degree of bias that matters, and and comparing to American sources, the BBC can be considered largely objective and unbiased, while representing the Western point of view (chew on that for a while to see if your mind can understand that the apparent contradiction is irrelevant)
Trust in the USA as a force for good, not self-centered evil.
Of course it could be that mechanism, and it could also be fairies. What we have is a situation where we have a very noticable increase in greenhouse gasses in the atmosphere, and increasingly stronger evidence that the earth is warming up. Note the order. We first saw the greenhouse gasses, and global warming has been predicted for a long time. Personally I am aware of the global warming threat since the beginning of the eighties. . This was based on our limited understanding of the effect of Co2 on the atmosphere. And now we are getting some very strong evidence that corroborate these predictions. Yes, it could still be fairies, but are you willing to bet your children's life on that?
This is totally different than an imploding financial sector due to cowboy 'investors' legally run up tabs of a couple of trillion. The cowboys need to be leashed in and kicked to a hedge fund where they can only kill off investor money, not wreck the economy, and banking should become boring again. But also that is not what's being done.
Yep, I think the innate American desire to be fucked over by their corporate overlords will prevent the US from sliding into a European style society. Americans just like to be indentured servants.
I kind of think that my dog by not pooping in my neighbour's garden has done more for world peace than the new laureate. It's interesting that 'anybody but Bush (as long as he is a Democrat)' gets an automatic Nobel prize. This hints to an interesting new strategy of 'good cop/bad cop' that should land the US a Nobel peace prize every other decade.
One thing you forget to cater in is that airplanes are typically much further from anyplace interesting than train stations. In Europe, train stations are usually down-town, and you can walk to a hotel from there. I understand that in a typical American city does not have such a down-town, but still, train stations can be much closer to whatever is the actual middle of nowhere in the typical American town. Contrast this with an airport which typically is 40 mins or more away from where people actually are.
He's been building AMD boxes for his customers. Now he built an AMD box for himself. Why is this not eating his own dogfood? Following your logic, only farmers can eat their own dogfood, not dogfood-makers.
I think this happened when we as a society decided that we should live by a system of law, rather than by force of arms alone. This happened a few centuries/millenia ago, same time when we decided to lock up Neanderthals that think killing people for trespassing alone is a good way to run a society. Although many people (most often from the South or Colorado) often claim that they will shoot on sight whenever somebody even but touches their lawn, remarkably few actually do.
Where are the Welch? They should be able to win this.
Honestly, thanks for the tip. Never realized that F5 would reload the page. In hindsight it makes sense, but I always used the mouse for reloading, not the keyboard.
It would be an interesting world in which multinational companies only have to abide the law of the country their headquarters is in: I expect that each Fortune-500 CEO would soon become el Presidente of some Pacific island-country.
And for each Google, how many other companies are out there that need a relational system to do their payroll, their management information system, their data warehouse, their inventory control, their ... Come to think of it, do you actually think Google runs their payroll from bigtable?
Not sure if there's a lot of argument, as the bias-variance decomposition does seem to give some critical insight in why ensemble techniques work. Couldn't find a good link, but loosely every statistical prediction method will make errors due to bias (being systematicaly wrong, always in the same direction), and errors due to variance (being highly dependent on the fitness data, overfitted). Error methods such as sum-of-squares can be readily decomposed into a term for bias and a term for variance.
Some methods will have errors mainly due to bias (linear methods). Others have error mainly due to variance (neural nets for instance). Whichever method has the lowest sum of the errors - bias + variance - wins. This holds for any prediction method, so also for an ensemble method. If you take for instance the ensemble method of bagging, you will use the same method on differently sampled data (Bootstrap AGGregated). This has as an effect that you average out all the error due to variance, and are left with the error due to bias. If you happen to use a low bias method (such as a neural network or a decision tree), this works fine.
For the recommender challenges, something even niftier is happening. Here many different methods are aggregated, each with different biases and different variances. So, in theory, when doing this, one should be able to average out both bias and variance error, and converge on the (Bayesian) optimal predictor. Note however that averaging is in itself a prediction method, so ensemble methods have their own bias and variance tradeoff. Fortunately, computing the mean is an unbiased and low variance method in its own right.
What I always found very interesting about ensemble methods is that they effectively contradict Occam's razor, in that the end result of an ensemble is not a single theory that predicts the data well, but a whole set of mutually contradicting theories that each hold some of the truth. The ensemble result might actually be huge, even when the system is simple.