I agree that population density is not a perfect predictor of the difficulty in implementing broadband, but it's much more relevant that just stating how big the country is - which is what the great-great-grandposter I was responding to was doing.
That indicates that while the average US population density is twice that of NZ, the imbalance (i.e., density of densely populated areas compared to density of sparsely populated areas) is far greater in the US.
That'd definitely a non-sequitur - you haven't provided any data on NZ to support that.
If we're going to be picky, then none of the population density data you or I could come up with will give a watertight prediction of the cost of implementing broadband. An empty field more sparsely populated (zero population density) than any county in the US. But note that a more heterogeneously distributed population is not necessarily more expensive to wire up than a more uniform one - quite the opposite, in fact. Vast uninhabited wastes do not need broadband at all; the individual cost of wiring up a small number of isolated people is high, but the cost per head of population in the country can be low. Meanwhile, the flip side of an uneven population distribution is that many people live close to each other, so are cheaper to wire up.
Wiring up the US is indeed a much bigger problem than wiring NZ, but you also have proportionally more resources to spend on it.
The economics of broadband depend on population density - it's more expensive to connect the same number of people over a greater distance.
Population of US: 320 million; area 4 million sq miles; population density 80 per square mile.
Population of NZ: 4 million, area 103 sq. miles, population density 40 people per square mile.
New Zealand has HALF the population density of the US, so connecting it up is actually MORE difficult (per capita).
And these things DO scale well: revenue, and workforce, is proportional to population size. Wiring up two towns is only twice the cost of wiring up one town
(Of course, things are slightly more complicated because you also need to connect up the two towns. However, those overheads only scale logarithmically with population size so are much less important. Also, there's an economy of scale I didn't take account of in the above: length of wiring is not proportional to area, but to linear size (square root area)).
Are there any health implications of micro plastics in salt? That was suspiciously left out of the article for some reason.
No it wasn't - the National Geographic article had a section titled "Is this harmful?"
It wasn't addressed in the original research article in "Environmental Science and Technology", but that's because it's a different question that requires different expertise and completely different sort of data. Hardly "suspicious" that the researchers addressed a valid question in their own area of expertise. Assessing the health risks of environmental exposure to microplastic is much more challenging, both economically and ethically. It would be profligate to do so without quantifying the exposure first.
You don't think consumers will want valid non-falsified facts about their electric cars too?
Won't they care about kilometers per kilowatt, or the expected lifetime of the power cells?
I would welcome this, and don't imagine this will show electric vehicles in a poor light. My 4 year old Nissan Leaf shows no significant battery degradation, and can still achieve better range than the official US Environmental Protection Agency figures. The manufacturers' figures are always an exaggeration, but the real-world carbon emissions of electric cars are still about a third of a fossil-fuel powered equivalent - and will only get better as electricity generation gets greener.
Your final point about the importance of random sampling is a good one, but on the way I'm afraid there are a number of misconceptions about how and why statistics works:
Statistical methods are based on what are known as "stable distributions".
No, statistical methods are based on whatever probability distribution is appropriate - they are not limited to stable distributions
A stable distribution is one where a subset of examples, selected randomly, will have the same characteristics as the full set.
Not quite: a stable distribution is one where the *sum* of a sample of independent random variables has the same distribution.
so if you have a bell curve population and you select a sample at random, then the sample mean will tend towards the population mean and the sample width will tend towards the population width.
The sample mean will always approach the population mean (in the limit of large sample size), whatever the distribution - stable or not. This is known as the "law of large numbers", and all that it requires is that the sample is unbiased. Similarly, the sample variance will also approach the population variance (if this is finite) - whatever the probability distribution.
It is this characteristic that lets us extend measurements of characteristics from a subset to the characteristics of the whole population.
No, statistical methods aren't (necessarily) based on the properties of the mean of a sample. The majority of them infer the properties of the population from a "likelihood" function, that quantifies the probability of getting the full set of measurements in terms of the parameters that describe the population. This doesn't require an underlying stable distribution - generalised linear models, for instance, can be based on any probability distribution you like. Bayesian methods don't even require a parametric form for the population distribution. It does rely on the sample being independent, or at least that the correlations between measurements is quantifiable.
I wish we could avoid this usage of the word "safely" - the plane may have landed without loss of life, but the landing was certainly dangerous. If I drive a car with my eyes closed, and happen not to crash, no-one would say that I was driving "safely".
I often read that a plane suffered incident "X" in flight, and then landed safely. I can't tell whether the writer means that incident "X" did not make the landing hazardous, or whether the landing was actually hazardous but resulted in no death or injury. Far clearer to say "the plane landed without further incident" if that is what you mean.
OMG you found the one conservative professor who is 73 who may still be teaching at Oxford!
It only took me a second or two to think of a very high profile conservative thinker who has recently been a professor in Oxford - enough for you to lose the bet you were prepared to make (that there had been none in 30 years).
liberal BS is killed by the antiseptic of objective truth
and sweeping, unsupported statements can be disproved by one data point. Can you point me to your evidence that less than 1% of Oxford academics are conservatives?
A couple more Republican's voted against it than Democrats.
"insightful"?
If you look at the data (scroll down to "grouped by vote position", all but three Republicans voted against it, and all but one democrats in favour of it. So, a lot more than "a couple".
that's the way the majority of Journals operate, most don't care that the author will hand out electronic copies to anyone who asks, many will post it on their university's web site.
Actually, the publishers DO care about generating income from their copyright, and for instance require extra subscription fees from universities in order for students to be allowed to copy articles that they already have access to. I'm repeatedly warned that my University could be fined if I copy a figure from a paper into my lecture slides and then "publish" them by distributing them to my students.
Academic publishing is a two way street the academics and journals need each other because "publish or perish" applies to both sides.
Academics do need to publish but they don't need commercial publishers to do so. There's a growing movement against the traditional journals - fuelled by the extortionate fees required for electronic access to the catalogue of publishing houses such as Springer and Elsevier. Universities cannot function without access to the literature, and this is your tax dollar being funnelled into the pockets of the publishers.
sure subscriptions are expensive compared to (say) people magazine, but if you pay peanuts for a job then only monkeys are going to apply.
The people who do the hard work behind the articles - the authors, the editors, and the reviewers - are not paid by the journal. The only paid staff are administrators and copy editors, who in my experience introduce more mistakes into the text than they correct.
Agreed, many trivial things can be handled by email. However, if you need to discuss or explain anything that's even slightly difficult conceptually then nothing beats an interactive conversation. It's much easier to gauge how much your peer understands (and therefore background you need to cover) when you can interact in real time.
The aircraft overran runway 4 while landing. A malfunction of the onboard flight computers prevented power from being reduced to idle, which inhibited thrust reverse and spoilers from being used. The offending engine was shut down, and brakes applied, but the aircraft was unable to stop before the end of the runway
I couldn't read the article you referenced ("server not responding"), but the accident report states this was caused by pilot error, not malfunctioning computers. From Wikipedia: "A selection by the pilot of the wrong mode on the onboard flight computers prevented power from being reduced to idle, which inhibited thrust reverse and spoilers from being used. The offending engine was shut down, and brakes applied, but the aircraft was unable to stop before the end of the runway".
There's a surprising number of people who believe that the high level of automation on Airbus is intrinsically more dangerous, but the figures show that the Airbus A320 is the safest narrow-body jet you can fly on. It's true that automated stuff can go wrong, but this can be more than compensated for by the ways it makes flying (or driving) safer.
3.3kW/h is TONS to keep any reasonably insulated house scorching hot
I assume you mean 3.3 kW, but I would be interested to see a source stating that this is enough to maintain a 20 degree (Celsius) difference between inside and outside for any "reasonably insulated" house.
And for the love of god you moron you didn't say "10kW/h" you said "10 kW/h for 10 hours"
There's no need to be offensive. It was you who said "10kW/h for 10 hours". You're using kW as a measure of energy, and kW/h as a measure of power. This is incorrect, as I've pointed out. "kW/h" would be a unit at the rate at which power changes.
How dumb are you?
Smart enough to know the difference between kW, kWh and kW/h. Here's a reference which may help you understand.
Firstly, according to your figures above (where you say 100 kWh is equivalent to driving 180 miles), 60 miles is equivalent to 33 kWh. This corresponds to 10 hours at 3.3 kW. 3.3kW is not enough to keep a house warm in subzero temperatures unless it has no windows.
Secondly, 10 kW/h means '10 kilowatts per hour', which has no physical meaning whatever. A Watt is a unit of power, not energy.
Note that kW is a unit of power, kWh is a unit of energy
Your figures weaken the support for your argument (that staying at home is greener than driving to work) significantly: they suggest it's greener to drive to work as long as it's less than 90 miles away from your home.
I disagree with your calculation, but the point remains broadly the same. Here's an alternative calculation. According to Wikipedia, 1 US gallon of gasoline is equivalent to 33 kWh, so 100kWh is equivalent to 3 gallons US. At 45 mpg US you can go 135 miles, not 180.
You can look at this the other way round. Driving 60 miles at 45mpg US is equivalent to 44kWh of energy usage. For it to be worth staying at home for 10 hours, you would have be able to keep your home warm with less than 4.4 kW. That's roughly the power output of two old-fashioned electric bar heaters. Do you get freezing temperatures where you live?
It's difficult to be more precise because the figures depend on lots of things - how cold is the place you live, how efficient is your car, how you define "equivalence" between gasoline and other energy forms - but you've illustrated the following point very well: most people wrongly assume that the energy required to heat their home is negligible relative to the cost of driving anywhere.
How much energy does it actually take to heat your house... because it's not 10kW/h.
"Kw/h" is not a unit of either energy or power. However, I can confirm that 10kW is approximately the power needed to keep my house comfortable in the winter. I know this because I know the ratings of the radiators in the house at 60 celcius, and I also know that they need to be kept at close to 60 celcius more or less constantly.
We have a small semi-detached house with cavity wall insulation, loft insulation, and double glazing. We still need over 10kW to heat in the cold months. If we're out of the house for 10 hours, we save roughly 100 kW hours (*). How far will your car go on that amount of energy?
(*) Of course, it's not that simple because we have to use extra heat to bring the house back to normal temperature when we get home. A more accurate analysis would compare the temperature-time graphs for the two scenarios and use Newton's law of cooling. Nevertheless, the above figures are roughly correct.
No necessarily. If you don't work at home, you also don't have to keep your home warm (or cool) enough to be comfortable during the day. The office, on the other hand, will be kept at a reasonable temperature whether you're there or not.
My wife's work is about 30 miles away, but she works from home most days. We calculated that, on the coldest winter days, the carbon cost of driving to work was about the same as the extra heating that would be needed if she stayed at home. If you have a shorter commute, or have a greener method of transport than driving a car, it's quite likely that it's greener to work in the office than to work at home.
they eat corn alright, as does most of the world, in the form of processed food. You find corn derivatives in a bewildering varieties of industrial foods.
You imply that the French, and indeed the rest of the world, eat significant amount of processed food. It's difficult to get hard data on this, but my impression (from having lived there) is that processed food is a much smaller part of their diet than in the US. This article says that Americans eat rather more processed food than other countries, but it's difficult to compare because "baked goods" and "ready-to-eat" in the US and in France are rather different.
On the other hand, "most of the world" is certainly not eating significant amounts of industrial food - in China and India it's almost unheard of.
On Airbus vehicles, if the avionics computers crash, the airplane crashes. There's exactly ZERO way to pilot the computer manually in such a failure.
Completely untrue. When the avionics 'crash', the flight system progresses through 'alternate' to 'direct' law where the pilot has direct control of the plane.
Moreover, the avionics system can and does overrule pilot input. So if you get sensor malfunctions like this, even if the pilot is trying desperately to save the plane, the computer can still crash you.
Have a look at the statistics (pages maintained by a pro-Boeing pilot, by the way) and you'll see (i) for all your hysterical fear of Airbus aircraft, the fly-by-wire Airbus aircraft (i.e. all except A300 and A310) are just as safe as their Boeing counterparts (ii) there are no examples of an Airbus crash caused by the computer overriding the will of the pilot.
I see your point, but if x1 and (x1+x2)/2 make very different predictions then you don't expect them both to be equally good at describing a data set generated by x1 - unless (i) your data are less informative than your prior, or (ii) unless your predictions are in a regime where the parameters are identifiable but the training data are in a regime where they are not. I admit that this might often be the case when trying to modelling real systems...
Sure, the posterior mean or maximum might be very different from the true parameter value, but the the true value should sit somewhere in the full posterior distribution. If parameters have non-identifiability issues then the posterior should be very flat, but if you base your predictions on the posterior then this will show up in the distribution of your predictions. I would have thought this would only lead to a bias if your prior were TOO informative, wouldn't it?
The phenomenon this guy has observed is nothing to do with chaos theory, as several posters think, but rather to do with error propagation and model uncertainty. This is an issue whether the model is chaotic or not. His mistake is to think that calibration has to choose a single set of parameters, and then one has to make a single prediction from the model. Statistical methods can take into account many sources of uncertainty, including the range of parameters that could have produced the original data and intrinsic stochasticity in the model. The best way to do this is using Bayesian techniques.
You're still limited by how realistic your model is, and this is likely to be the real problem with economic models. However, Carter's argument (that it's fundamentally impossible to fit a model to itself and then make consistent predictions) is wrong.
Researchers and peer reviewers are not paid for their work but academic publishers have said such a business model is required to maintain quality.
The publishers are lying here to protect their cash cow. What maintains quality is the peer review system (which the journals do not pay for). The transfer of copyright to the publisher allows them to hold Universities to ransom - universities cannot function without access to the literature (present and past), and the costs of online access to journals have been spiraling over the past few years at a time when the publishers' actual costs are going down. After all, they don't pay for the research to be carried out, nor do they pay the academic editors or the reviewers, nor do they even need to typeset the document now that everyone submits a machine-readable copy.
Slashdot Mirror
What you say is a non sequitur.
I agree that population density is not a perfect predictor of the difficulty in implementing broadband, but it's much more relevant that just stating how big the country is - which is what the great-great-grandposter I was responding to was doing.
That indicates that while the average US population density is twice that of NZ, the imbalance (i.e., density of densely populated areas compared to density of sparsely populated areas) is far greater in the US.
That'd definitely a non-sequitur - you haven't provided any data on NZ to support that.
If we're going to be picky, then none of the population density data you or I could come up with will give a watertight prediction of the cost of implementing broadband. An empty field more sparsely populated (zero population density) than any county in the US. But note that a more heterogeneously distributed population is not necessarily more expensive to wire up than a more uniform one - quite the opposite, in fact. Vast uninhabited wastes do not need broadband at all; the individual cost of wiring up a small number of isolated people is high, but the cost per head of population in the country can be low. Meanwhile, the flip side of an uneven population distribution is that many people live close to each other, so are cheaper to wire up.
Wiring up the US is indeed a much bigger problem than wiring NZ, but you also have proportionally more resources to spend on it.
The economics of broadband depend on population density - it's more expensive to connect the same number of people over a greater distance.
Population of US: 320 million; area 4 million sq miles; population density 80 per square mile.
Population of NZ: 4 million, area 103 sq. miles, population density 40 people per square mile.
New Zealand has HALF the population density of the US, so connecting it up is actually MORE difficult (per capita).
And these things DO scale well: revenue, and workforce, is proportional to population size. Wiring up two towns is only twice the cost of wiring up one town
(Of course, things are slightly more complicated because you also need to connect up the two towns. However, those overheads only scale logarithmically with population size so are much less important. Also, there's an economy of scale I didn't take account of in the above: length of wiring is not proportional to area, but to linear size (square root area)).
Are there any health implications of micro plastics in salt? That was suspiciously left out of the article for some reason.
No it wasn't - the National Geographic article had a section titled "Is this harmful?"
It wasn't addressed in the original research article in "Environmental Science and Technology", but that's because it's a different question that requires different expertise and completely different sort of data. Hardly "suspicious" that the researchers addressed a valid question in their own area of expertise. Assessing the health risks of environmental exposure to microplastic is much more challenging, both economically and ethically. It would be profligate to do so without quantifying the exposure first.
You don't think consumers will want valid non-falsified facts about their electric cars too? Won't they care about kilometers per kilowatt, or the expected lifetime of the power cells?
I would welcome this, and don't imagine this will show electric vehicles in a poor light. My 4 year old Nissan Leaf shows no significant battery degradation, and can still achieve better range than the official US Environmental Protection Agency figures. The manufacturers' figures are always an exaggeration, but the real-world carbon emissions of electric cars are still about a third of a fossil-fuel powered equivalent - and will only get better as electricity generation gets greener.
Your final point about the importance of random sampling is a good one, but on the way I'm afraid there are a number of misconceptions about how and why statistics works:
Statistical methods are based on what are known as "stable distributions".
No, statistical methods are based on whatever probability distribution is appropriate - they are not limited to stable distributions
A stable distribution is one where a subset of examples, selected randomly, will have the same characteristics as the full set.
Not quite: a stable distribution is one where the *sum* of a sample of independent random variables has the same distribution.
so if you have a bell curve population and you select a sample at random, then the sample mean will tend towards the population mean and the sample width will tend towards the population width.
The sample mean will always approach the population mean (in the limit of large sample size), whatever the distribution - stable or not. This is known as the "law of large numbers", and all that it requires is that the sample is unbiased. Similarly, the sample variance will also approach the population variance (if this is finite) - whatever the probability distribution.
It is this characteristic that lets us extend measurements of characteristics from a subset to the characteristics of the whole population.
No, statistical methods aren't (necessarily) based on the properties of the mean of a sample. The majority of them infer the properties of the population from a "likelihood" function, that quantifies the probability of getting the full set of measurements in terms of the parameters that describe the population. This doesn't require an underlying stable distribution - generalised linear models, for instance, can be based on any probability distribution you like. Bayesian methods don't even require a parametric form for the population distribution. It does rely on the sample being independent, or at least that the correlations between measurements is quantifiable.
One-wing F-15, landed safely...
I wish we could avoid this usage of the word "safely" - the plane may have landed without loss of life, but the landing was certainly dangerous. If I drive a car with my eyes closed, and happen not to crash, no-one would say that I was driving "safely".
I often read that a plane suffered incident "X" in flight, and then landed safely. I can't tell whether the writer means that incident "X" did not make the landing hazardous, or whether the landing was actually hazardous but resulted in no death or injury. Far clearer to say "the plane landed without further incident" if that is what you mean.
OMG you found the one conservative professor who is 73 who may still be teaching at Oxford!
It only took me a second or two to think of a very high profile conservative thinker who has recently been a professor in Oxford - enough for you to lose the bet you were prepared to make (that there had been none in 30 years).
liberal BS is killed by the antiseptic of objective truth
and sweeping, unsupported statements can be disproved by one data point. Can you point me to your evidence that less than 1% of Oxford academics are conservatives?
I am willing to bet that you couldn't find a single conservative professor at Oxford in the last 30 years.
How about Roger Scruton?
A couple more Republican's voted against it than Democrats.
"insightful"?
If you look at the data (scroll down to "grouped by vote position", all but three Republicans voted against it, and all but one democrats in favour of it. So, a lot more than "a couple".
that's the way the majority of Journals operate, most don't care that the author will hand out electronic copies to anyone who asks, many will post it on their university's web site.
Actually, the publishers DO care about generating income from their copyright, and for instance require extra subscription fees from universities in order for students to be allowed to copy articles that they already have access to. I'm repeatedly warned that my University could be fined if I copy a figure from a paper into my lecture slides and then "publish" them by distributing them to my students.
Academic publishing is a two way street the academics and journals need each other because "publish or perish" applies to both sides.
Academics do need to publish but they don't need commercial publishers to do so. There's a growing movement against the traditional journals - fuelled by the extortionate fees required for electronic access to the catalogue of publishing houses such as Springer and Elsevier. Universities cannot function without access to the literature, and this is your tax dollar being funnelled into the pockets of the publishers.
sure subscriptions are expensive compared to (say) people magazine, but if you pay peanuts for a job then only monkeys are going to apply.
The people who do the hard work behind the articles - the authors, the editors, and the reviewers - are not paid by the journal. The only paid staff are administrators and copy editors, who in my experience introduce more mistakes into the text than they correct.
Agreed, many trivial things can be handled by email. However, if you need to discuss or explain anything that's even slightly difficult conceptually then nothing beats an interactive conversation. It's much easier to gauge how much your peer understands (and therefore background you need to cover) when you can interact in real time.
A320 crashes http://www.airdisaster.com/cgi-bin/view_details.cgi?date=03221998®=RP-C3222&airline=Philippine+Airlines
The aircraft overran runway 4 while landing. A malfunction of the onboard flight computers prevented power from being reduced to idle, which inhibited thrust reverse and spoilers from being used. The offending engine was shut down, and brakes applied, but the aircraft was unable to stop before the end of the runway
I couldn't read the article you referenced ("server not responding"), but the accident report states this was caused by pilot error, not malfunctioning computers. From Wikipedia: "A selection by the pilot of the wrong mode on the onboard flight computers prevented power from being reduced to idle, which inhibited thrust reverse and spoilers from being used. The offending engine was shut down, and brakes applied, but the aircraft was unable to stop before the end of the runway".
There's a surprising number of people who believe that the high level of automation on Airbus is intrinsically more dangerous, but the figures show that the Airbus A320 is the safest narrow-body jet you can fly on. It's true that automated stuff can go wrong, but this can be more than compensated for by the ways it makes flying (or driving) safer.
3.3kW/h is TONS to keep any reasonably insulated house scorching hot
I assume you mean 3.3 kW, but I would be interested to see a source stating that this is enough to maintain a 20 degree (Celsius) difference between inside and outside for any "reasonably insulated" house.
And for the love of god you moron you didn't say "10kW/h" you said "10 kW/h for 10 hours"
There's no need to be offensive. It was you who said "10kW/h for 10 hours". You're using kW as a measure of energy, and kW/h as a measure of power. This is incorrect, as I've pointed out. "kW/h" would be a unit at the rate at which power changes.
How dumb are you?
Smart enough to know the difference between kW, kWh and kW/h. Here's a reference which may help you understand.
Firstly, according to your figures above (where you say 100 kWh is equivalent to driving 180 miles), 60 miles is equivalent to 33 kWh. This corresponds to 10 hours at 3.3 kW. 3.3kW is not enough to keep a house warm in subzero temperatures unless it has no windows.
Secondly, 10 kW/h means '10 kilowatts per hour', which has no physical meaning whatever. A Watt is a unit of power, not energy.
Note that kW is a unit of power, kWh is a unit of energy
Your figures weaken the support for your argument (that staying at home is greener than driving to work) significantly: they suggest it's greener to drive to work as long as it's less than 90 miles away from your home.
I disagree with your calculation, but the point remains broadly the same. Here's an alternative calculation. According to Wikipedia, 1 US gallon of gasoline is equivalent to 33 kWh, so 100kWh is equivalent to 3 gallons US. At 45 mpg US you can go 135 miles, not 180.
You can look at this the other way round. Driving 60 miles at 45mpg US is equivalent to 44kWh of energy usage. For it to be worth staying at home for 10 hours, you would have be able to keep your home warm with less than 4.4 kW. That's roughly the power output of two old-fashioned electric bar heaters. Do you get freezing temperatures where you live?
It's difficult to be more precise because the figures depend on lots of things - how cold is the place you live, how efficient is your car, how you define "equivalence" between gasoline and other energy forms - but you've illustrated the following point very well: most people wrongly assume that the energy required to heat their home is negligible relative to the cost of driving anywhere.
How much energy does it actually take to heat your house... because it's not 10kW/h.
"Kw/h" is not a unit of either energy or power. However, I can confirm that 10kW is approximately the power needed to keep my house comfortable in the winter. I know this because I know the ratings of the radiators in the house at 60 celcius, and I also know that they need to be kept at close to 60 celcius more or less constantly.
Perhaps some numbers would be helpful here.
We have a small semi-detached house with cavity wall insulation, loft insulation, and double glazing. We still need over 10kW to heat in the cold months. If we're out of the house for 10 hours, we save roughly 100 kW hours (*). How far will your car go on that amount of energy?
(*) Of course, it's not that simple because we have to use extra heat to bring the house back to normal temperature when we get home. A more accurate analysis would compare the temperature-time graphs for the two scenarios and use Newton's law of cooling. Nevertheless, the above figures are roughly correct.
Get used to it people. It's a green initiative.
No necessarily. If you don't work at home, you also don't have to keep your home warm (or cool) enough to be comfortable during the day. The office, on the other hand, will be kept at a reasonable temperature whether you're there or not.
My wife's work is about 30 miles away, but she works from home most days. We calculated that, on the coldest winter days, the carbon cost of driving to work was about the same as the extra heating that would be needed if she stayed at home. If you have a shorter commute, or have a greener method of transport than driving a car, it's quite likely that it's greener to work in the office than to work at home.
they eat corn alright, as does most of the world, in the form of processed food. You find corn derivatives in a bewildering varieties of industrial foods.
You imply that the French, and indeed the rest of the world, eat significant amount of processed food. It's difficult to get hard data on this, but my impression (from having lived there) is that processed food is a much smaller part of their diet than in the US. This article says that Americans eat rather more processed food than other countries, but it's difficult to compare because "baked goods" and "ready-to-eat" in the US and in France are rather different.
On the other hand, "most of the world" is certainly not eating significant amounts of industrial food - in China and India it's almost unheard of.
On Airbus vehicles, if the avionics computers crash, the airplane crashes. There's exactly ZERO way to pilot the computer manually in such a failure.
Completely untrue. When the avionics 'crash', the flight system progresses through 'alternate' to 'direct' law where the pilot has direct control of the plane.
Moreover, the avionics system can and does overrule pilot input. So if you get sensor malfunctions like this, even if the pilot is trying desperately to save the plane, the computer can still crash you.
Have a look at the statistics (pages maintained by a pro-Boeing pilot, by the way) and you'll see (i) for all your hysterical fear of Airbus aircraft, the fly-by-wire Airbus aircraft (i.e. all except A300 and A310) are just as safe as their Boeing counterparts (ii) there are no examples of an Airbus crash caused by the computer overriding the will of the pilot.
I see your point, but if x1 and (x1+x2)/2 make very different predictions then you don't expect them both to be equally good at describing a data set generated by x1 - unless (i) your data are less informative than your prior, or (ii) unless your predictions are in a regime where the parameters are identifiable but the training data are in a regime where they are not. I admit that this might often be the case when trying to modelling real systems...
Sure, the posterior mean or maximum might be very different from the true parameter value, but the the true value should sit somewhere in the full posterior distribution. If parameters have non-identifiability issues then the posterior should be very flat, but if you base your predictions on the posterior then this will show up in the distribution of your predictions. I would have thought this would only lead to a bias if your prior were TOO informative, wouldn't it?
The phenomenon this guy has observed is nothing to do with chaos theory, as several posters think, but rather to do with error propagation and model uncertainty. This is an issue whether the model is chaotic or not. His mistake is to think that calibration has to choose a single set of parameters, and then one has to make a single prediction from the model. Statistical methods can take into account many sources of uncertainty, including the range of parameters that could have produced the original data and intrinsic stochasticity in the model. The best way to do this is using Bayesian techniques.
You're still limited by how realistic your model is, and this is likely to be the real problem with economic models. However, Carter's argument (that it's fundamentally impossible to fit a model to itself and then make consistent predictions) is wrong.
Researchers and peer reviewers are not paid for their work but academic publishers have said such a business model is required to maintain quality.
The publishers are lying here to protect their cash cow. What maintains quality is the peer review system (which the journals do not pay for). The transfer of copyright to the publisher allows them to hold Universities to ransom - universities cannot function without access to the literature (present and past), and the costs of online access to journals have been spiraling over the past few years at a time when the publishers' actual costs are going down. After all, they don't pay for the research to be carried out, nor do they pay the academic editors or the reviewers, nor do they even need to typeset the document now that everyone submits a machine-readable copy.
A number with an infinite amount of decimal places cannot be defined.
On the contrary, it can easily be defined.
thus does not exist.
No numbers actually `exist' - they're all an abstraction, even the integers.