Then it is now. For that matter, that kind of data rate is going to seriously screw with many existing LAN and cluster fabrics - very few are designed to support that kind of on-the-wire rate. You'd need awe-inspiring hardware filtering and buffering to be able to convert between speeds on one side and speeds on the other. (The value would be that you could build one hell of a "fat tree" network if a single fibre is enough to guarantee that the total bandwidth of 20 downstream nodes is equal to the total upstream bandwidth.)
That's 50 terabits per second, multiplexed. Since they're using older, clunkier technology to handle the lasers than this chip, the throughput that is technically possible will logically be much greater -- with the system for doing so being smaller and more energy-efficient.
Then the cops are doing the wrong job. The CORRECT job description for a cop is to ensure that society (and, by extension, the individuals within it) suffer least harm on aggregate. Clearly, it is impossible to follow a least-harm doctrine if a cop is only trying to maximize their "clean-up" rate by arresting the innocent and allowing the guilty to go free.
ANY police force, doesn't matter where, doesn't matter when, that follows a doctrine of "look good" rather than "DO good" is a police force society is better off without. The problem is never with authority or government, the problem is with ILLUSION of authority and ILLUSION of government. The failure to tell reality from illusion is why corruption exists at all.
Being hopeful means there is a non-zero probability of being right. Being hopeful but skeptical means that probability rises to above 50% (since skepticism guarantees eliminating any solution that is certain to be flawed). Being a cynic GUARANTEES a 100% probability of being wrong. I'd rather be a hopeful skeptic than a moronic cynic.
There has never been a pure democratic, communist OR socialist government, so I can be 100% certain you're ignorant on the qualities of any of them.
The BBC is, in effect, an international body (since it has federalized and has branches in many lands, strange and otherwise, and it is not under the control of any nation - UK included) that DOES accomplish things AND is less politically motivated than the US government.
UNESCO is certainly an international body, it protects heritage in many countries (so is definitely accomplishing things) and it has shown a willingness to completely ignore the views of ANY nation - US included - that violate the purpose of the body, no matter what the cost to itself.
Generically, I am collecting scans of photos and negatives ranging from 1860 to the modern day for (currently five) strands of the family tree, which I am attempting to organize by date and location. That way, I can see how people, places, culture/society change with place and time, in addition to having as complete as possible conventional pictorial biography of everyone in the family tree. It is a nightmare to organize (most of the negative packets have little or no labeling) or index (databases are not good at searching blobs), let alone store.
I'm having to scan the negatives at very high resolution (12000 dpi) because some are extremely old and show signs of decay, which means I am -also- essentially making electronic backup copies of the negatives. I can't simply scan at a "sensible" visual resolution and then re-scan anything of interest at a higher resolution later.
To simplify identifying people, I'm uploading very low-res copies of the older images onto Google+ photos and passing links round the family, but the permission system (and documentation) is painfully crude and I have not yet figured out exactly how to authorize specific individuals to be able to tag - and their face identifier is flaky at best.
If it hits, it wipes out the major cause of habitat destruction, global warming and talk radio. If it misses, the sales of tinfoil hats and doomsday billboards will restore the global economy.
I have an extremely good offline backup. Mind you, I've just filled 100 DVDs to capacity and expect to fill 1000 more. No, not from downloads. Family history project. A big family history project. A big and extremely EXPENSIVE family history project.
And the hard disk space I'm using isn't even a fraction of the capacity of a modern hard drive.
Offline backup, these days, is getting very difficult to justify.
Nationalism is almost immaterial. It is far more efficient to have a multi-tier topology in government (local stuff can be handled locally, national stuff can be handled nationally, international stuff can be handled internationally). I'd rather the boundaries be drawn according to cultures rather than according to 20th century national identities, since cultures tend to reflect the needs of that region, but something is better than nothing. Usually.
For each type of network, there's a latency per hop and a (longer) latency at the endpoints for packet processing. (Pingpong tests generally use zero payload, so will just give you latency per hop.)
At the hardware level, cheap telecos might connect up using T2 or T4 lines (these are teleco-to-teleco lines, you can't buy them). These will have very different latencies to inter-teleco SDSL or inter-teleco optic fibre.
The on-the-wire protocol also matters. There's going to be a big difference between using ATM, Frame Relay or SONET, for example. On top of the on-the-wire protocol, there are next-level protocols that also make a big difference - MPLS, for example.
Finally, there is the topology. Many cheap telecos use tree topologies (which minimize their costs but maximize your hop count), some use simple mesh topologies (typically shortens distances to about half those in a tree), a few might use more complex meshes (reduces hop counts still further) or other topologies.
The difference in performance overall between the best-case and worst-case can be two or three orders of magnitude. For moderate improvements in performance, the difference in cost will be a fraction of the difference in improvement in service. For major improvements in improvements in performance, you will pay through the nose.
There is no -theoretical- reason why you could not achieve 8ms per hop latencies. The WAN hardware for this exists. There is no -theoretical- reason why backbone providers couldn't handle 10 terabit/second rates per fibre - again the hardware exists. There is no -theoretical- reason why you could not get 1 gigabit to the home - it's already done in some cities in some countries. In practice, private ISPs are run for the benefit of their shareholders, NOT for the benefit of the consumer. They will deliver only what they absolutely have to in order to keep the profits up.
It would be better if you also used pchar (a traceroute variant that reports effective bandwidth and percent packet loss on any given hop) but it is no longer maintained and the patches to the current Linux networking stack - due to go in - never did. You have to hack the flags. Having said that, most ISPs have never heard of the pathchar family of programs. Fools.
Your mission, should you choose to accept it, is to FIND a non-US TLD......and, then, find which one of those is outside of the control of ICANN as the root domain...
A quasi-independent international body is perfectly capable of handling disputes. Just because an international body is not a national body doesn't mean it's incapable of making decisions or making rulings.
The really annoying thing is that many, many technical gurus were against international control of the Internet and for US control of it in order to "protect the freedom". Yeah. Right. The rest of us have been posting for a long time now that all this was to be expected and that, whatever the faults of the UN and other international bodies, they're still better at safeguarding things like freedom.
ICANN should be dismantled completely and all control transferred to a quasi-independent international body. THAT is how to protect freedom.
Though I sincerely doubt that we'll see a single apology from those who protested against such a move the last time for what is ultimately their fault. Yes. Theirs. They are responsible for their actions and their actions include pressuring the US Govt to retain the power of dictatorship over the Internet. They owe, big time, to the rest of us for the inevitable and obvious consequences.
That was the beauty of IPSS (international packet switch stream) - you dialed a local PAD and could then connect to an X.25-to-IP relay (NSFNet ran several). Much cheaper than long-distance calls, although still not as cheap as people would have liked. British Telecom, especially, gouged heavily. This resulted in several online gamers later being arrested for theft and corruption in order to pay the bills they'd rack up.
A lot (ok, maybe in the order of a few thousand rather than a few million) of Europeans were using IPSS for gaming in the 80s. Lots of online MMORGs were available to those who knew either the IP addresses or PADs. That's where the big market was at the time. Richard Bartle was a household name at the time - the John Carmack of his era. Everything from "Commodore and Vegetable Games" to "PCW" and "Practical Computing" were covering the online gaming industry at the time.
Oh, gopher is still around. You see the clients being listed on Freshmeat/Freecode from time to time. Alas, Archie seems to have died. It truly wasn't evil, unlike the search engines of today. WAIS was another early system that has since met the Great Bit Bucket in the Sky. FSP (an alternative to FTP) never really made it big, which always surprised me, but is also occasionally still maintained.
Telnet and FTP were the dominant protocols of way-back, yes, though Finger was occasionally in use. Less so after the Great Internet Worm fiasco, as admins were confronted with the choice of either disabling the service or (gasp!) having to apply security measures. However, they weren't the only protocols. NNTP was around (Usenet - a wonderland of intelligent conversation until AOL discovered it). Clients using Sun's RPC protocol were thinner on the ground, but they did exist for some applications.
A number of European nations (such as France) opted to have a private information network instead, so you saw networks spring up that allowed train schedules to be looked up, tickets to concerts to be bought, etc. eCommerce started a loooong time before CERN's webserver was written. Those were not Internet, though. They ran through teletext operations or via a single central service. Not a million miles from the "network computer" concept Sun was to later promote.
The Internet was publicly available in the 1980s. Dunno about grandparent post, but anyone in the 80s was quite capable of dialing into the Internet via an IPSS-to-IP gateway (NSFNet offered several that were free). Indeed, the modern Internet in Europe -is- largely just IPSS rebadged, so one could argue that that alone was Internet.
However, all evidence is that people want cheap rather than good, so I don't expect these to come on market any time soon. Besides, those with no hearing will complain (as indeed I note they already have) that they can't hear the difference, actual measurements notwithstanding.
Having said that, it would be interesting to see what material science can do to work on the concept. Spider silk is good but fragile - making it suitable for a single performance at The Proms at the Albert Hall but not really useful for musicians in general. A tougher synthetic version is definitely needed, along with other strings of other materials with other sounds.
Ideally, there would be a single model you could produce where you could feed in the material properties and then synthesize the tonal properties. People shopping for strings could then choose by hearing rather than by popular opinion.
I know a witch-doctor (well, several), and I think his cure for one person's hangover would probably be different to his cure for another person's, depending on the case. For your study, you'd want all the treatments to be the same, right? I don't think it's going to happen, not when it is so personal. So I guess we're not going to get the straightforward proof you're looking for. But whose fault is that? If you require very strong independent standards of objective proof, you're excluding all treatments which might work but are difficult to study or analyse.
Well, that depends on what you'd want to study. I tend to be of the school of thought that you want like treatments for like causes, not necessarily like treatments for like effects. Thus, if there are N ways to get a headache, I would expect N treatments. In other words, I would expect that if I looked at the underlying cause (which will presumably be electrical, chemical and/or blood pressure related, where any possible magnitude and permutation of those is entirely valid) that if two people had absolutely identical underlying causes that the treatment would always be absolutely identical (ie: the treatment isn't random) and that if I monitored that underlying cause, there would be a change in state between before treatment and after treatment.
This isn't too difficult to study, merely time-consuming. The law of large numbers only works if, well, you've large numbers. In science, you'll sometimes see people talk about the "confidence level" or the "sigma level" of a result. The systems are slightly different but they boil down to the same thing - the odds that what you are getting isn't a real result but just pure statistical fluke.
The more numbers you have, the greater your confidence that what you are seeing is real. Now, my way of thinking is absurdly simple - if you can demonstrate a high level of confidence that f(given treatment, given underlying cause) = (given underlying effect) and f(!given treatment, given underlying case) = (!given underlying effect) then you have established that the treatment - when applied to that cause - produces that effect. How it does so is a matter for theorists.
Now, I'm very careful to state "underlying cause" and "underlying effect" because there can be zero correlation between what underlies a symptom and the symptom itself. That is entirely valid -- and rather common. The greater the distance between what is actually happening and what is actually being observed, the greater the chance of you missing whatever is actually going on. True, you can't always make direct observations of the underlying cause, so you do the best you can and minimize the gap. You can safely ignore any study where the gap is ignored or even increased because you know for a fact that the researchers cannot have considered the uncertainties introduced and therefore cannot have an accurate idea of how confident you can be in the results. If you don't know how good the results are, what use are they?
Now to go back to your other point:
But just to point out that human perception is all that we've got. It is our only direct input, despite its flaws. Everything else is second- or third-hand.
To some extent, I've covered this in discussing the level of directness (or lack thereof). However, it is worth examining this point a little closer. Yes, as Descartes (and indeed R. D. Laing) noted, human perception is indeed all that we've got. Furthermore, indirectness (as noted) increases the number of places errors can be introduced. As a result, you want to keep things as direct as possible.
Equally, though, quantitative data can be more precisely compared than qualitative data. When comparing qualities, you introduce all kinds of other types of error. Unquantifiable kinds. It's another reason I regard symptom-based treatments
Human perception is never good enough for anything. Humans perceive things that are unreal all the time (see The Guardian's experiment on false memories that they recently ran as an example).
Anything that exists can be measured by a physical instrument. Not necessarily at the time that you deduce that it exists (particle physics frequently deduces the existence of particles long before direct observation becomes possible) but even then you can firmly establish the constraints on that physical instrument and thus prove that such an instrument must be buildable.
Unlike some, I am not a skeptic of "alternative medicine" merely because it is alternative. I accept entirely the premise that there may be alternative approaches to medicine that are superior to conventional Western medicines. If a witch-doctor could demonstrate repeatable cures for hangovers then I would want to know how any why, but I wouldn't deny evidence merely because of the label on the packet. However, I absolutely require that such approaches be shown with the same scientific rigour that I would expect of any other kind of medicine (or, indeed, any other kind of phenomenon). Equally, I am highly skeptical of Western medicine where that scientific rigour is absent or dubious. There can be only one standard and there can be no excuses made for any industry, be it mainstream or traditional.
(Equally, in surveys of scientists where a high percentage of fraud is claimed, no excuses can be made for scientists who commit fraud. Fraud is fraud, whether carried out by bankers, Harley Street doctors, scientists or the local shaman. I don't care who, I don't care why.)
My chief complaint for those who pick sides rather than pick standards is that they are not looking at the evidence, no matter what side they pick. Science isn't about "home team" vs. "the other guys", science is about methodology, rigour and standards of evidence. (For those who give a rat's arse, I apply that to all things. I don't like bands, I like songs. I don't like countries, I like factors. I don't like authors, I like stories. I don't give a damn about the source, I care about the results. The source doesn't enter into it. Ever.)
I can find nothing that indicates when the Travois was invented. Want to give some data on that?
(Our evidence for the age of the wheel is limited to a "latest", not an earliest, so the summary and article are dubious. We know that the age of the words "wheel" and "axle" are 5,500 years old, but we don't know that the invention itself was named at the time of invention, nor do we know that the invention was by a proto-Indo-European society. It is entirely plausible to imagine that the invention was earlier by a culture that never transmitted the technology and was invaded by a proto-Indo-European group. I'm not arguing that was the case, merely that the author's rebuttal in the book mentioned in the summary of that idea did not sound sufficient to me.)
We don't know what the Greeks did or did not have in the way of mathematics or logic. Ever since the Archimedes Palimpsest showed that most conventional wisdom on what the Greeks knew was wrong, it has been clear that assumptions about how they derived their results are flawed.
We do know, however, that Plato's Republic was very much data-driven. His writings were not theoretical but based on actual observations of actual political systems, where they failed in practice and how they evolved in practice. That is not philosophy, that is empirical science much as we practice it today (drawing up falsifiable hypotheses based on available data).
The easy solution is to raise the level of education (and thus understanding) of everyone. You still produce median leaders, but the median is now better than it was. So long as the median is always kept at or above the level required to tackle the abstract risks of the day, you will win.
Slashdot Mirror
Then it is now. For that matter, that kind of data rate is going to seriously screw with many existing LAN and cluster fabrics - very few are designed to support that kind of on-the-wire rate. You'd need awe-inspiring hardware filtering and buffering to be able to convert between speeds on one side and speeds on the other. (The value would be that you could build one hell of a "fat tree" network if a single fibre is enough to guarantee that the total bandwidth of 20 downstream nodes is equal to the total upstream bandwidth.)
Combine that with this story - a single channel rate of 400 Gbps (512 Gbps including error-correction) over very long distances. (Wikipedia says 128 channels per fibre is available.)
http://www.bbc.co.uk/news/technology-17271797
That's 50 terabits per second, multiplexed. Since they're using older, clunkier technology to handle the lasers than this chip, the throughput that is technically possible will logically be much greater -- with the system for doing so being smaller and more energy-efficient.
Do you know how rare *good* bug reports are?
Then the cops are doing the wrong job. The CORRECT job description for a cop is to ensure that society (and, by extension, the individuals within it) suffer least harm on aggregate. Clearly, it is impossible to follow a least-harm doctrine if a cop is only trying to maximize their "clean-up" rate by arresting the innocent and allowing the guilty to go free.
ANY police force, doesn't matter where, doesn't matter when, that follows a doctrine of "look good" rather than "DO good" is a police force society is better off without. The problem is never with authority or government, the problem is with ILLUSION of authority and ILLUSION of government. The failure to tell reality from illusion is why corruption exists at all.
It can't hurt to try these. It might well reduce accidents - it would certainly reduce stupid court cases.
Being hopeful means there is a non-zero probability of being right. Being hopeful but skeptical means that probability rises to above 50% (since skepticism guarantees eliminating any solution that is certain to be flawed). Being a cynic GUARANTEES a 100% probability of being wrong. I'd rather be a hopeful skeptic than a moronic cynic.
There has never been a pure democratic, communist OR socialist government, so I can be 100% certain you're ignorant on the qualities of any of them.
The BBC is, in effect, an international body (since it has federalized and has branches in many lands, strange and otherwise, and it is not under the control of any nation - UK included) that DOES accomplish things AND is less politically motivated than the US government.
UNESCO is certainly an international body, it protects heritage in many countries (so is definitely accomplishing things) and it has shown a willingness to completely ignore the views of ANY nation - US included - that violate the purpose of the body, no matter what the cost to itself.
Generically, I am collecting scans of photos and negatives ranging from 1860 to the modern day for (currently five) strands of the family tree, which I am attempting to organize by date and location. That way, I can see how people, places, culture/society change with place and time, in addition to having as complete as possible conventional pictorial biography of everyone in the family tree. It is a nightmare to organize (most of the negative packets have little or no labeling) or index (databases are not good at searching blobs), let alone store.
I'm having to scan the negatives at very high resolution (12000 dpi) because some are extremely old and show signs of decay, which means I am -also- essentially making electronic backup copies of the negatives. I can't simply scan at a "sensible" visual resolution and then re-scan anything of interest at a higher resolution later.
To simplify identifying people, I'm uploading very low-res copies of the older images onto Google+ photos and passing links round the family, but the permission system (and documentation) is painfully crude and I have not yet figured out exactly how to authorize specific individuals to be able to tag - and their face identifier is flaky at best.
If it hits, it wipes out the major cause of habitat destruction, global warming and talk radio. If it misses, the sales of tinfoil hats and doomsday billboards will restore the global economy.
I have an extremely good offline backup. Mind you, I've just filled 100 DVDs to capacity and expect to fill 1000 more. No, not from downloads. Family history project. A big family history project. A big and extremely EXPENSIVE family history project.
And the hard disk space I'm using isn't even a fraction of the capacity of a modern hard drive.
Offline backup, these days, is getting very difficult to justify.
Nationalism is almost immaterial. It is far more efficient to have a multi-tier topology in government (local stuff can be handled locally, national stuff can be handled nationally, international stuff can be handled internationally). I'd rather the boundaries be drawn according to cultures rather than according to 20th century national identities, since cultures tend to reflect the needs of that region, but something is better than nothing. Usually.
For each type of network, there's a latency per hop and a (longer) latency at the endpoints for packet processing. (Pingpong tests generally use zero payload, so will just give you latency per hop.)
At the hardware level, cheap telecos might connect up using T2 or T4 lines (these are teleco-to-teleco lines, you can't buy them). These will have very different latencies to inter-teleco SDSL or inter-teleco optic fibre.
The on-the-wire protocol also matters. There's going to be a big difference between using ATM, Frame Relay or SONET, for example. On top of the on-the-wire protocol, there are next-level protocols that also make a big difference - MPLS, for example.
Finally, there is the topology. Many cheap telecos use tree topologies (which minimize their costs but maximize your hop count), some use simple mesh topologies (typically shortens distances to about half those in a tree), a few might use more complex meshes (reduces hop counts still further) or other topologies.
The difference in performance overall between the best-case and worst-case can be two or three orders of magnitude. For moderate improvements in performance, the difference in cost will be a fraction of the difference in improvement in service. For major improvements in improvements in performance, you will pay through the nose.
There is no -theoretical- reason why you could not achieve 8ms per hop latencies. The WAN hardware for this exists. There is no -theoretical- reason why backbone providers couldn't handle 10 terabit/second rates per fibre - again the hardware exists. There is no -theoretical- reason why you could not get 1 gigabit to the home - it's already done in some cities in some countries. In practice, private ISPs are run for the benefit of their shareholders, NOT for the benefit of the consumer. They will deliver only what they absolutely have to in order to keep the profits up.
It would be better if you also used pchar (a traceroute variant that reports effective bandwidth and percent packet loss on any given hop) but it is no longer maintained and the patches to the current Linux networking stack - due to go in - never did. You have to hack the flags. Having said that, most ISPs have never heard of the pathchar family of programs. Fools.
Your mission, should you choose to accept it, is to FIND a non-US TLD... ...and, then, find which one of those is outside of the control of ICANN as the root domain...
(Cue Mission: Impossible theme)
A quasi-independent international body is perfectly capable of handling disputes. Just because an international body is not a national body doesn't mean it's incapable of making decisions or making rulings.
The really annoying thing is that many, many technical gurus were against international control of the Internet and for US control of it in order to "protect the freedom". Yeah. Right. The rest of us have been posting for a long time now that all this was to be expected and that, whatever the faults of the UN and other international bodies, they're still better at safeguarding things like freedom.
ICANN should be dismantled completely and all control transferred to a quasi-independent international body. THAT is how to protect freedom.
Though I sincerely doubt that we'll see a single apology from those who protested against such a move the last time for what is ultimately their fault. Yes. Theirs. They are responsible for their actions and their actions include pressuring the US Govt to retain the power of dictatorship over the Internet. They owe, big time, to the rest of us for the inevitable and obvious consequences.
That was the beauty of IPSS (international packet switch stream) - you dialed a local PAD and could then connect to an X.25-to-IP relay (NSFNet ran several). Much cheaper than long-distance calls, although still not as cheap as people would have liked. British Telecom, especially, gouged heavily. This resulted in several online gamers later being arrested for theft and corruption in order to pay the bills they'd rack up.
A lot (ok, maybe in the order of a few thousand rather than a few million) of Europeans were using IPSS for gaming in the 80s. Lots of online MMORGs were available to those who knew either the IP addresses or PADs. That's where the big market was at the time. Richard Bartle was a household name at the time - the John Carmack of his era. Everything from "Commodore and Vegetable Games" to "PCW" and "Practical Computing" were covering the online gaming industry at the time.
Oh, gopher is still around. You see the clients being listed on Freshmeat/Freecode from time to time. Alas, Archie seems to have died. It truly wasn't evil, unlike the search engines of today. WAIS was another early system that has since met the Great Bit Bucket in the Sky. FSP (an alternative to FTP) never really made it big, which always surprised me, but is also occasionally still maintained.
Telnet and FTP were the dominant protocols of way-back, yes, though Finger was occasionally in use. Less so after the Great Internet Worm fiasco, as admins were confronted with the choice of either disabling the service or (gasp!) having to apply security measures. However, they weren't the only protocols. NNTP was around (Usenet - a wonderland of intelligent conversation until AOL discovered it). Clients using Sun's RPC protocol were thinner on the ground, but they did exist for some applications.
A number of European nations (such as France) opted to have a private information network instead, so you saw networks spring up that allowed train schedules to be looked up, tickets to concerts to be bought, etc. eCommerce started a loooong time before CERN's webserver was written. Those were not Internet, though. They ran through teletext operations or via a single central service. Not a million miles from the "network computer" concept Sun was to later promote.
The Internet was publicly available in the 1980s. Dunno about grandparent post, but anyone in the 80s was quite capable of dialing into the Internet via an IPSS-to-IP gateway (NSFNet offered several that were free). Indeed, the modern Internet in Europe -is- largely just IPSS rebadged, so one could argue that that alone was Internet.
However, all evidence is that people want cheap rather than good, so I don't expect these to come on market any time soon. Besides, those with no hearing will complain (as indeed I note they already have) that they can't hear the difference, actual measurements notwithstanding.
Having said that, it would be interesting to see what material science can do to work on the concept. Spider silk is good but fragile - making it suitable for a single performance at The Proms at the Albert Hall but not really useful for musicians in general. A tougher synthetic version is definitely needed, along with other strings of other materials with other sounds.
Ideally, there would be a single model you could produce where you could feed in the material properties and then synthesize the tonal properties. People shopping for strings could then choose by hearing rather than by popular opinion.
Well, that depends on what you'd want to study. I tend to be of the school of thought that you want like treatments for like causes, not necessarily like treatments for like effects. Thus, if there are N ways to get a headache, I would expect N treatments. In other words, I would expect that if I looked at the underlying cause (which will presumably be electrical, chemical and/or blood pressure related, where any possible magnitude and permutation of those is entirely valid) that if two people had absolutely identical underlying causes that the treatment would always be absolutely identical (ie: the treatment isn't random) and that if I monitored that underlying cause, there would be a change in state between before treatment and after treatment.
This isn't too difficult to study, merely time-consuming. The law of large numbers only works if, well, you've large numbers. In science, you'll sometimes see people talk about the "confidence level" or the "sigma level" of a result. The systems are slightly different but they boil down to the same thing - the odds that what you are getting isn't a real result but just pure statistical fluke.
http://en.wikipedia.org/wiki/Standard_deviation
The more numbers you have, the greater your confidence that what you are seeing is real. Now, my way of thinking is absurdly simple - if you can demonstrate a high level of confidence that f(given treatment, given underlying cause) = (given underlying effect) and f(!given treatment, given underlying case) = (!given underlying effect) then you have established that the treatment - when applied to that cause - produces that effect. How it does so is a matter for theorists.
Now, I'm very careful to state "underlying cause" and "underlying effect" because there can be zero correlation between what underlies a symptom and the symptom itself. That is entirely valid -- and rather common. The greater the distance between what is actually happening and what is actually being observed, the greater the chance of you missing whatever is actually going on. True, you can't always make direct observations of the underlying cause, so you do the best you can and minimize the gap. You can safely ignore any study where the gap is ignored or even increased because you know for a fact that the researchers cannot have considered the uncertainties introduced and therefore cannot have an accurate idea of how confident you can be in the results. If you don't know how good the results are, what use are they?
Now to go back to your other point:
To some extent, I've covered this in discussing the level of directness (or lack thereof). However, it is worth examining this point a little closer. Yes, as Descartes (and indeed R. D. Laing) noted, human perception is indeed all that we've got. Furthermore, indirectness (as noted) increases the number of places errors can be introduced. As a result, you want to keep things as direct as possible.
Equally, though, quantitative data can be more precisely compared than qualitative data. When comparing qualities, you introduce all kinds of other types of error. Unquantifiable kinds. It's another reason I regard symptom-based treatments
Human perception is never good enough for anything. Humans perceive things that are unreal all the time (see The Guardian's experiment on false memories that they recently ran as an example).
Anything that exists can be measured by a physical instrument. Not necessarily at the time that you deduce that it exists (particle physics frequently deduces the existence of particles long before direct observation becomes possible) but even then you can firmly establish the constraints on that physical instrument and thus prove that such an instrument must be buildable.
Unlike some, I am not a skeptic of "alternative medicine" merely because it is alternative. I accept entirely the premise that there may be alternative approaches to medicine that are superior to conventional Western medicines. If a witch-doctor could demonstrate repeatable cures for hangovers then I would want to know how any why, but I wouldn't deny evidence merely because of the label on the packet. However, I absolutely require that such approaches be shown with the same scientific rigour that I would expect of any other kind of medicine (or, indeed, any other kind of phenomenon). Equally, I am highly skeptical of Western medicine where that scientific rigour is absent or dubious. There can be only one standard and there can be no excuses made for any industry, be it mainstream or traditional.
(Equally, in surveys of scientists where a high percentage of fraud is claimed, no excuses can be made for scientists who commit fraud. Fraud is fraud, whether carried out by bankers, Harley Street doctors, scientists or the local shaman. I don't care who, I don't care why.)
My chief complaint for those who pick sides rather than pick standards is that they are not looking at the evidence, no matter what side they pick. Science isn't about "home team" vs. "the other guys", science is about methodology, rigour and standards of evidence. (For those who give a rat's arse, I apply that to all things. I don't like bands, I like songs. I don't like countries, I like factors. I don't like authors, I like stories. I don't give a damn about the source, I care about the results. The source doesn't enter into it. Ever.)
I can find nothing that indicates when the Travois was invented. Want to give some data on that?
(Our evidence for the age of the wheel is limited to a "latest", not an earliest, so the summary and article are dubious. We know that the age of the words "wheel" and "axle" are 5,500 years old, but we don't know that the invention itself was named at the time of invention, nor do we know that the invention was by a proto-Indo-European society. It is entirely plausible to imagine that the invention was earlier by a culture that never transmitted the technology and was invaded by a proto-Indo-European group. I'm not arguing that was the case, merely that the author's rebuttal in the book mentioned in the summary of that idea did not sound sufficient to me.)
We don't know what the Greeks did or did not have in the way of mathematics or logic. Ever since the Archimedes Palimpsest showed that most conventional wisdom on what the Greeks knew was wrong, it has been clear that assumptions about how they derived their results are flawed.
We do know, however, that Plato's Republic was very much data-driven. His writings were not theoretical but based on actual observations of actual political systems, where they failed in practice and how they evolved in practice. That is not philosophy, that is empirical science much as we practice it today (drawing up falsifiable hypotheses based on available data).
Dictatorship is worse than democracy, for the same reason single-core vector computers are worse than clusters of multi-core computers.
The easy solution is to raise the level of education (and thus understanding) of everyone. You still produce median leaders, but the median is now better than it was. So long as the median is always kept at or above the level required to tackle the abstract risks of the day, you will win.
They're androids, so component works just fine.