Whoa there, the hole in the ozone layer WOULD have continued to expand (basically until there was no ozone left), but for a heroic international effort to dramatically reduce the amount of CFC's escaping into the atmosphere. Fortunately this was successful.
The problem with global warming is that by the time that the effect is big enough to be attributable to human intervention beyond any question of doubt (many decades away, probably) the damage by then would be so great and so hard to reverse that it would be an absolute catastrophe.
On other words, the potential economic effect of not buying into it are so huge that it might not even be just an economic effect anymore, it may end up as an extinction effect.
The reality is that current theories of climate, even though they may not be as rigorous as one would like, say that climate change is and will happen. Most of the argument is on how much and how soon, rather than if. Sure, there is a chance that the current theories are completely wrong and there will be no climate change at all - are you willing to bet your life (and the life of everything on the planet) on that though?
He is referring to methodology - producing rigorously tested data, versus anecdotes, belief in the divine authority of the shaman, etc, etc.
That doesn't mean that everything the shaman does is useless, it just provides a mechanism for distinguishing what is useful and what is not.
The vast majority of folklore that is incorrect too. The whole purpose of science is to find out which is correct and which isn't. The fact that people can use this knowledge to make theories about how things work, and thus invent new things, is really secondary to the process of science itself.
Compare
"I have tested this herbal remedy by reading thousands of ancient scrolls in which the ancient Egyptians expressed belief that this plant was given to them by Amon, and therefore has divine properties"
with
"I have tested this herbal remedy in a double-blind study of thousands of people and it is statistically likely that it has a desirable effect"
If you want to find know the properties of the herbal remedy, the first statement is not useful to you. But that does not mean that the herbal remedy doesn't work (although you should not be surprised if it does not).
There is at least one peer-reviewed journal article about this experiment, and probably the guy is at least under the impression that the result is correct.
That doesn't make it less of a scam though, it just means that there are some very trashy journals who have very low standards for referee selection. And, calling the author a "scientist" is probably a bit much of a stretch.
I even came across a theoretical paper demonstrating that the effect can exist, given the right conditions. The abstract stated boldly that the effect is a natural consequence of relativity plus some quantum effects, without stating that the model of gravity they were using was completely artificial, I think Euclidean space with massive gravitons and lots of other weirdness. (For the non-physicists, having massive gravitons means that gravity would decay exponentially fast, rather than the usual inverse-square law). Indeed, the paper served as a good demonstration that it is very unlikely that such an effect could happen in reality.
By the way, i don't think NASA gave him 1 million, I think rather they spent that trying to reproduce the experiment themselves. I vaguely remember a story that he was the only one who knew enough details to repeat the experiment, so NASA ended up getting him involved (which basically amounts to giving him the million, I guess).
The only other case that springs to mind of an experiment where only one person can "reproduce" it, is (ahem) Jan-Hendrik Schon.
X-Rays were discussed in the scientific literature long before there were any patents issued. Even then, patents only effect a specific device for generating X-rays, not the X-rays themselves (otherwise you had better kill yourself immediately - or pay me lots of money, those background X-rays passing through your body right now are in violation of my IP rights!)
Patents were intended to be for inventions, that is, a device that is physically realizable and practical to build. It really must be a condition of a patent application that you can demonstrate that the invention works, and performs as set out in the papent claim.
That the USPTO has prostituted itself to big money over the years doesn't change that intent at least, even if they broke the mechanism (hopefully tempoarily).
I agree that in principle, there is no harm (apart from waste of resources) in patenting impossible inventions. The problems come though when the holders of the patent use it as a club against other, related inventions that might actually work. So in practice it can be a very bad thing.
In no sense did Farnsworth work in isolation. The fact that the inventor of TV is a rather disputed claim, and there were a whole host of inventions that led up to it (Logie-Baird's work, for example) is enough evidence for that.
Just because someone works alone doesn't mean that they don't read the technical literature, attend conferences or exhibitions, or borrow ideas from other people.
The important factor is to make sure that enough samples are taken out of the entire population such that the variance in the result is "small enough".
If you want to test what colour my hair is you only need to take one sample because the population size for this experiment is 1.
If you want to test for bias in a set of 100 coin tosses, a sample size of 1 is certainly not adequate, even 10 samples (10% of the population) might not be enough, if the bias is small. But if the bias is big (say, all 10 samples comes up "heads"), its a fairly strong indicator. But, it is still a statistical estimate; you might just have been very unlucky, and a 1/1024 chance will happen, well, one time in 1024:-)
I would imagine that for such a small effect such as someone getting a mild disease and ending up obease (which surely happens in a some fraction of cases anyway, and provides no evidence for a causal relationship between the disease and obeasity), the required sample size for a significant result would be very big. After all, surely very many people have suffered a minor disease at some stage of their life. How many of them are obease?
I think what he means is not so much that all of the "small" science has been done, but rather the period of time where one very bright person could make a lot of progress in isolation.
I've sometimes heard this as "most of the easy science has already been done", but that is also not very well worded. What it means is that it is no longer possible to make fundamental discoveries (or even any discoveries at all, apart from personal ones) by rolling balls down an inclined plane. Instead, you need either a large scale machine, say a particle accelerator, or even a small scale machine, say a STM machine or similar. Both of these are quite complex devices, which one person could not build and operate by themselves.
But, I would argue the "easy" point though. In the 17th century, Galelleo's experiment was surely not "easy", although it is easy to reproduce with modern equipment.
Actually, that isn't true. The whole point of wave-particle duality is that a single photon will behave in a similar way to a wave - and any wave can be considered to be made up of one or more particles.
If you attempt to measure the polarization of a single photon but your detector is not at the correct angle, then you will get a probabilistic result. This is equally true for a single photon or many photons.
I guess your point though is that if the beam is polarized, then the photons will have consistent polarization so you can rotate the detector until you get 100% in the same direction. But most photon sources emit photons with random polarization.
That is only partially true. True, most (but by no means all) of Einstein's important papers had a single author, but by no means did he work in isolation. He frequently visited other scientists, gave lectures at other universities, attended conferences etc, exactly as scientists do today.
He certainly was not suppressed, at least not by the scientific community. When General Relativity was published, for example, many other people working in the field immediately realized the significance and jumped on it. IIRC, Minkowsky himself was only months away from figuring it out himself.
It is absolutely untrue that no one accepted the theories until the eclipse experiment (which, ironically, was later shown to to be "probably fraudulant" in fact! A great example of post-war anglo-german cooperation, publicised to the point that they HAD to get a result so much that it was ultimately faked). By that time, Einstein was already very famous, and relativity was by far the most likely contender for the theory of gravitation.
Of course, there were (and still are) competing theories. A major test of relativity occurred in the late 70s / early 80s when the rotational speed of a binary pulsar was measured to be slowing down, in accordance with the loss of energy from radiating gravitational waves, predicted by general relativity. At the time, there were several theories that reproduced the previously tested predictions of relativity but predicted a different rate of slowing down. The 1993 nobel prize was given for this in fact.
Sure, a quick glance in/usr/lib suggests that most libs have at least two versions.
But that is very different to having copy for every process, which is what the OP was suggesting. Currently ps shows ~150 processes currently loaded, I guess most of them are linked against libc. Using a different copy of the lib for each process would increase the amount of RAM I need by well over 100MB, and that is only for a single lib.
As he says, 'There is no scientific claim so preposterous that a scientist cannot be found to vouch for it'. What he doesn't say is that there are plenty more who will invest in it or base legislation on it."
From the article, the full paragraph of the quote is:
There is, alas, no scientific claim so preposterous that a scientist cannot be found to vouch for it. And many such claims end up in a court of law after they have cost some gullible person or corporation a lot of money. How are juries to evaluate them?
The very next sentence indicates that there are very many people who are willing to invest or base laws on bad science!
Consuder how many processes are loaded on a typical 'doze system. Now imagine what would happen if every one of them had a different copy of USER32.DLL (or whatever).
Well, on a typical system there might be a hundred or so processes. Add 4MB to all of them and its suddenly not insignificant anymore.
Plus, if two running processes are sharing a shared library, then a task swap doesn't completely blow away the cache.
Finally, I think with Windoze I think there is a mode in which DLL's can have global data, shared among all programs using it. Not sure though, its years since I did any windoze programming.
This will solve some problems where installing software blows away a specific version of some DLL that is required by some other piece of software, but at the expense of introducing hard to solve problems where two programs A and B expect to use the same DLL, but end up using different ones. Knowing M$, it will probably not be possible to manually change which version is linked in.
Where did slashdot claim that apple had a *software* patent on the trashcan? All I see is a link to an article on software patents that has some relevance because it talks about " the difference between utility and design patents".
Interestingly, from http://www.uspto.gov/web/offices/pac/design/defini tion.htm on the criteria for getting a design patent:... In addition, 35 U.S.C. 171 requires that a design to be patentable must be "original". Clearly a design which simulates a well known, or naturally occurring object or person is not original as required by the statute.
I find it hard to believe that a picture of a trash can can fall into the "original" category.
Well, infrastructure improvements can be great in the long term, as long as its the correct infrastructure.
Actually I don't know whether the infrastructure investments in Japan have been good, or maybe the increased basic research funding was just slop left over from masses of useless dam building projects;)
I'm now waiting to see what reaction the 'bankers' comment gets!
The GNU multiprecision library (GNUmp) is an excellent choice for this, very fast, although it is a bit cumbersome to use directly. I use a very simple C++ wrapper, but a more sophisticated one is now included in the library. The URL is in the story I believe.
Most non-numerical people use matlab, maple or mathematica when they want to solve a relatively simple problem. For bigger problems, in my experience it is easier and faster to just write (or get someone else to write) a C or C++ or Fortran program to do it, rather than use higher level tools, * unless the problem can be expressed particularly simply in mathematica and not in a general purpose programming language. I've only ever met one person who attempted running a mathematica problem on a supercomputer (although I know that people do this all the time): he basically ended up as an alcoholic dropout.
Only the old guys use Fortran.
Only misguided (IMO) people use C for anything big.
Libraries: I use GMP, BLAS (usually in the form of ATLAS), LAPACK, ARPACK, MPI, pthreads (OpenMP is usually a better choice, and nowdays has pretty good compiler support - unfortunately I wrote my main codes a couple of years to early). All of these are solid, reliable and fast. GSL looks good but I havn't used it.
Note though that none of the previous are actually C++ libraries. ARPACK++ is crap, most C++ linear algebra libraries are crap, the C++ bindings for MPI are crap, pretty much all existing scientific libraries in C++ are crap. However, some recent libraries have potential, such as the stuff on boost.org (eg, uBLAS), maybe MTL will end up being good but the existing release is not IMHO. Blitz++ is good, but the transition away from a single-developer project has been long, AFAICT it is still basically unmaintained. Pooma appears to be really solid and usable, but I've never gone beyond reading some of the web tutorials.
But, I practically never use a non-C++ library 'raw', I always put at least a minimal C++ wrapper. With careful use of templates, you can get some pretty nice looking (and efficient) code. This is especially important for Fortran. Apart from needing platform-dependent hacks to actually call Fortran functions, I havn't yet seen a Fortran interface that I actually like. Too many parameters, obscure, historical conventions, the dreaded "reverse communication interface" etc etc.
I also use xmgrace, recently figured out how to script it and that is a very powerful tool. I use Gnuplot for 3d plots, but I havn't really figured out how to work it yet.
I also use Perl a bit, but mostly for job control sort of things (copying files across before the start of an MPI calculation, submitting jobs to the queue, etc).
Slashdot Mirror
The problem with global warming is that by the time that the effect is big enough to be attributable to human intervention beyond any question of doubt (many decades away, probably) the damage by then would be so great and so hard to reverse that it would be an absolute catastrophe.
On other words, the potential economic effect of not buying into it are so huge that it might not even be just an economic effect anymore, it may end up as an extinction effect.
The reality is that current theories of climate, even though they may not be as rigorous as one would like, say that climate change is and will happen. Most of the argument is on how much and how soon, rather than if. Sure, there is a chance that the current theories are completely wrong and there will be no climate change at all - are you willing to bet your life (and the life of everything on the planet) on that though?
That doesn't mean that everything the shaman does is useless, it just provides a mechanism for distinguishing what is useful and what is not.
The vast majority of folklore that is incorrect too. The whole purpose of science is to find out which is correct and which isn't. The fact that people can use this knowledge to make theories about how things work, and thus invent new things, is really secondary to the process of science itself.
Compare
"I have tested this herbal remedy by reading thousands of ancient scrolls in which the ancient Egyptians expressed belief that this plant was given to them by Amon, and therefore has divine properties"
with
"I have tested this herbal remedy in a double-blind study of thousands of people and it is statistically likely that it has a desirable effect"
If you want to find know the properties of the herbal remedy, the first statement is not useful to you. But that does not mean that the herbal remedy doesn't work (although you should not be surprised if it does not).
That doesn't make it less of a scam though, it just means that there are some very trashy journals who have very low standards for referee selection. And, calling the author a "scientist" is probably a bit much of a stretch.
I even came across a theoretical paper demonstrating that the effect can exist, given the right conditions. The abstract stated boldly that the effect is a natural consequence of relativity plus some quantum effects, without stating that the model of gravity they were using was completely artificial, I think Euclidean space with massive gravitons and lots of other weirdness. (For the non-physicists, having massive gravitons means that gravity would decay exponentially fast, rather than the usual inverse-square law). Indeed, the paper served as a good demonstration that it is very unlikely that such an effect could happen in reality.
By the way, i don't think NASA gave him 1 million, I think rather they spent that trying to reproduce the experiment themselves. I vaguely remember a story that he was the only one who knew enough details to repeat the experiment, so NASA ended up getting him involved (which basically amounts to giving him the million, I guess).
The only other case that springs to mind of an experiment where only one person can "reproduce" it, is (ahem) Jan-Hendrik Schon.
Since when were brave guys with little rockets at the cutting edge of science?
You are right, I replied to the wrong comment.
Patents were intended to be for inventions, that is, a device that is physically realizable and practical to build. It really must be a condition of a patent application that you can demonstrate that the invention works, and performs as set out in the papent claim.
That the USPTO has prostituted itself to big money over the years doesn't change that intent at least, even if they broke the mechanism (hopefully tempoarily).
I agree that in principle, there is no harm (apart from waste of resources) in patenting impossible inventions. The problems come though when the holders of the patent use it as a club against other, related inventions that might actually work. So in practice it can be a very bad thing.
Huh? (he says, gibbering madly)
In no sense did Farnsworth work in isolation. The fact that the inventor of TV is a rather disputed claim, and there were a whole host of inventions that led up to it (Logie-Baird's work, for example) is enough evidence for that.
Just because someone works alone doesn't mean that they don't read the technical literature, attend conferences or exhibitions, or borrow ideas from other people.
The important factor is to make sure that enough samples are taken out of the entire population such that the variance in the result is "small enough".
:-)
If you want to test what colour my hair is you only need to take one sample because the population size for this experiment is 1.
If you want to test for bias in a set of 100 coin tosses, a sample size of 1 is certainly not adequate, even 10 samples (10% of the population) might not be enough, if the bias is small. But if the bias is big (say, all 10 samples comes up "heads"), its a fairly strong indicator. But, it is still a statistical estimate; you might just have been very unlucky, and a 1/1024 chance will happen, well, one time in 1024
I would imagine that for such a small effect such as someone getting a mild disease and ending up obease (which surely happens in a some fraction of cases anyway, and provides no evidence for a causal relationship between the disease and obeasity), the required sample size for a significant result would be very big. After all, surely very many people have suffered a minor disease at some stage of their life. How many of them are obease?
I've sometimes heard this as "most of the easy science has already been done", but that is also not very well worded. What it means is that it is no longer possible to make fundamental discoveries (or even any discoveries at all, apart from personal ones) by rolling balls down an inclined plane. Instead, you need either a large scale machine, say a particle accelerator, or even a small scale machine, say a STM machine or similar. Both of these are quite complex devices, which one person could not build and operate by themselves. But, I would argue the "easy" point though. In the 17th century, Galelleo's experiment was surely not "easy", although it is easy to reproduce with modern equipment.
Actually, that isn't true. The whole point of wave-particle duality is that a single photon will behave in a similar way to a wave - and any wave can be considered to be made up of one or more particles. If you attempt to measure the polarization of a single photon but your detector is not at the correct angle, then you will get a probabilistic result. This is equally true for a single photon or many photons. I guess your point though is that if the beam is polarized, then the photons will have consistent polarization so you can rotate the detector until you get 100% in the same direction. But most photon sources emit photons with random polarization.
He certainly was not suppressed, at least not by the scientific community. When General Relativity was published, for example, many other people working in the field immediately realized the significance and jumped on it. IIRC, Minkowsky himself was only months away from figuring it out himself. It is absolutely untrue that no one accepted the theories until the eclipse experiment (which, ironically, was later shown to to be "probably fraudulant" in fact! A great example of post-war anglo-german cooperation, publicised to the point that they HAD to get a result so much that it was ultimately faked). By that time, Einstein was already very famous, and relativity was by far the most likely contender for the theory of gravitation.
Of course, there were (and still are) competing theories. A major test of relativity occurred in the late 70s / early 80s when the rotational speed of a binary pulsar was measured to be slowing down, in accordance with the loss of energy from radiating gravitational waves, predicted by general relativity. At the time, there were several theories that reproduced the previously tested predictions of relativity but predicted a different rate of slowing down. The 1993 nobel prize was given for this in fact.
Sure, a quick glance in /usr/lib suggests that most libs have at least two versions.
But that is very different to having copy for every process, which is what the OP was suggesting. Currently ps shows ~150 processes currently loaded, I guess most of them are linked against libc. Using a different copy of the lib for each process would increase the amount of RAM I need by well over 100MB, and that is only for a single lib.
As he says, 'There is no scientific claim so preposterous that a scientist cannot be found to vouch for it'. What he doesn't say is that there are plenty more who will invest in it or base legislation on it."
From the article, the full paragraph of the quote is:
There is, alas, no scientific claim so preposterous that a scientist cannot be found to vouch for it. And many such claims end up in a court of law after they have cost some gullible person or corporation a lot of money. How are juries to evaluate them?
The very next sentence indicates that there are very many people who are willing to invest or base laws on bad science!
The number of copies in memory, for one thing.
Consuder how many processes are loaded on a typical 'doze system. Now imagine what would happen if every one of them had a different copy of USER32.DLL (or whatever).
Well, on a typical system there might be a hundred or so processes. Add 4MB to all of them and its suddenly not insignificant anymore.
Plus, if two running processes are sharing a shared library, then a task swap doesn't completely blow away the cache.
Finally, I think with Windoze I think there is a mode in which DLL's can have global data, shared among all programs using it. Not sure though, its years since I did any windoze programming.
This will solve some problems where installing software blows away a specific version of some DLL that is required by some other piece of software, but at the expense of introducing hard to solve problems where two programs A and B expect to use the same DLL, but end up using different ones. Knowing M$, it will probably not be possible to manually change which version is linked in.
Well, considering that I had never heard of "design patents" until yesterday, I thought I was doing OK! :-)
Where did slashdot claim that apple had a *software* patent on the trashcan? All I see is a link to an article on software patents that has some relevance because it talks about " the difference between utility and design patents".
i tion.htm on the criteria for getting a design patent: ... In addition, 35 U.S.C. 171 requires that a design to be patentable must be "original". Clearly a design which simulates a well known, or naturally occurring object or person is not original as required by the statute.
Interestingly, from http://www.uspto.gov/web/offices/pac/design/defin
I find it hard to believe that a picture of a trash can can fall into the "original" category.
Well, infrastructure improvements can be great in the long term, as long as its the correct infrastructure.
;)
Actually I don't know whether the infrastructure investments in Japan have been good, or maybe the increased basic research funding was just slop left over from masses of useless dam building projects
I'm now waiting to see what reaction the 'bankers' comment gets!
Yes, but a big part of Japans efforts to improve the economy have been focussed around increasing research, especially basic reasearch.
:-(
Is slashdot full of bankers now? It used to be full of techie geeks
Dieses ist ein Fehler, dort ist KEIN Problem mit amerikanischem Forschung Ausgang!
The GNU multiprecision library (GNUmp) is an excellent choice for this, very fast, although it is a bit cumbersome to use directly. I use a very simple C++ wrapper, but a more sophisticated one is now included in the library. The URL is in the story I believe.
This is what we use here:
Most non-numerical people use matlab, maple or mathematica when they want to solve a relatively simple problem. For bigger problems, in my experience it is easier and faster to just write (or get someone else to write) a C or C++ or Fortran program to do it, rather than use higher level tools, * unless the problem can be expressed particularly simply in mathematica and not in a general purpose programming language. I've only ever met one person who attempted running a mathematica problem on a supercomputer (although I know that people do this all the time): he basically ended up as an alcoholic dropout.
Only the old guys use Fortran.
Only misguided (IMO) people use C for anything big.
Libraries: I use GMP, BLAS (usually in the form of ATLAS), LAPACK, ARPACK, MPI, pthreads (OpenMP is usually a better choice, and nowdays has pretty good compiler support - unfortunately I wrote my main codes a couple of years to early). All of these are solid, reliable and fast. GSL looks good but I havn't used it.
Note though that none of the previous are actually C++ libraries. ARPACK++ is crap, most C++ linear algebra libraries are crap, the C++ bindings for MPI are crap, pretty much all existing scientific libraries in C++ are crap. However, some recent libraries have potential, such as the stuff on boost.org (eg, uBLAS), maybe MTL will end up being good but the existing release is not IMHO. Blitz++ is good, but the transition away from a single-developer project has been long, AFAICT it is still basically unmaintained. Pooma appears to be really solid and usable, but I've never gone beyond reading some of the web tutorials.
But, I practically never use a non-C++ library 'raw', I always put at least a minimal C++ wrapper. With careful use of templates, you can get some pretty nice looking (and efficient) code. This is especially important for Fortran. Apart from needing platform-dependent hacks to actually call Fortran functions, I havn't yet seen a Fortran interface that I actually like. Too many parameters, obscure, historical conventions, the dreaded "reverse communication interface" etc etc.
I also use xmgrace, recently figured out how to script it and that is a very powerful tool. I use Gnuplot for 3d plots, but I havn't really figured out how to work it yet.
I also use Perl a bit, but mostly for job control sort of things (copying files across before the start of an MPI calculation, submitting jobs to the queue, etc).