The first settlers in America experienced enormous causalities their first years they were in America. Entire colonies were lost.
There is little or no evidence as to the fate of the first colonies in America, although we know that ultimately some of the most enterprising and capable must have survived. It was over ten thousand years ago, after all.
IAANP, and from what I can tell from the article, the design looks similar in concept to the boiling water CANDU, which didn't quite live up to its promise.
There are a couple of problems with steam-moderated designs:
maintaining a given density is not trivial
steam pipes tend to erode
The first problem is a polite way of saying there may be reactor stability issues. If CAESAR uses super-heated steam this may be less of an issue, but otherwise the ratio of steam to liquid in the cooling circuit is a function of pressure and temperature in ways that can create problems if there is an unexpected excursion.
The second problem is a major issue, especially when coupled with the long-term affects of radiation on materials. Intense neutron bombardment is a good way to introduce defects in metallic lattices. Defects are a good place for corrosion, cracking and other bad things to start. Ergo, the odds of a sealed reactor lasting for more than a decade or so are not good. Retubing of reactors is an ongoing maintenance problem even in conventional designs.
Nuclear power is an option that we may in the end decide we have to go with, and it's good that advanced reactor designs continue to get consideration, but the engineering challenges are still severe and the proliferation potential is large.
The important thing about this "teleportation" process to remember is: if you stick your hand into the region between the transmitter and reciever you will still get a hole burned in it by the perfectly ordinary beam of energetic, physical photons that is "teleporting" the information.
Tonnes of plutonium are "missing" in the U.S. for pretty much the same reason. Process accounting is just not all that accurate, and keeping track of things at the 1% level means that there's lots of opportunity for error.
All this story is saying that that the measured amount is 3% lower than the estimated amount, and for anyone who has ever tried to calculate the results of a complex process with an accuracy of 1%, this is not a big surprise.
On the other hand, these sorts of "accounting losses" do provide an opportunity for terrorists to get their hands on the stuff secretly. This is very much like "coin clipping" back in the days when money was made out of gold or silver--get a coin, pare a bit off the edge, pass it on at face value, and eventually you've enhanced your wealth by a significant amount. Milling of coin edges was introduced to prevent this practice, with indifferent success.
Unfortunately, there's no practical defense against this kind of thing, which should make us just a little leary about the prospects of "peaceful" nuclear power. My own attitude is "nuclear power if necessary, but not necessarily nuclear power."
The VTK (http://www.kitware.com) framework unit tests visualizations by dumping a png of the visualization and diffing it against a "correct" version. This is is a pretty extreme test--one pixel out of place and it fails--but it does ensure stability of the rendering code.
A chessboard is 8x8, meaning 64 spaces. However, each space can contain a pawn, a rook, a bishop, a knight, a king or a queen of either colour. The best estimate for the number of states the board can be in is 2.99x1041.
Not all of these states are reachable by legal moves, though. The number of reachable states is still vast, but far smaller than the number of total states, and far harder to calculate.
This asshole says she doesn't do anything dishonest. In particular:
She doesn't forge or falsify the message headers;
But at the far end of the article we read about her computer guy:
...he's found people are more likely to open e-mail if it appears to be from a real person, so he types his friends' names on "from" lines. "The trick is to make it look personal," he said as he tapped out commands on his computer. "You want to make it look like it comes from the guy in the cubicle down the hall."
Ok, so isn't the "from" line in in some narrow, literal, technical sense, part of the message header?
--Tom
Scientists are extremely uptight about exact numerical analysis. We get the data and compare them to a tightly parameterized model which includes everything we know about our detector response as well as the probable sources of the events we are detecting. Good models have small numbers of parameters and many constraints compared to the richness of the data. "With enough free parameters you can fit an elephant", the saying goes, which indicates how important it is to scientists to keep the number of parameters small--no one wants to see a comment like that on a referee's report!
With regard to gravitational observatories, the data are very rich: polarization, amplitude, phase and frequency spectra will be available, possibly from several detectors with different orientations. Detector response is also extremely well understood. The theoretical physics of the sources--general relativity--is also very well understood, and models of stellar collapse, neutron star collisions, etc, contain few parameters (masses, angular momenta, impact parameter...that's about it.)
As such, we can compare model to reality and produce a statistically valid likelihood that the model is false. The Baysians in the audience will point out that relative to our prior knowledge we can also produce the probability that the model is true.
So it isn't a matter of getting something that "roughly fits"--the analysis either produces a fit within error or it does not. If it does not, we dig more deeply into the possible sources of disagreement. The data are sufficienty rich that many, many types of cross-checking and internal consistency checking will be available.
To a hardened skeptic, this of course will not do. But hardened skepticism is an anti-scientific attitude. Scientists are open-minded skeptics, who are able to keep the contingent nature of their beliefs in mind while at the same time maintaining a commitment to distinguish clearly between probable truth and probable falsehood.
A number of comments posted here are mistaken about the nature of shape in the quantum world. It has nothing to do with the probability distribution of the proton's location; it has to do with the probability distributions of the quarks within the proton.
Only entities that are extended in space have shape. Electrons, for example, are to the best of our knowledge pointlike, and therefore we say they do not have shape. Atoms, on the other hand, are extended in space and therefore do have shape. Their shape is given to them by the probability distributions for the locations of the electrons that surround them, not by the probability distribution of their own location.
Shape is described in nuclear physics using the mathematical device of the "structure function", which is just a function of the three spatial co-ordinates that describes the size and (a)symmetries of the nucleus relative to some interaction. Structure functions are ususally expressed as linearly weighted sums of spherical harmonics or Legendre polynomials, which capture physically interesting processes in different terms.
Protons are believed to have a spherically symmetric structure function. On a number of both theoretical and experimental grounds I'd be extremely doubtful that the dynamical structure function of the proton is asymmetric, and the article in fact suggests that the result is a purely kinematic one due to relativistic effects, which is suprising but in the final analysis not the kind of earth-shattering news dynamical asymmetry would be.
In particular, a dynamical asymmetry of the proton would imply naively a comparable asymmetry of the neutron, and this is known not to exist experimentally at very high precision (and whose existence at a very small level is predicted by electro-weak theory.)
I thank my graduate student, Daniel Solli, for pointing out this cross term in the expansion of the minimal-coupling Hamiltonian to me.
I guess this is better than no acknowledgement at all, but where I come from the person who provides the fundamental insight that is the basis for a new piece of work usually gets co-authorship.
--Tom
Re:Guilt By Association, don't buy it
on
Monsanto and PCBs
·
· Score: 1
Who cares that some GM crops in some circumstances might sometimes be good, when we know perfectly well that those GM crops are exactly the ones that neither corporations (interested in profits) nor governments (interested in control) will ever be motivated to produce?
That fact that water in some form is necessary for life is no use to someone drowning in mid-ocean, and mid-ocean is pretty much where we are with regard to GM crops.
I've been doing OO programming in physics and engineering (mostly Monte Carlo simulation, but also statistical analyis of data) since the late '80's. The trick is to reconceptualize your problem, which is independent of the language you're writing in. My early OO efforts were in FORTRAN77 and C, now long abandoned in favor of C++.
I've since seen a lot of what passes for OO programming in the sciences, and most of it isn't--it's thinly disguised procedural programming with clean and well-specfiied interfaces.
The difference between OO and procedural programming at the design level is the way in which programs are structured. Procedural programs are at their core a collection of algorithms that operate on data. OO programs (should) have at their core representations of real-world objects, with neither data nor algorithms being the primary abstraction. The be an object a thing must have both state (data) and behavior (algorithms).
We are still in very early days yet in applying OO principles to science and engineering problems. Even "OO" numerical libraries are to my mind a very, very small step: they represent traditional, procedural algorithms in slightly richer and more maintainable ways. Having a matrix "object" for instance, is a definite gain ove having a flat array and bunch of procedures that work on it, but the success of the STL suggests that generic programming is a much better approach to this sort of problem than OO programming.
The place where OO programming will come to the fore is in allowing rich and intelligent representations of real-world objects in engineering software.
For example, I once wrote an OO layer overtop of a Levenberg-Marquardt minimizer that allows me to represent the objective function in terms of physically meaningful components that can take constraint objects that prevent the minimizer from entering physically impossible parts of the parameter space. In one sense this is a trivial problem, but in a larger sense it allows me to quickly specify components and constraints without ever having to think about array indices and other error-prone nonsense.
It is this aspect of OO--its ability to produce rich and natural representations of complex problems--that we have hardly begun to explore, and where the real strength of the approach lies for science and engineering.
PMTs are remarkably robust. When we were building SNO, we tested a bunch of PMTs. Amongst other tests, we pressurized a tank to over 80 PSI and tried to smash the tubes inside using a rod that pushed through a pressure fitting.
The outcome of this test was that mostly the PMTs did not implode. There was a strong tendency for the rod to simply punch a rod-sized (about 1 cm) hole in the target tube. Putting a blunt block on the end of the rod did eventually produce an instance where the target tube smashed. The adjacent tubes, which were mounted in closer proximity to the target tube than they are in the real detector (and much closer than the tubes are in SuperK) were not damaged, despite being visibly twisted in their mountings.
Caveate: the tubes used in SuperK are about twice the size of those used in SNO, and therefore correspondingly more fragile. But having handled these tubes a good deal, I can say that it takes more than a small bang to break them. Whatever happened at SuperK (the NYT story is weirdly uninformative, to the extent that I wonder if they don't have all the major facts wrong) it is unlikely to be as simple as a chain reaction of imploding tubes.
Big Bang nucleosynthesis determines the amount of baryonic ("normal") matter because it allows us to predict the primordial abundances of light elements as a function of that amount, most dramatically the He/H ratio, which is known from observation and inference to have been about 1:3.
It works like this: early in the Big Bang, everything was so hot that quarks were essentially free, forming a big, hot slurry called a quark-gluon plasma. As the primordial fireball expanded and cooled, protons and neutrons condensed out of the quark-gluon plasma like drops of mist forming in wet, cooling air.
The temperature of the universe at this time was still very high, which means the average energy of these protons and neutrons was much higher than the binding energy of light nuclei, so no sooner would a proton and neutron bind to form a deuterium nucleus than they would get knocked apart by another particle wacking into them.
The universe cooled further, and eventually reached a temperature where the average particle energy was low enough that light nuclei could form and not get disrupted. Originally, roughly equal numbers of neutrons and protons were formed, but free neutrons decay with a lifetime of about 15 minutes, so by the time the universe cooled enough for neutrons to bind permanently, there were a lot fewer of them.
The fraction of neutrons and the total density of neutrons and protons at the time the universe reached this temperature determines the amount of helium in the early universe, which we can measure by correcting for the subsequent production of helium in stars. Thus, we can determine the density of the universe when the light nuclei were being formed. We also know the radius of the universe at that time by various chains of inference. Knowing the radius and the density gives us the mass: that is, the total amount of normal matter.
The details of Big Bang nucleosythesis are one of the great triumphs of 20th century physics, for which Fowler got the Nobel Prize back in the '60's. There have been anomalies detected (lithium is particularly difficult to deal with) but the theory has so far stood up to a lot of critical examination.
Dark matter (or modified gravitational potentials, which I find kind of appealing) get invoked as explanations for anomalies in galactic dynamics precisely because Big Bang nucleosynthesis stands on such solid foundations.
Slashdot Mirror
There is little or no evidence as to the fate of the first colonies in America, although we know that ultimately some of the most enterprising and capable must have survived. It was over ten thousand years ago, after all.
--Tom
IAANP, and from what I can tell from the article, the design looks similar in concept to the boiling water CANDU, which didn't quite live up to its promise.
There are a couple of problems with steam-moderated designs:
The first problem is a polite way of saying there may be reactor stability issues. If CAESAR uses super-heated steam this may be less of an issue, but otherwise the ratio of steam to liquid in the cooling circuit is a function of pressure and temperature in ways that can create problems if there is an unexpected excursion.
The second problem is a major issue, especially when coupled with the long-term affects of radiation on materials. Intense neutron bombardment is a good way to introduce defects in metallic lattices. Defects are a good place for corrosion, cracking and other bad things to start. Ergo, the odds of a sealed reactor lasting for more than a decade or so are not good. Retubing of reactors is an ongoing maintenance problem even in conventional designs.
Nuclear power is an option that we may in the end decide we have to go with, and it's good that advanced reactor designs continue to get consideration, but the engineering challenges are still severe and the proliferation potential is large.
--Tom
The important thing about this "teleportation" process to remember is: if you stick your hand into the region between the transmitter and reciever you will still get a hole burned in it by the perfectly ordinary beam of energetic, physical photons that is "teleporting" the information.
--Tom
Tonnes of plutonium are "missing" in the U.S. for pretty much the same reason. Process accounting is just not all that accurate, and keeping track of things at the 1% level means that there's lots of opportunity for error.
All this story is saying that that the measured amount is 3% lower than the estimated amount, and for anyone who has ever tried to calculate the results of a complex process with an accuracy of 1%, this is not a big surprise.
On the other hand, these sorts of "accounting losses" do provide an opportunity for terrorists to get their hands on the stuff secretly. This is very much like "coin clipping" back in the days when money was made out of gold or silver--get a coin, pare a bit off the edge, pass it on at face value, and eventually you've enhanced your wealth by a significant amount. Milling of coin edges was introduced to prevent this practice, with indifferent success.
Unfortunately, there's no practical defense against this kind of thing, which should make us just a little leary about the prospects of "peaceful" nuclear power. My own attitude is "nuclear power if necessary, but not necessarily nuclear power."
--Tom
--Tom
Not all of these states are reachable by legal moves, though. The number of reachable states is still vast, but far smaller than the number of total states, and far harder to calculate.
--Tom
She doesn't forge or falsify the message headers;
But at the far end of the article we read about her computer guy:
Ok, so isn't the "from" line in in some narrow, literal, technical sense, part of the message header? --Tom
Scientists are extremely uptight about exact numerical analysis. We get the data and compare them to a tightly parameterized model which includes everything we know about our detector response as well as the probable sources of the events we are detecting. Good models have small numbers of parameters and many constraints compared to the richness of the data. "With enough free parameters you can fit an elephant", the saying goes, which indicates how important it is to scientists to keep the number of parameters small--no one wants to see a comment like that on a referee's report!
With regard to gravitational observatories, the data are very rich: polarization, amplitude, phase and frequency spectra will be available, possibly from several detectors with different orientations. Detector response is also extremely well understood. The theoretical physics of the sources--general relativity--is also very well understood, and models of stellar collapse, neutron star collisions, etc, contain few parameters (masses, angular momenta, impact parameter...that's about it.)
As such, we can compare model to reality and produce a statistically valid likelihood that the model is false. The Baysians in the audience will point out that relative to our prior knowledge we can also produce the probability that the model is true.
So it isn't a matter of getting something that "roughly fits"--the analysis either produces a fit within error or it does not. If it does not, we dig more deeply into the possible sources of disagreement. The data are sufficienty rich that many, many types of cross-checking and internal consistency checking will be available.
To a hardened skeptic, this of course will not do. But hardened skepticism is an anti-scientific attitude. Scientists are open-minded skeptics, who are able to keep the contingent nature of their beliefs in mind while at the same time maintaining a commitment to distinguish clearly between probable truth and probable falsehood.
--Tom
A number of comments posted here are mistaken about the nature of shape in the quantum world. It has nothing to do with the probability distribution of the proton's location; it has to do with the probability distributions of the quarks within the proton.
Only entities that are extended in space have shape. Electrons, for example, are to the best of our knowledge pointlike, and therefore we say they do not have shape. Atoms, on the other hand, are extended in space and therefore do have shape. Their shape is given to them by the probability distributions for the locations of the electrons that surround them, not by the probability distribution of their own location.
Shape is described in nuclear physics using the mathematical device of the "structure function", which is just a function of the three spatial co-ordinates that describes the size and (a)symmetries of the nucleus relative to some interaction. Structure functions are ususally expressed as linearly weighted sums of spherical harmonics or Legendre polynomials, which capture physically interesting processes in different terms.
Protons are believed to have a spherically symmetric structure function. On a number of both theoretical and experimental grounds I'd be extremely doubtful that the dynamical structure function of the proton is asymmetric, and the article in fact suggests that the result is a purely kinematic one due to relativistic effects, which is suprising but in the final analysis not the kind of earth-shattering news dynamical asymmetry would be.
In particular, a dynamical asymmetry of the proton would imply naively a comparable asymmetry of the neutron, and this is known not to exist experimentally at very high precision (and whose existence at a very small level is predicted by electro-weak theory.)
--Tom
I guess this is better than no acknowledgement at all, but where I come from the person who provides the fundamental insight that is the basis for a new piece of work usually gets co-authorship.
--Tom
Who cares that some GM crops in some circumstances might sometimes be good, when we know perfectly well that those GM crops are exactly the ones that neither corporations (interested in profits) nor governments (interested in control) will ever be motivated to produce?
That fact that water in some form is necessary for life is no use to someone drowning in mid-ocean, and mid-ocean is pretty much where we are with regard to GM crops.
--Tom
I've been doing OO programming in physics and engineering (mostly Monte Carlo simulation, but also statistical analyis of data) since the late '80's. The trick is to reconceptualize your problem, which is independent of the language you're writing in. My early OO efforts were in FORTRAN77 and C, now long abandoned in favor of C++.
I've since seen a lot of what passes for OO programming in the sciences, and most of it isn't--it's thinly disguised procedural programming with clean and well-specfiied interfaces.
The difference between OO and procedural programming at the design level is the way in which programs are structured. Procedural programs are at their core a collection of algorithms that operate on data. OO programs (should) have at their core representations of real-world objects, with neither data nor algorithms being the primary abstraction. The be an object a thing must have both state (data) and behavior (algorithms).
We are still in very early days yet in applying OO principles to science and engineering problems. Even "OO" numerical libraries are to my mind a very, very small step: they represent traditional, procedural algorithms in slightly richer and more maintainable ways. Having a matrix "object" for instance, is a definite gain ove having a flat array and bunch of procedures that work on it, but the success of the STL suggests that generic programming is a much better approach to this sort of problem than OO programming.
The place where OO programming will come to the fore is in allowing rich and intelligent representations of real-world objects in engineering software.
For example, I once wrote an OO layer overtop of a Levenberg-Marquardt minimizer that allows me to represent the objective function in terms of physically meaningful components that can take constraint objects that prevent the minimizer from entering physically impossible parts of the parameter space. In one sense this is a trivial problem, but in a larger sense it allows me to quickly specify components and constraints without ever having to think about array indices and other error-prone nonsense.
It is this aspect of OO--its ability to produce rich and natural representations of complex problems--that we have hardly begun to explore, and where the real strength of the approach lies for science and engineering.
--Tom
PMTs are remarkably robust. When we were building SNO, we tested a bunch of PMTs. Amongst other tests, we pressurized a tank to over 80 PSI and tried to smash the tubes inside using a rod that pushed through a pressure fitting.
The outcome of this test was that mostly the PMTs did not implode. There was a strong tendency for the rod to simply punch a rod-sized (about 1 cm) hole in the target tube. Putting a blunt block on the end of the rod did eventually produce an instance where the target tube smashed. The adjacent tubes, which were mounted in closer proximity to the target tube than they are in the real detector (and much closer than the tubes are in SuperK) were not damaged, despite being visibly twisted in their mountings.
Caveate: the tubes used in SuperK are about twice the size of those used in SNO, and therefore correspondingly more fragile. But having handled these tubes a good deal, I can say that it takes more than a small bang to break them. Whatever happened at SuperK (the NYT story is weirdly uninformative, to the extent that I wonder if they don't have all the major facts wrong) it is unlikely to be as simple as a chain reaction of imploding tubes.
--Tom
A troll is worth a thousand words :-)
Big Bang nucleosynthesis determines the amount of baryonic ("normal") matter because it allows us to predict the primordial abundances of light elements as a function of that amount, most dramatically the He/H ratio, which is known from observation and inference to have been about 1:3.
It works like this: early in the Big Bang, everything was so hot that quarks were essentially free, forming a big, hot slurry called a quark-gluon plasma. As the primordial fireball expanded and cooled, protons and neutrons condensed out of the quark-gluon plasma like drops of mist forming in wet, cooling air.
The temperature of the universe at this time was still very high, which means the average energy of these protons and neutrons was much higher than the binding energy of light nuclei, so no sooner would a proton and neutron bind to form a deuterium nucleus than they would get knocked apart by another particle wacking into them.
The universe cooled further, and eventually reached a temperature where the average particle energy was low enough that light nuclei could form and not get disrupted. Originally, roughly equal numbers of neutrons and protons were formed, but free neutrons decay with a lifetime of about 15 minutes, so by the time the universe cooled enough for neutrons to bind permanently, there were a lot fewer of them.
The fraction of neutrons and the total density of neutrons and protons at the time the universe reached this temperature determines the amount of helium in the early universe, which we can measure by correcting for the subsequent production of helium in stars. Thus, we can determine the density of the universe when the light nuclei were being formed. We also know the radius of the universe at that time by various chains of inference. Knowing the radius and the density gives us the mass: that is, the total amount of normal matter.
The details of Big Bang nucleosythesis are one of the great triumphs of 20th century physics, for which Fowler got the Nobel Prize back in the '60's. There have been anomalies detected (lithium is particularly difficult to deal with) but the theory has so far stood up to a lot of critical examination.
Dark matter (or modified gravitational potentials, which I find kind of appealing) get invoked as explanations for anomalies in galactic dynamics precisely because Big Bang nucleosynthesis stands on such solid foundations.
--Tom