* A total SU(3) is a neat idea, but how does it jibe with experimental data? Given the difficulties with technicolor, my guess is that it doesn't.
* Gravitons are spin 2, but it's not true that spin 2 particles are gravitons. Can you identify your rank-2 symmetry generators with the Lorentz group? Finally, is it renormalizable?
To me endeavors like this are the most perfect expression of man. Vonnegut wrote in Breakfast of Champions,
Our awareness is all that is alive and maybe sacred in any of us. Everything else about us is dead machinery.
To plunge into the unknown is a moral imperative for any thinking being.
If all you care about are material practicalities, this thing is roughly 1/50th the current cost of a certain misadventure in the Middle East, and is more likely to produce cool stuff. One particularly exciting bit of technology already is the LHC's grid computing infrastructure.
* What does the quarternion formulation tell us that the standard Standard Model formulation does not? I understand that it provides a unified framework for treating the different groups, but particles in the Standard Model are still charged separately under electroweak and strong -- is this a high energy theory, where we expect gauge coupling unification somewhere?
* I don't understand your concept of inertial mass breaking gauge symmetries. The Standard Model is Lorentz invariant, and gauge particles have to be massless to preserve gauge invariance. A Higgs condensate breaks the SU(2) gauge symmetry by making the gauge fields massive. How does your idea work?
It's our results that need to be replicated, not our methods. Anyone can do the same thing and get the same result. Doing it by a different method and getting the same result is a much more rigorous validation. As long as it's a different box, it doesn't matter if one or both are black. This is indeed the common practice, for codes that are short enough to duplicate without too much effort. However, if you have a collider that requires millions of lines of code just to do event storage and tagging, it will be unique. This means it must be open source.
You could argue that Mathematica provides a nice high level, black box platform for writing small private codes that can be duplicated on Maple or Maxima. Yes, in 99+% of cases they will get the same results and this is just fine. In the remainder, the way to validate the underlying tool is to port the same code. Even then, you would only validate for this particular task.
Mathematica, being a CAS without peer, cannot be validated in general -- you would need a tool like Maxima to generate the enormous (infinite?) number of validation tests to check the black box inputs and outputs, and Maxima has a ways to go to catch up to Mathematica. It would be satisfactory if Wolfram exposed the core API of atomic operations and published unit tests, then made everything on top of it open source.
I should've said that vectorization paths are not used very much, if at all. Because of the work put into validation, you'll see in the literature that only recently have people publishing public codes been trying out SSE and the like.
AFAIK, you need to jump through hoops to get gcc to do automatic vectorization. (For example.) I think it will be a while before it's included into the "-O#" optimizations that physicists are comfortable with.
Nobody in physics, at least, uses commercial packages for published claims. They either write their own codes or use packages like ROOT that are open source.
(IDL has some traction in astronomy and aeronomy, but these are the very people leading the SciPy development effort.)
The only way to ensure complete coverage is by unit testing -- this requires atomic operations to be exposed, with everything else open source on top of it.
The working standard is: how do we ensure that the makers of some instrument (hardware or software) is performing to the advertised spec?
For hardware, it's published specs and calibration data; if there are questions, open design documents can be scrutinized, as well as the devices themselves.
For software, it's unit, regression and validation tests. If there are questions, open source can be scrutinized.
wxMaxima and SciPy could round out a nice open-source scientific suite. SciPy uses wxWidgets for its GUIs as well. (I prefer GNUstep personally, but whatever works.)
In the distant future, some kind of integrated portable platform for both high-performance numerical analysis and reliable symbolic/special functions calculation would be fantastic.
The Hubble telescope and LHC have produced many hundreds, if not thousands, of papers on instrumentation: design, construction and calibration. Moreover, there are hundreds of eyes from competing institutions looking at various components during the entire process.
Mathematica's programming language is transparent, but the code behind all the special functions and numerical routines have only been seen by people working toward Wolfram's bottom line. To their credit, they cite scientific references for their algorithms, but they do not distribute their codes for examination by others.
There is also a serious scientific issue in using closed-source software for data analysis or theoretical calculations. All scientific work should be transparent to review and reproducible, from first principles, in order to find validate any findings. A black box code is antithetical to this principle.
Mathematica, wonderful as it is, should only be used for prototyping.
Has anyone tracked utilization of Vit.D that's self-generated vs consumed as a supplement?? I wonder how much difference that might make. It's probably a moot point. The biochemistry of Vitamin D is well known, and the D3 you eat is exactly the same chemically as what's produced by the skin. Any difference in bioavailability between digestion and direct production is handled by monitoring blood levels.
Vitamin D2 in humans is a bit trickier, but its biochemistry is also well-known from studying animals.
Since I spent 2 to 5 hours a day in the morning and late-afternoon sun, I doubt I'm in any serious deficit.:) Make sure it is outdoors, and not just basking in the sunlight from behind a window. Glass doesn't block UVA, so you can get a tan, but it does block UVB, which is necessary for Vitamin D production.
Toxicity cases are rare, and occur when blood levels reach multiples of the recommended upper limit. You have to abuse prescription strength supplements or guzzle cod liver oil to achieve the necessary high doses.
The toxic mechanism is hypercalcemia: increased calcium mobility and reabsorption through the kidneys, which can get damaged.
You make a good point: the amount of Vitamin D you get from normal diet is very small compared to the amount of Vitamin D obtained by exposure to UV-B radiation, but this amount is zero if your Vitamin D blood level is already maxed out due to chemical equilibrium in the skin.
Really, prescription supplements or cod liver oil are only appropriate if there's a demonstrated deficiency like pain/osteomalacia/rickets.
- The dosage a white person gets in the sun is something like 20,000 IU after only 20 minutes. (Increased sun exposure does not lead to toxicity because if the blood level of Vitamin D is high, any extra produced stays in the skin and is broken down.)
- Vitamin D is fat soluble. One you are depleted, the fatty tissues (e.g. liver) must be saturated before the blood level rises.
I took 50,000 IU of Vitamin D2 three times a week for one month, and since then once every 10 days. Note that Vitamin D2 has 40% the efficacy of Vitamin D3, and so requires 2.5 times the D3 blood level. (DISCLAIMER: I am not a medical doctor! I just see one.)
* Dietary allowances
The RDA for Vitamin D is what is prescribed to white children to prevent rickets, usually manifesting over the winter after they've grown without enough available Vitamin D. If you don't get enough sun for your skin tone the requirements will be far, far higher year 'round.
I have Vitamin D deficiency, and it came close to ruining my life. I am a scientist, but I also have dark skin that never burns. Even though I don't own a car, I just can't get the 2-3 hours of sunshine daily needed to fulfill my Vitamin D requirement; white folks only need 20 minutes. Moreover, it's kind of chilly where I live, so I wear long pants and sleeves much of the time.
Over time, I developed a pain that just sucked the life out of me -- like I was playing four quarters of football daily, with the flu. Even with powerful pain killers I couldn't sleep, and woke up every day feeling I was hit by a bus.
The link to cancer is still an open question, but the pain is a hard fact.
PS: The only way to overdose on Vitamin D is to abuse prescription-strength supplements or cod liver oil.
I don't think that post-1997 admittees are underqualified, only that it's a mistake for MIT to seek the same students as the Ivies. MIT is in a niche of its own that is incredibly important from a historical perspective; "well-roundedness" is a good thing, but is not the last word in what is valuable to an institution or to the future success of an individual. Over the long-term, MIT risks becoming an also-ran by pursuing the same admissions strategies as Harvard and Princeton, because it can never achieve the same kind of cachet.
As an MIT alum admitted prior to her regime, I'm pleased that she will be replaced. While her lies are a black eye for the institute, her admissions policies and personal philosophy had done more damage. In an attempt to admit "well-rounded" students and compete with the Harvards of the world, she chipped away at the identity that makes MIT unique: academic excellence, creativity and fun. If that makes MIT too "geeky," so be it.
Moreover, her outspokenness reduced the dignity of her position and the process. Admissions should serve the principles of the school -- period.
You could make the same argument about tangible goods that are more sophisticated than food.
As our economy has evolved food productions has rightfully shifted to countries with cheaper labor, as our intellectual goods have a more far-reaching impact.
You've made the case that the current IP regime is broken (patents too easily granted, prior art, overly long patent/copyright extension), not that IP is fundamentally broken. [You may wish to read the other branch under my first reply.]
Specifically, software and drug patents are more contentious than machine and business process patents.
Slashdot Mirror
* A total SU(3) is a neat idea, but how does it jibe with experimental data? Given the difficulties with technicolor, my guess is that it doesn't.
* Gravitons are spin 2, but it's not true that spin 2 particles are gravitons. Can you identify your rank-2 symmetry generators with the Lorentz group? Finally, is it renormalizable?
To plunge into the unknown is a moral imperative for any thinking being.
If all you care about are material practicalities, this thing is roughly 1/50th the current cost of a certain misadventure in the Middle East, and is more likely to produce cool stuff. One particularly exciting bit of technology already is the LHC's grid computing infrastructure.
Questions:
* What does the quarternion formulation tell us that the standard Standard Model formulation does not? I understand that it provides a unified framework for treating the different groups, but particles in the Standard Model are still charged separately under electroweak and strong -- is this a high energy theory, where we expect gauge coupling unification somewhere?
* I don't understand your concept of inertial mass breaking gauge symmetries. The Standard Model is Lorentz invariant, and gauge particles have to be massless to preserve gauge invariance. A Higgs condensate breaks the SU(2) gauge symmetry by making the gauge fields massive. How does your idea work?
You could argue that Mathematica provides a nice high level, black box platform for writing small private codes that can be duplicated on Maple or Maxima. Yes, in 99+% of cases they will get the same results and this is just fine. In the remainder, the way to validate the underlying tool is to port the same code. Even then, you would only validate for this particular task.
Mathematica, being a CAS without peer, cannot be validated in general -- you would need a tool like Maxima to generate the enormous (infinite?) number of validation tests to check the black box inputs and outputs, and Maxima has a ways to go to catch up to Mathematica. It would be satisfactory if Wolfram exposed the core API of atomic operations and published unit tests, then made everything on top of it open source.
You're right about CPUs.
I should've said that vectorization paths are not used very much, if at all. Because of the work put into validation, you'll see in the literature that only recently have people publishing public codes been trying out SSE and the like.
AFAIK, you need to jump through hoops to get gcc to do automatic vectorization. (For example.) I think it will be a while before it's included into the "-O#" optimizations that physicists are comfortable with.
Yes, everyone uses an open source toolchain. On top of that, everyone advertising their code for scientific use has a validation suite.
Mass-market CPUs are single data/single instruction. So, between validation and open source, it is easy to catch hardware bugs and faults.
Nobody in physics, at least, uses commercial packages for published claims. They either write their own codes or use packages like ROOT that are open source.
(IDL has some traction in astronomy and aeronomy, but these are the very people leading the SciPy development effort.)
The only way to ensure complete coverage is by unit testing -- this requires atomic operations to be exposed, with everything else open source on top of it.
The working standard is: how do we ensure that the makers of some instrument (hardware or software) is performing to the advertised spec?
For hardware, it's published specs and calibration data; if there are questions, open design documents can be scrutinized, as well as the devices themselves.
For software, it's unit, regression and validation tests. If there are questions, open source can be scrutinized.
Huh -- somebody above mentioned SAGE, and it's intense: written in Python, uses NumPy, Maxima and matplotlib.
wxMaxima and SciPy could round out a nice open-source scientific suite. SciPy uses wxWidgets for its GUIs as well. (I prefer GNUstep personally, but whatever works.)
In the distant future, some kind of integrated portable platform for both high-performance numerical analysis and reliable symbolic/special functions calculation would be fantastic .
No.
The Hubble telescope and LHC have produced many hundreds, if not thousands, of papers on instrumentation: design, construction and calibration. Moreover, there are hundreds of eyes from competing institutions looking at various components during the entire process.
Mathematica's programming language is transparent, but the code behind all the special functions and numerical routines have only been seen by people working toward Wolfram's bottom line. To their credit, they cite scientific references for their algorithms, but they do not distribute their codes for examination by others.
There is also a serious scientific issue in using closed-source software for data analysis or theoretical calculations. All scientific work should be transparent to review and reproducible, from first principles, in order to find validate any findings. A black box code is antithetical to this principle.
Mathematica, wonderful as it is, should only be used for prototyping.
Vitamin D2 in humans is a bit trickier, but its biochemistry is also well-known from studying animals. Since I spent 2 to 5 hours a day in the morning and late-afternoon sun, I doubt I'm in any serious deficit.
Toxicity cases are rare, and occur when blood levels reach multiples of the recommended upper limit. You have to abuse prescription strength supplements or guzzle cod liver oil to achieve the necessary high doses.
The toxic mechanism is hypercalcemia: increased calcium mobility and reabsorption through the kidneys, which can get damaged.
You make a good point: the amount of Vitamin D you get from normal diet is very small compared to the amount of Vitamin D obtained by exposure to UV-B radiation, but this amount is zero if your Vitamin D blood level is already maxed out due to chemical equilibrium in the skin.
Really, prescription supplements or cod liver oil are only appropriate if there's a demonstrated deficiency like pain/osteomalacia/rickets.
Please also see my other comment, in response to some questions.
* Dosing
One must note two facts:
- The dosage a white person gets in the sun is something like 20,000 IU after only 20 minutes. (Increased sun exposure does not lead to toxicity because if the blood level of Vitamin D is high, any extra produced stays in the skin and is broken down.)
- Vitamin D is fat soluble. One you are depleted, the fatty tissues (e.g. liver) must be saturated before the blood level rises.
I took 50,000 IU of Vitamin D2 three times a week for one month, and since then once every 10 days. Note that Vitamin D2 has 40% the efficacy of Vitamin D3, and so requires 2.5 times the D3 blood level. (DISCLAIMER: I am not a medical doctor! I just see one.)
* Dietary allowances
The RDA for Vitamin D is what is prescribed to white children to prevent rickets, usually manifesting over the winter after they've grown without enough available Vitamin D. If you don't get enough sun for your skin tone the requirements will be far, far higher year 'round.
Very well, but how do you recoup the design costs if not by charging for reproductions?
I think we agree where the rubber meets the road -- the school and admittees should be suited to one another.
I had a problem with Marilee Jones having a far wider agenda than that.
I have Vitamin D deficiency, and it came close to ruining my life. I am a scientist, but I also have dark skin that never burns. Even though I don't own a car, I just can't get the 2-3 hours of sunshine daily needed to fulfill my Vitamin D requirement; white folks only need 20 minutes. Moreover, it's kind of chilly where I live, so I wear long pants and sleeves much of the time.
Over time, I developed a pain that just sucked the life out of me -- like I was playing four quarters of football daily, with the flu. Even with powerful pain killers I couldn't sleep, and woke up every day feeling I was hit by a bus.
The link to cancer is still an open question, but the pain is a hard fact.
PS: The only way to overdose on Vitamin D is to abuse prescription-strength supplements or cod liver oil.
You're attacking a strawman; my response to a similarly confused individual here:
9 15029
http://slashdot.org/comments.pl?sid=232631&pid=18
I don't think that post-1997 admittees are underqualified, only that it's a mistake for MIT to seek the same students as the Ivies. MIT is in a niche of its own that is incredibly important from a historical perspective;
"well-roundedness" is a good thing, but is not the last word in what is valuable to an institution or to the future success of an individual. Over the long-term, MIT risks becoming an also-ran by pursuing the same admissions strategies as Harvard and Princeton, because it can never achieve the same kind of cachet.
As an MIT alum admitted prior to her regime, I'm pleased that she will be replaced. While her lies are a black eye for the institute, her admissions policies and personal philosophy had done more damage. In an attempt to admit "well-rounded" students and compete with the Harvards of the world, she chipped away at the identity that makes MIT unique: academic excellence, creativity and fun. If that makes MIT too "geeky," so be it.
Moreover, her outspokenness reduced the dignity of her position and the process. Admissions should serve the principles of the school -- period.
You could make the same argument about tangible goods that are more sophisticated than food.
As our economy has evolved food productions has rightfully shifted to countries with cheaper labor, as our intellectual goods have a more far-reaching impact.
You've made the case that the current IP regime is broken (patents too easily granted, prior art, overly long patent/copyright extension), not that IP is fundamentally broken. [You may wish to read the other branch under my first reply.]
Specifically, software and drug patents are more contentious than machine and business process patents.