Right, but the simple idea of doing this presumes that the ISPs would go along. I think that if the cost of internet went up by much my Mother would cancel her ISP and do without. Add to that people who cancel because it is offensive, and then multiply each by thousands, and you have ISPs loosing tons of money.
there is actually a constant stream of muons, neutrons, gammas, and electrons streaming down from above. The number passing through a chip is much higher than what you calculate on a per minute basis. In general neutrons and gammas will pass through a thin item (like a chip or gas chamber) and when they do deposit energy, there will be a lot deposited right there. Muons and elections on the other hand leave a constant trail of ionized particles behind them.
The real question is, how much energy deposition is necessary to cause a bit flip. To answer this, there are many papers on the bit flip rates to various chips to various kinds of memory measured at accelerators and the like. There are also maps of cosmic ray (by particle and energy) density across the US and the world. Canada is actually right below the magnetic North and so it gets the highest flux rate because of the minimum amount of deflection, and the US is a close second.
right, but in what sense is the energy renewable. carbon from ground -> coal power plant -> fuel -> car -> air. This isn't that much different than other systems that are also non-renewable.
I've known few Ph.D. statisticians happy to use a canned procedure when they can write their own. It isn't reinventing the wheel, it's owning the knowledge of how the bits an pieces actually work. Now once they do that, it makes sense to use the canned procedure (as long as it gets the same results as the "hand rolled" procedure)--and trust me, it's out there in R.
The reason so many statisticians use R is that you feel very "close" to data, so many things are so easy to do--any general purpose programing language would either require reinventing the wheel on all the basic canned procedures or feeling "further" from the data. The cost of the general purpose programing language is that when you want to do the detail work you are left with tools that can do everything, not the stuff that's great at doing what you want quickly.
I think your point amounts to this: computer science undergrads love Python, so surely Ph.D. mathematicians will like it better than the language they already know and love. General purpose languages are great for general purpose jobs. Specific purpose languages are great for specific jobs--like if you have a Ph.D., you are probably locked in to the one field for quite a while, you should probably learn the specific purpose language.
Also, the only thing I know of that's quirky in R is how it is largely seamless "do what I mean" scoping, but then in a few cases the programmers want you to mean something else and you have to learn the function that fixes that.
hmm, so I'd have to learn some third hybrid syntax to use R in SAGE? The things that I notice most are (1) matlab style vectors (i.e. [1,2,3,4]), (2) python style objects calls, (3) remapping function names, and (4) a non-R function called in the R context. The second one is not a big deal at all until that day when you get to something that makes it impossible to do that way, the rest are just style choices that I don't love, but SAGE would probably enjoy.
I love the idea of the project and hope it goes well for you. What I'd really like to see (and maybe you already have this) is arbitrary precision math and a great symbolic editor for an excellent differentiation/integration engine. If I could have that, I'd be straight OSS for my math software--and I would be so happy. Well, maybe also a nice canned ode and pde solver that does okay for a broad range of problems.
> a <- c(1,2,3,4) > a [1] 1 2 3 4 > length(a) [1] 4 > dim(a) <- c(2,2) > a [,1] [,2] [1,] 1 3 [2,] 2 4 > t.text(0:99) Error: could not find function "t.text" > t.test(0:99)
One Sample t-test
data: 0:99 t = 17.0622, df = 99, p-value < 2.2e-16 alternative hypothesis: true mean is not equal to 0 95 percent confidence interval: 43.74349 55.25651 sample estimates: mean of x 49.5
What are possible advantages of Python? Do you think R is a, "special-purpose math language"? I don't see it as being that much different than other languages except that it has lots of goodies for statisticians.
Since R has a great user base and is so easy to add extensions to, wouldn't it make more sense to put sage inside of R? I also don't understand what it could mean to include R in sage since it's a pretty complete programing language for what it does. The great thing about R is how easy it is to write stats software in it, now all the great functions it already has--except the extensions, those are great.
Are you aware of R it's intended as a OSS version of S-plus but I'd say it's better than matlab in many ways. You have to dig into the packages to get really interesting functionality, but out of the box I prefer it to matlab.
Well, there's also the refining (preparing the fuel to be burned in the nuclear reactor) and the solvents used to do this have a global warming potential that is huge so that even small releases of a few tones thousand tones per year can catch up with the millions of tons of CO2 that do not get generated.
I don't understand, the B option is for how aggressive to be in power save. When you set it to one, you are saying, "treat electricity like its water in a desert." Locking after every read or write would be the best response to this, in my opinion. Why not back off from one a little?
plus, for all we know, this is mostly atime updates.
it seems to me that in an increasingly laptop centric world, there would be lots of though going into reducing fsync calls and managing small writes in a memory cache that the HD doesn't even get told about.
And that's why flop performance is usually reported for a very specific matrix opperation. The other one is called the theoretical speed and nobody cares about it.
What do you use that can, "keep comments, styles, formating etc straight." The only program that I know that does two of the three is LaTeX which does an excellent job of styles and formating.
No de facto--my desktop has been up for 32 days and has paged out a few million pages. Also, Apple and MS (if one of the other repliers is correct) have figured out ways to use swap to change performance. If there were a good reason not to do this, I'm sure the OS writers would figure out how to do it, but with HDs very reliable and cheap, why not optimize performance?
One of the worst things is that many schools will accept the best SAT/ACT score that you submit not because it's the best measure, but because then they get to report that SAT/ACT score in their statistics. In fact, some will even cherry pick the best verbal and the best math from the set. It's just insane. There's an economics paper on the topic, titled, "Retaking the SAT" By Vigdor and Clotfelter in the journal of human resources, 2002ish.
As someone who has a BA in chemistry from a liberal arts college, I strongly disagree that the degree didn't help me on the job market. Specifically, it got me my first job easy and in physics. Very few of my humanities major friends had that experience, and almost all of my chemistry major friends got great jobs without hardly trying.
Where exactly is that in the wikipedia article exactly? Look at Eric Hall's Taylor lecture to the Health Physics Society, he shows does down to 100 mrem (1 mSv) having effects, deflates having an threshold to a threshold nobody cares about.
mod parent up. I was a radiation physicist for 6 years and parent knows far more than the author of the article--and says exactly what I thought when I read the article.
Oh get off trying to misunderstand him. The point is that if you can see ionizing radiation, your probably in for it. The exception would be Cherenkov radiation--which might not kill you if the setup is right.
Slashdot Mirror
Did it ever occur to you that any site with CmdrTaco on it may not be right for you?
This just isn't a political reality.
The real question is, how much energy deposition is necessary to cause a bit flip. To answer this, there are many papers on the bit flip rates to various chips to various kinds of memory measured at accelerators and the like. There are also maps of cosmic ray (by particle and energy) density across the US and the world. Canada is actually right below the magnetic North and so it gets the highest flux rate because of the minimum amount of deflection, and the US is a close second.
Considering doing the same. Do you use TeX? If so, is there anything as good as TeXShop is for OS X?
right, but in what sense is the energy renewable. carbon from ground -> coal power plant -> fuel -> car -> air. This isn't that much different than other systems that are also non-renewable.
The reason so many statisticians use R is that you feel very "close" to data, so many things are so easy to do--any general purpose programing language would either require reinventing the wheel on all the basic canned procedures or feeling "further" from the data. The cost of the general purpose programing language is that when you want to do the detail work you are left with tools that can do everything, not the stuff that's great at doing what you want quickly.
Also, the only thing I know of that's quirky in R is how it is largely seamless "do what I mean" scoping, but then in a few cases the programmers want you to mean something else and you have to learn the function that fixes that.
What are possible advantages of Python? Do you think R is a, "special-purpose math language"? I don't see it as being that much different than other languages except that it has lots of goodies for statisticians.
Since R has a great user base and is so easy to add extensions to, wouldn't it make more sense to put sage inside of R? I also don't understand what it could mean to include R in sage since it's a pretty complete programing language for what it does. The great thing about R is how easy it is to write stats software in it, now all the great functions it already has--except the extensions, those are great.
Are you aware of R it's intended as a OSS version of S-plus but I'd say it's better than matlab in many ways. You have to dig into the packages to get really interesting functionality, but out of the box I prefer it to matlab.
Well, there's also the refining (preparing the fuel to be burned in the nuclear reactor) and the solvents used to do this have a global warming potential that is huge so that even small releases of a few tones thousand tones per year can catch up with the millions of tons of CO2 that do not get generated.
I don't understand, the B option is for how aggressive to be in power save. When you set it to one, you are saying, "treat electricity like its water in a desert." Locking after every read or write would be the best response to this, in my opinion. Why not back off from one a little?
it seems to me that in an increasingly laptop centric world, there would be lots of though going into reducing fsync calls and managing small writes in a memory cache that the HD doesn't even get told about.
And that's why flop performance is usually reported for a very specific matrix opperation. The other one is called the theoretical speed and nobody cares about it.
What do you use that can, "keep comments, styles, formating etc straight." The only program that I know that does two of the three is LaTeX which does an excellent job of styles and formating.
that's a good point, why don't I have the 3117 they advertise on my site
3094, using 0 MB (well, probably less than 1/2 MB). Maybe you have to have free space to get more.
No de facto--my desktop has been up for 32 days and has paged out a few million pages. Also, Apple and MS (if one of the other repliers is correct) have figured out ways to use swap to change performance. If there were a good reason not to do this, I'm sure the OS writers would figure out how to do it, but with HDs very reliable and cheap, why not optimize performance?
One of the worst things is that many schools will accept the best SAT/ACT score that you submit not because it's the best measure, but because then they get to report that SAT/ACT score in their statistics. In fact, some will even cherry pick the best verbal and the best math from the set. It's just insane. There's an economics paper on the topic, titled, "Retaking the SAT" By Vigdor and Clotfelter in the journal of human resources, 2002ish.
As someone who has a BA in chemistry from a liberal arts college, I strongly disagree that the degree didn't help me on the job market. Specifically, it got me my first job easy and in physics. Very few of my humanities major friends had that experience, and almost all of my chemistry major friends got great jobs without hardly trying.
Where exactly is that in the wikipedia article exactly? Look at Eric Hall's Taylor lecture to the Health Physics Society, he shows does down to 100 mrem (1 mSv) having effects, deflates having an threshold to a threshold nobody cares about.
Excellent job at misunderstanding me again! Cherenkov radiation, of course, can not exist without ionizing radiation.
mod parent up. I was a radiation physicist for 6 years and parent knows far more than the author of the article--and says exactly what I thought when I read the article.
Oh get off trying to misunderstand him. The point is that if you can see ionizing radiation, your probably in for it. The exception would be Cherenkov radiation--which might not kill you if the setup is right.