Yes!! Now I can find the next / tomorrow's WiFi hotspot using my laptop while sitting at a Starbucks! Totally makes sense!
As in, I'm driving another 400 miles this afternoon and would like to find the next spot to check weather conditions, road closures, construction, etc., while I sip on coffee. Beats the heck out of Mobile/Shell/etc. -- Paul
I try to unblock ads to my favorite small sites (e.g., sourceforge, slashdot, overclockers, ocforums), especially as survival is not so guaranteed for the smaller sites. -- Paul
On the civil side, if you're at fault for the accident, be prepared to face the victim in court to pay for the consequences and damages. (i.e., compensatory and punative damages) Doing dumb things that contribute to the severity of the accident tend to make you more at fault in such a thing.
I think this is where legislation and/or court precedent could really shine. Accidents are accidents, but if you lose customer data, you're responsible for any and all consequences. Doing dumb things like sending all that data around unencrypted and unprotected could be viewed as knowingly keeping your brakes in disrepair or yapping on a cell phone: they'd serve to show you were even more grossly negligent. -- Paul
I'll bite. Here are two I can think of off the top of my head:
Application one: A collection of bacterial species was selected to detect specific toxins and/or biological agents. This can quantify how they would respond, aiding in the design of a biotoxin and/or chemical detector.
Application two: Species X is known to break down chemical agent Y. A computer model is used to determine the optimal temperature, oxygen, and other environmental settings to use this species in cleanup after a chemical attack, e.g., on a water supply.
There are many reasons. For one thing, not all physical quantities are easily and accurately measurable. But in a situation like this, you can measure whatever you simulated and gain new insights, right down to the molecular level. Chemical reaction rates in individual cells, intracellular pressures and mechanical stresses, the effects of different mutation rates, the distribution of nutrient concentrations both within the cells and between them, and so forth.
Then there's the issue of experimental control. Want to test the effect of a different mutation rate in gene X? No problem. Vary that effect in your model only and don't change anything else. There's some good science to be conducted once the numerical implementation and the biophysical models have been validated.
Then there's cost. You can run many simulations and obtain a vast amount of data in the same time and for much less cost than it would take to run a few physical experiments. So, run a few hundred computer sims to determine which physical experiments are worth conducting; now physical lab time is more focused and more efficient. -- Paul
From the article: It takes five passwords to boot up my laptop and check my e-mail.
One of those passwords is over 50 characters long.
The first day he wakes up with some memory loss is going to be rough! Password-protecting your laptop is not only a good idea, but essential. But this is a just a little over the top. -- Paul
From TFP: Our original 256-bit key designs were designed to use the round function to lower the design, implementation and cryptanalysis time. However, all of our attempts were either weak against reduced round related keys attacks or were too inefficient for on-the-fly computation. As a result for this design we reduced the key size to 128-bits.
In general (not just in cryptography, which is certainly not my field), it's a good thing to have an idea how to extend an algorithm when designing it. Here, however, that doesn't seem to be the case.
Presumably, advances in computing hardware will eventually render this 128-bit algorithm unsecure, and it would be necessary to extend the algorithm to a higher-bit cipher. However, the quote above seems to indicate that they don't really know how to extend it to higher bits and still provide the necessary cryptoanalysis and implement it well. That doesn't sound like a good thing in a design.
In contrast, many other crypto algorithms are fairly easy to strengthen over time just by increasing the key size, since such algorithms already have a substantial amount of cryptoanalysis and it's known how large the keys need to be with a given amount of computational power (with known attacks, granted). I'd be curious to see whether or not the problems they encountered are insurmountable. -- Paul
From TFA, it seems clear that one of the issues is whether or not HP was (illegally) abusing the contracting as a way to get full-time employees without paying full-time benefits and taxes.
What isn't clear are the terms of the contract. Were they compensated for the lack of benefits? Even if so, did the contract violate any laws? We don't have any of these details, so it's certainly hard to say.
And even if they did have a contract (and I'd assume they did), that doesn't mean that it was a valid contract. If a company hired me as a contractor and had a clause requiring that I don't vote in Presidential elections, certainly that clause wouldn't be valid. Even if I had gone along with the clause for a few years before learning that it wasn't legal, it wouldn't have been legal. And within the statute of limitations, I'd have every right to sue for it when I learned that what had been stipulated was illegal.
So, let the judicial process do its work. If there are no merits, it'll be tossed, and contract law will be a little more precise in the future with the precedent. If there is merit, then it will benefit all contract workers out there (at least in the long term) by making their rights clearer, and it'll benefit all taxpayers as HP would better pull its own weight in paying for defense, highways, etc. -- Paul
You'll never get below ambient temperature with an aircooled heatsink. (In this house, we obey the rules of thermodynamics!) -- Paul
Re:Deck prisms and SOLF tubes
on
Sunlight in a Tube
·
· Score: 2, Interesting
I remember playing with lots of fun stuff like that at 3M when I had a brief fiber optics internship there. One of my favorites was a longish (maybe 15 feet) solid plastic tube of high refractive index and a translucent cladding, about a half inch in diameter. (Just a big multimode fiber, really.)
We'd have it all coiled up, point one end at the indoor lights, and point the other end wherever we wanted, and it made a fairly bright spot. It was pretty cool. -- Paul
In the fiber optics community, this is called a multimode fiber: a core of material with a higher index of refraction surrounding by a cladding of lower refractive index. The ratio of core radius to cladding radius is high, and so a large number of modes of EM radiation are supported (i.e., most wavelengths of light are transmitted through the fiber.)
In fact, the language is precisely that of fiber optics: at these scales, the size of the fiber core is much greater than that of the wavelength of the light, and so the ray-like properties of light dominate. (i.e., the light beams "bounce back and forth on the walls".)
In single-moded fibers, the ratio of the core radius to cladding radius is extremely low: on the order of the wavelength of the transmitted light. At this scale the wave-like nature of light dominates. (You need to characterize the behavior using Maxwell's equations, rather than simpler "bouncing" notions.)
The downside is that a multimode fiber has a high leakage and is not suitable for long-distance transmission. Fortunately, that's not a problem here, since the light only need to be transmitted on the order of meters to tens of meters. -- Paul
Reading the parts of the article on MP3 player displays really leaves me tempted:
Why not add a switch so that the display only shows when it's pressed? (I know that some players do this anyway, but not my Rio S10.) I rarely look at the screen anyway. I bet it wouldn't be that hard to open it up and solder one in... (Then again, maybe it'd be better to just try a firmware hack instead...)
I was a little fuzzy on the article on how they could dim parts of a normal LCD monitor screen, however. Isn't there only one backlight, so it's all or nothing? Are they proposing a grid of backlights instead of just one large one? Or is it that when the pixels are dimmed, the transistors use less power? That part wasn't very clear to me after reading TFA. -- Paul
At a given temperature, a gas has a certain pressure and root mean speed (norm of velocity from its kinetic energy). (A bit of calculation can show it to be (3kT/m)^(1/2), where k is Boltzmann's constant, T is temp in Kelvin, and m the gas molecule's mass.)
If the root mean square of the gas is comparable to the escape velocity (2GM/R)^(1/2), the the majority of the gas will only stick around for a few days (if v_{esc} / v_{rms} is around 1), or maybe a few years. In fact, for the majority of the gas to be retained by the planet for several billion years, we need v_{esc} / v_{rms} around 10 or more.
It turns out that v_{esc} / v_{rms} for Mars for most gases is too low. Water, ammonia, and methane, as well as helium and hydrogen are too light to be retained for long. (Although it turns out that water is just a bit too light, so it might stick around for thousands or millions or years.) However, it does appear oxygen, nitrogen, and carbon dioxide might be just heavy enough to be retained.
This means that if there had ever been a significant amount of liquid water on Mars, it would not have stuck around long. CO2, and O2, on the other hand, have a shot. (So I guess we could design a breathable atmosphere, but water would be a problem.)
Interestingly enough, these figures change (for the worse) if temperature increases on Mars (increases the kinetic energy of the gases), so making Mars more hospitable, temperature-wise, may make it less long-term hospitable, desirable molecule-wise.
I got a lot of this info from my undergrad astronomy/astrophysics text: Introductory Astronomy and Astrophysics, 4th ed, by Zeilik and Gregory. -- Paul
Who on earth needs a terabyte of storage? And more importantly, Why would we want it on a non-hard disk. The massive storage would be so much better on a hard disk. I can't imagine wanting to carry a terabyte with me on a disk!
Anybody who does scientific work, for instance.
It's not hard to generate a few GB of data in a fluid mechanics simulation. People doing rendering (e.g., Pixar) also run into this... -- Paul
Too bad that a hard disk would be nowhere near to keeping up with a 1GB/s transfer rate. Heck, IIRC (and please correct me if I don't!) RAM would have trouble keeping up with that... -- Paul
wasn't really a cartoon-based game, but an add-on to Quake III. I just loved the [url=http://www.planetquake.com/polycount/info/qua ke3/homer/homer.shtml]Homer player model[/url] model that could be downloaded and used in the arena.
I used to start a custom game with all Homers and watch the AI characters fight. Just watching 6 Homers jumping all over the place, yelling "Woo-Hoo", or "My legs... I can't feel my legs", or doing the victory dance "I am evil Homer! I am evil Homer!" when taunting other characters just left me in tears I was laughing so hard. (Obviously, graduate school and fatigue make the stupidest things feel pretty hilarious.:) -- Paul
Actually, this might be interesting to use for underclocking. Take your Mac Mini 1.42, underclock it a bit to 1.25, and it's even less likely that the included fan will turn on. Makes it all the more of a silent computing solution.
They do this type of thing all the time at silentpcreview, although they gain a bit more (by way of silence) because they can also adjust the CPU voltage down once the clockspeed is reduced. Just a thought. -- Paul
Intel claims all their busses are "quad-pumped". Those 533mhz busses are really 133mhz.
But they still send 4 instructions per clock, versus one instruction per clock. You're not really comparing apples to apples here. If you wanted to apply the same logic, PC3200 DDR memory really only runs at 200 MHz. (Except for the fact that the 2 signals per clock make it run at a functional equivalent of 400 MHz.) Same concept.
The P4 architecture definitely has its problems, but this isn't one of them. -- Paul
And then there's a whole different class of langauges: math packages that allow programming from within. Heck, minesweeper is even included in Matlab as an example. You an define classes, new data types, etc. I know that quite a bit can be done in Mathematica and Maple as well.
It'd be interesting to see where these fit in, since they have some elements of various languages. (e.g., in Matlab, indexing starts at 1, but much notation is C-like, and some other things are very Java-like.) -- Paul
Our research group recently bought a small cluster (around 40 processors), and as the project moved forward, it found that finding a good place to put it with sufficient cooling and power infrastructure was quite a bit more costly than originally assumed.
The idea of renting a lot of computing power without bothering with these issues is very attractive. -- Paul
That is interesting. I was very tempted towards the AFL, but ultimately steered towards the more-familiar LGPL. I forgot to post one more link that I also found useful:
It gave a nice, concise grid of options with some further explanation. Too bad I didn't know of this book, though. It sounds like just what I needed! -- Paul
Slashdot Mirror
More like
Yes!! Now I can find the next / tomorrow's WiFi hotspot using my laptop while sitting at a Starbucks! Totally makes sense!
As in, I'm driving another 400 miles this afternoon and would like to find the next spot to check weather conditions, road closures, construction, etc., while I sip on coffee. Beats the heck out of Mobile/Shell/etc. -- Paul
Well, there is some truth to it.
I try to unblock ads to my favorite small sites (e.g., sourceforge, slashdot, overclockers, ocforums), especially as survival is not so guaranteed for the smaller sites. -- Paul
From the article:
"Operations at the launch facility would include development testing, pre-flight processing and flight, landing and recovery activities." -- Paul
Indeed. And that's just the criminal side.
On the civil side, if you're at fault for the accident, be prepared to face the victim in court to pay for the consequences and damages. (i.e., compensatory and punative damages) Doing dumb things that contribute to the severity of the accident tend to make you more at fault in such a thing.
I think this is where legislation and/or court precedent could really shine. Accidents are accidents, but if you lose customer data, you're responsible for any and all consequences. Doing dumb things like sending all that data around unencrypted and unprotected could be viewed as knowingly keeping your brakes in disrepair or yapping on a cell phone: they'd serve to show you were even more grossly negligent. -- Paul
I'll bite. Here are two I can think of off the top of my head:
Application one: A collection of bacterial species was selected to detect specific toxins and/or biological agents. This can quantify how they would respond, aiding in the design of a biotoxin and/or chemical detector.
Application two: Species X is known to break down chemical agent Y. A computer model is used to determine the optimal temperature, oxygen, and other environmental settings to use this species in cleanup after a chemical attack, e.g., on a water supply.
-- Paul
There are many reasons. For one thing, not all physical quantities are easily and accurately measurable. But in a situation like this, you can measure whatever you simulated and gain new insights, right down to the molecular level. Chemical reaction rates in individual cells, intracellular pressures and mechanical stresses, the effects of different mutation rates, the distribution of nutrient concentrations both within the cells and between them, and so forth.
Then there's the issue of experimental control. Want to test the effect of a different mutation rate in gene X? No problem. Vary that effect in your model only and don't change anything else. There's some good science to be conducted once the numerical implementation and the biophysical models have been validated.
Then there's cost. You can run many simulations and obtain a vast amount of data in the same time and for much less cost than it would take to run a few physical experiments. So, run a few hundred computer sims to determine which physical experiments are worth conducting; now physical lab time is more focused and more efficient. -- Paul
From the article:
It takes five passwords to boot up my laptop and check my e-mail.
One of those passwords is over 50 characters long.
The first day he wakes up with some memory loss is going to be rough! Password-protecting your laptop is not only a good idea, but essential. But this is a just a little over the top. -- Paul
From TFP:
Our original 256-bit key designs were designed to use the round function to lower the design, implementation and cryptanalysis time. However, all of our attempts were either weak against reduced round related keys attacks or were too inefficient for on-the-fly computation. As a result for this design we reduced the key size to 128-bits.
In general (not just in cryptography, which is certainly not my field), it's a good thing to have an idea how to extend an algorithm when designing it. Here, however, that doesn't seem to be the case.
Presumably, advances in computing hardware will eventually render this 128-bit algorithm unsecure, and it would be necessary to extend the algorithm to a higher-bit cipher. However, the quote above seems to indicate that they don't really know how to extend it to higher bits and still provide the necessary cryptoanalysis and implement it well. That doesn't sound like a good thing in a design.
In contrast, many other crypto algorithms are fairly easy to strengthen over time just by increasing the key size, since such algorithms already have a substantial amount of cryptoanalysis and it's known how large the keys need to be with a given amount of computational power (with known attacks, granted). I'd be curious to see whether or not the problems they encountered are insurmountable. -- Paul
From TFA, it seems clear that one of the issues is whether or not HP was (illegally) abusing the contracting as a way to get full-time employees without paying full-time benefits and taxes.
What isn't clear are the terms of the contract. Were they compensated for the lack of benefits? Even if so, did the contract violate any laws? We don't have any of these details, so it's certainly hard to say.
And even if they did have a contract (and I'd assume they did), that doesn't mean that it was a valid contract. If a company hired me as a contractor and had a clause requiring that I don't vote in Presidential elections, certainly that clause wouldn't be valid. Even if I had gone along with the clause for a few years before learning that it wasn't legal, it wouldn't have been legal. And within the statute of limitations, I'd have every right to sue for it when I learned that what had been stipulated was illegal.
So, let the judicial process do its work. If there are no merits, it'll be tossed, and contract law will be a little more precise in the future with the precedent. If there is merit, then it will benefit all contract workers out there (at least in the long term) by making their rights clearer, and it'll benefit all taxpayers as HP would better pull its own weight in paying for defense, highways, etc. -- Paul
You'll never get below ambient temperature with an aircooled heatsink. (In this house, we obey the rules of thermodynamics!) -- Paul
I remember playing with lots of fun stuff like that at 3M when I had a brief fiber optics internship there. One of my favorites was a longish (maybe 15 feet) solid plastic tube of high refractive index and a translucent cladding, about a half inch in diameter. (Just a big multimode fiber, really.)
We'd have it all coiled up, point one end at the indoor lights, and point the other end wherever we wanted, and it made a fairly bright spot. It was pretty cool. -- Paul
Mod parent up.
In the fiber optics community, this is called a multimode fiber: a core of material with a higher index of refraction surrounding by a cladding of lower refractive index. The ratio of core radius to cladding radius is high, and so a large number of modes of EM radiation are supported (i.e., most wavelengths of light are transmitted through the fiber.)
In fact, the language is precisely that of fiber optics: at these scales, the size of the fiber core is much greater than that of the wavelength of the light, and so the ray-like properties of light dominate. (i.e., the light beams "bounce back and forth on the walls".)
In single-moded fibers, the ratio of the core radius to cladding radius is extremely low: on the order of the wavelength of the transmitted light. At this scale the wave-like nature of light dominates. (You need to characterize the behavior using Maxwell's equations, rather than simpler "bouncing" notions.)
The downside is that a multimode fiber has a high leakage and is not suitable for long-distance transmission. Fortunately, that's not a problem here, since the light only need to be transmitted on the order of meters to tens of meters. -- Paul
Reading the parts of the article on MP3 player displays really leaves me tempted:
... (Then again, maybe it'd be better to just try a firmware hack instead ...)
Why not add a switch so that the display only shows when it's pressed? (I know that some players do this anyway, but not my Rio S10.) I rarely look at the screen anyway. I bet it wouldn't be that hard to open it up and solder one in
I was a little fuzzy on the article on how they could dim parts of a normal LCD monitor screen, however. Isn't there only one backlight, so it's all or nothing? Are they proposing a grid of backlights instead of just one large one? Or is it that when the pixels are dimmed, the transistors use less power? That part wasn't very clear to me after reading TFA. -- Paul
Some good things that have eaten more memory and cycles (all of which have improved the user experience, as opposed to what the summary states):
1 Programs that check your work as you go (e.g.: autocalculate on spreadsheets)
2 More help dialogs, things watching for cameras, and whatnot to smooth the user experience.
3 More use of IM and other software in the background much of the time.
4 Services running so that it's faster to sort and search files, open your favorite programs, etc.
In short, lots of stuff running to make your experience smoother, even if it doesn't look like it's doing much more.
Some bad things:
1 More viruses, etc.
2 The mandantory virus scanner that has to run in the background all the time because of (1)
3 All the crap adware that installed more than it used to be.
These are just a few of the trends I can think of . -- Paul
At a given temperature, a gas has a certain pressure and root mean speed (norm of velocity from its kinetic energy). (A bit of calculation can show it to be (3kT/m)^(1/2), where k is Boltzmann's constant, T is temp in Kelvin, and m the gas molecule's mass.)
If the root mean square of the gas is comparable to the escape velocity (2GM/R)^(1/2), the the majority of the gas will only stick around for a few days (if v_{esc} / v_{rms} is around 1), or maybe a few years. In fact, for the majority of the gas to be retained by the planet for several billion years, we need v_{esc} / v_{rms} around 10 or more.
It turns out that v_{esc} / v_{rms} for Mars for most gases is too low. Water, ammonia, and methane, as well as helium and hydrogen are too light to be retained for long. (Although it turns out that water is just a bit too light, so it might stick around for thousands or millions or years.) However, it does appear oxygen, nitrogen, and carbon dioxide might be just heavy enough to be retained.
This means that if there had ever been a significant amount of liquid water on Mars, it would not have stuck around long. CO2, and O2, on the other hand, have a shot. (So I guess we could design a breathable atmosphere, but water would be a problem.)
Interestingly enough, these figures change (for the worse) if temperature increases on Mars (increases the kinetic energy of the gases), so making Mars more hospitable, temperature-wise, may make it less long-term hospitable, desirable molecule-wise.
I got a lot of this info from my undergrad astronomy/astrophysics text: Introductory Astronomy and Astrophysics, 4th ed, by Zeilik and Gregory. -- Paul
Who on earth needs a terabyte of storage? And more importantly, Why would we want it on a non-hard disk. The massive storage would be so much better on a hard disk. I can't imagine wanting to carry a terabyte with me on a disk!
Anybody who does scientific work, for instance.
It's not hard to generate a few GB of data in a fluid mechanics simulation. People doing rendering (e.g., Pixar) also run into this ... -- Paul
Too bad that a hard disk would be nowhere near to keeping up with a 1GB/s transfer rate. Heck, IIRC (and please correct me if I don't!) RAM would have trouble keeping up with that ... -- Paul
I used to start a custom game with all Homers and watch the AI characters fight. Just watching 6 Homers jumping all over the place, yelling "Woo-Hoo", or "My legs ... I can't feel my legs", or doing the victory dance "I am evil Homer! I am evil Homer!" when taunting other characters just left me in tears I was laughing so hard. (Obviously, graduate school and fatigue make the stupidest things feel pretty hilarious. :) -- Paul
Actually, this might be interesting to use for underclocking. Take your Mac Mini 1.42, underclock it a bit to 1.25, and it's even less likely that the included fan will turn on. Makes it all the more of a silent computing solution.
They do this type of thing all the time at silentpcreview, although they gain a bit more (by way of silence) because they can also adjust the CPU voltage down once the clockspeed is reduced. Just a thought. -- Paul
Intel claims all their busses are "quad-pumped". Those 533mhz busses are really 133mhz.
But they still send 4 instructions per clock, versus one instruction per clock. You're not really comparing apples to apples here. If you wanted to apply the same logic, PC3200 DDR memory really only runs at 200 MHz. (Except for the fact that the 2 signals per clock make it run at a functional equivalent of 400 MHz.) Same concept.
The P4 architecture definitely has its problems, but this isn't one of them. -- Paul
And then there's a whole different class of langauges: math packages that allow programming from within. Heck, minesweeper is even included in Matlab as an example. You an define classes, new data types, etc. I know that quite a bit can be done in Mathematica and Maple as well.
It'd be interesting to see where these fit in, since they have some elements of various languages. (e.g., in Matlab, indexing starts at 1, but much notation is C-like, and some other things are very Java-like.) -- Paul
Indeed!
Our research group recently bought a small cluster (around 40 processors), and as the project moved forward, it found that finding a good place to put it with sufficient cooling and power infrastructure was quite a bit more costly than originally assumed.
The idea of renting a lot of computing power without bothering with these issues is very attractive. -- Paul
That is interesting. I was very tempted towards the AFL, but ultimately steered towards the more-familiar LGPL. I forgot to post one more link that I also found useful:
License Comparison
This one gives a bit more analysis per license (and groups them nicely), but doesn't have such a nice, concise grid as the other link. -- Paul
http://www.croftsoft.com/library/tutorials/opensou rce/
It gave a nice, concise grid of options with some further explanation. Too bad I didn't know of this book, though. It sounds like just what I needed! -- Paul"Because everybody deserves a second chance!" (TM) :) -- Paul