Sorry for all the posts: I now really hate the "HTML formatted" box.
The standard press description is a little confusing. A good way to think about the subject is that atomic clocks are extremely good frequency standards, which incidentally makes them good time standards as well (if I have a pendulum that oscillates once per second, I can measure time by counting the number of oscillations).
The idea behind all atomic clocks is that atoms are very picky about the kinds of light they absorb and emit (that's how astronomers can tell what kinds of atoms make up stars). There are some frequencies of light that interact very strongly with any given kind of atom, and some frequencies where the light barely interacts at all. When the atom absorbs a photon, it jumps to a higher energy state, when it emits a photon, it jumps to a lower energy state.
If you look carefully at the spectrum of light that an atom absorbs or emits, you'll find that the atom isn't equally picky about every kind of transition that it can make. There are some transitions (in cesium, they're called hyperfine transitions) where the atom isn't just picky, it's positively fastidious. What you'll find is that if you want to excite these transitions, you'll have to shine light that is exactly the right frequency, plus or minus a tiny amount (the "linewidth" -- literally, if you plotted absorbtion vs. frequency, the width of the peak you would see on the graph.)
So reasoning backwards, if I'm shining a laser at a cavity of cesium atoms and I measure that they're strongly absorbing the light, then I know the frequency of the laser has to be *exactly* the frequency that excites the atom, plus or minus a tiny linewidth. So I can count the oscillations of my laser and figure out how much time has elapsed. That's basically how an atomic clock works.
But if you wanted to get really anal about it, you could point out that I really don't know the exact frequency of my laser at all -- all I know is that it's the frequency of the atomic transition, plus or minus the linewidth. So if a hypothetical Alice and Bob in adjacent laboratories had lasers locked to the same transition, it's possible that Alice could have her laser locked at the atomic transition frequency minus a linewidth, while Bob has his locked at the atomic transition frequency plus a linewidth. (The linewidth was chosen to be really small, but it's still not zero: also, really narrow lines are hard to lock a laser to, so there's always a tradeoff involved.) So Alice and Bob's clocks will drift a tiny bit relative to each other. But because the linewidth is so small, it will take an insane number of oscillations before Bob measures that one more second has passed than Alice measures. The "1 second in 30 billion years" is just a reflection of this: it measures the linewidth of the transition relative to the frequency of light involved.
The appeal of using Mercury or Strontium atoms (small world: one of my best friends is also working on a Strontium time standard) is that they have a special transition that is even narrower relative to the transition frequency than Cesium's hyperfine transition.
New submission: now with super paragraph breaks!
The standard press description is a little confusing. A good way to think about the subject is that atomic clocks are extremely good frequency standards, which incidentally makes them good time standards as well (if I have a pendulum that oscillates once per second, I can measure time by counting the number of oscillations).
The idea behind all atomic clocks is that atoms are very picky about the kinds of light they absorb and emit (that's how astronomers can tell what kinds of atoms make up stars). There are some frequencies of light that interact very strongly with any given kind of atom, and some frequencies where the light barely interacts at all. When the atom absorbs a photon, it jumps to a higher energy state, when it emits a photon, it jumps to a lower energy state.
If you look carefully at the spectrum of light that an atom absorbs or emits, you'll find that the atom isn't equally picky about every kind of transition that it can make. There are some transitions (in cesium, they're called hyperfine transitions) where the atom isn't just picky, it's positively fastidious. What you'll find is that if you want to excite these transitions, you'll have to shine light that is exactly the right frequency, plus or minus a tiny amount (the "linewidth" -- literally, if you plotted absorbtion vs. frequency, the width of the peak you would see on the graph.)
So reasoning backwards, if I'm shining a laser at a cavity of cesium atoms and I measure that they're strongly absorbing the light, then I know the frequency of the laser has to be *exactly* the frequency that excites the atom, plus or minus a tiny linewidth. So I can count the oscillations of my laser and figure out how much time has elapsed. That's basically how an atomic clock works.
But if you wanted to get really anal about it, you could point out that I really don't know the exact frequency of my laser at all -- all I know is that it's the frequency of the atomic transition, plus or minus the linewidth. So if a hypothetical Alice and Bob in adjacent laboratories had lasers locked to the same transition, it's possible that Alice could have her laser locked at the atomic transition frequency minus a linewidth, while Bob has his locked at the atomic transition frequency plus a linewidth. (The linewidth was chosen to be really small, but it's still not zero: also, really narrow lines are hard to lock a laser to, so there's always a tradeoff involved.) So Alice and Bob's clocks will drift a tiny bit relative to each other. But because the linewidth is so small, it will take an insane number of oscillations before Bob measures that one more second has passed than Alice measures. The "1 second in 30 billion years" is just a reflection of this: it measures the linewidth of the transition relative to the frequency of light involved.
The appeal of using Mercury or Strontium atoms (small world: one of my best friends is also working on a Strontium time standard) is that they have a special transition that is even narrower relative to the transition frequency than Cesium's hyperfine transition.
The standard press description is a little confusing. A good way to think about the subject is that atomic clocks are extremely good frequency standards, which incidentally makes them good time standards as well (if I have a pendulum that oscillates once per second, I can measure time by counting the number of oscillations).
The idea behind all atomic clocks is that atoms are very picky about the kinds of light they absorb and emit (that's how astronomers can tell what kinds of atoms make up stars). There are some frequencies of light that interact very strongly with any given kind of atom, and some frequencies where the light barely interacts at all. When the atom absorbs a photon, it jumps to a higher energy state, when it emits a photon, it jumps to a lower energy state.
If you look carefully at the spectrum of light that an atom absorbs or emits, you'll find that the atom isn't equally picky about every kind of transition that it can make. There are some transitions (in cesium, they're called hyperfine transitions) where the atom isn't just picky, it's positively fastidious. What you'll find is that if you want to excite these transitions, you'll have to shine light that is exactly the right frequency, plus or minus a tiny amount (the "linewidth" -- literally, if you plotted absorbtion vs. frequency, the width of the peak you would see on the graph.)
So reasoning backwards, if I'm shining a laser at a cavity of cesium atoms and I measure that they're strongly absorbing the light, then I know the frequency of the laser has to be *exactly* the frequency that excites the atom, plus or minus a tiny linewidth. So I can count the oscillations of my laser and figure out how much time has elapsed. That's basically how an atomic clock works.
But if you wanted to get really anal about it, you could point out that I really don't know the exact frequency of my laser at all -- all I know is that it's the frequency of the atomic transition, plus or minus the linewidth. So if a hypothetical Alice and Bob in adjacent laboratories had lasers locked to the same transition, it's possible that Alice could have her laser locked at the atomic transition frequency minus a linewidth, while Bob has his locked at the atomic transition frequency plus a linewidth. (The linewidth was chosen to be really small, but it's still not zero: also, really narrow lines are hard to lock a laser to, so there's always a tradeoff involved.) So Alice and Bob's clocks will drift a tiny bit relative to each other. But because the linewidth is so small, it will take an insane number of oscillations before Bob measures that one more second has passed than Alice measures. The "1 second in 30 billion years" is just a reflection of this: it measures the linewidth of the transition relative to the frequency of light involved.
The appeal of using Mercury or Strontium atoms (small world: one of my best friends is also working on a Strontium time standard) is that they have a special transition that is even narrower relative to the transition frequency than Cesium's hyperfine transition.
Was the disparity between the areas you'd consider important if your only source of information was popular science (ie, most people until their couple years of college) and the areas considered important by scientists themselves.
For example, my scientific "grandfather" (advisor's advisor) Ugo Fano wrote a tremendously significant paper that got ranked here at #3. Yet I'd never heard of the man before grad school.
A moderately annoying, but extremely common procedure when I'm browsing is to have a specific destination in mind, say Baseball Primer
http://www.baseballthinkfactory.org/files/primer/
Now, because this has a lot of discussions, when I start typing basebal... I get a lot of urls in the autocompletion field like
http://www.baseballthinkfactory.org/files/primer/o racle/
or even unrelated baseball sites. So it's not uncommon for me to have to press downarrow several times. A very useful application of machine learning would be to order the autocompletion possibilities so that my average number of downarrow presses is minimized.
It seems to me that there's another problem with computerized voting machines that hasn't been touched on much.
While it's possible to design a computerized system with a verifiable audit trail, paper records, what have you; the techniques used to ensure that these work properly are not intuitive and are difficult to explain to an interested bystander. Think of trying to explain an MD5 sum to your mother.
In contrast, during the Florida recount *everyone* knew the difference between a hanging chad and a pregnant chad, and most people had passionate opinions about how well they conveyed a voter's intent. Same thing with the infamous "butterfly ballot" in Palm Beach.
I have this nightmare of the next disputed presidential race ending up like the breast implant lawsuits. One side is totally, utterly wrong, but their "experts" sound just as confusing as the other side's experts, and all the gobbledegook provides just enough of a fig leaf for people to believe what they want to believe anyway.
For that reason, I won't support any voting system unless it passes the "USA Today" test -- every important aspect of the system should be completely explainable by one of those infographics that USA Today likes to put on its front page -- ideally, the only words needed would be captions at the bottom.
Punch card systems, despite their problems, pass the USA Today test. So do optical (scantron) systems. Lever voting boxes -- even voting by raised hands. Computerized systems fail miserably.
First, I'd like to note that my group's funding, so far as I know, doesn't involve anything from the DOD. I don't know about any of the other groups at my institute (JILA - see jilawww.colorado.edu), because it basically doesn't matter -- people write their *own* grant applications, after all, including choosing where to apply for them. If you read the article and disregard the rhetoric about "polluted" money, you'll find that it says basically this -- people do neat stuff, and the Pentagon offers to give them money.
I think this process is at least as much a beaurocratic effect as anything else -- the Pentagon has a big budget for research, but there's only so much research you can really do that has direct military applications, so they protect the budget by funding a lot of stuff that might not be militarily important in any obvious way but is pretty neat and could pay off in the long run (again, this is pretty much exactly what the director of DARPA said in the article). The effect is a lot of stuff getting funded that has immediate, concrete civilian benefits (example -- the IR map of the Milky Way, funded by SDI) but military benefits which are less immediate, less concrete, and further off (they never actually built SDI, so the military payoff is a little nebulous, but the papers still got published).
To give an example without the emotional baggage of military funding, consider that I worked last summer on a NASA grant. The area I worked on consisted of field theory applied to ultracold quantum gases. Now, to be perfectly honest, I have no idea how ultracold quantum gases might prove useful to NASA. Going one step further, I have no idea whether there's anything ultracold gases might be useful to which might be useful to NASA. Now, if you were a Village Voice reporter describing my summer, would you say that I was in the gray area of the space-industrial complex, or would you conclude that NASA chose to fund something that didn't directly relate to space, but which is interesting nonetheless?
And I believe that when Joseph Fourier presented his work to the academy of sciences showing that any function could be represented as an infinite sum of sine and cosine functions, the result was a big yawn from everyone.
Actually, Fourier's proof was extremely controversial at the time, and has arguably had a larger impact on the subsequent development of mathematics than anything else in the 19th century not invented by Gauss.
Consider a square wave. It's a discontinuous function that by Fourier's theorem can be represented as an infinite series of continuous functions -- and yet it's trivial to show that any sum of continuous functions must itself be continuous. So which is it -- continuous or discontinuous?
The problem in this specific instance results from a failure to distinguish between pointwise convergence (looks at local behavior -- whether two functions give the same answer at the same point) and functional convergence (loosely, that the functions behave the same over the entire range being considered). But the real problem was that there was enough slop in 19th century definitions and standards of proof that it was possible to "prove" a theorem true or false using equally valid arguments.
There were other problems cropping up at the same time, of course, but the problems of Fourier analysis were a major if not the major cause of the movement for rigor that redefined math in the 20th century.
Connecting all this to things the average Slashdotter will have heard of, the famous Hilbert program was a prominent part of the movement towards rigor -- a series of important questions that had to be answered if rigor were to be possible. Goedel's Incompleteness Theorem and the Turing machine were both answers to Hilbert problems.
My mom publishes a small-town newspaper, and I was surprised to find out how little of the revenue comes from subscriptions -- barely enough to pay for the paper it's printed on(!) Almost all of the take-home money comes from ad sales.
I took a class that focused on the history of newspapers, and found out that there was a big shift in the 1850s toward this model of business -- early newspapers were more or less supported by the money that readers paid the publisher. With the advent of national advertising (which was actually seen as a big democratizing influence at time, although it was very controversial for two or three decades) and cheaper printing technologies, the economies shifted, to where larger circulation (and hence ad revenues) dominated the money that could be made through extracting a profit per physical paper sold.
This is a good reason to be skeptical when people criticise the free content/supported by ads business model -- it's actually much closer to the model embraced by the (successful) newspapers than a subscription based system.
If you're looking for somebody *really* new, check out Ken MacLeod. _The Star Fraction_ and _The Sky Road_ are both different than anything else you're likely to find out there.
If you're looking for somebody newer than Heinlein, Asimov, etc, Kim Stanley Robinson is a good bet. Of course, the Mars series probably makes him so well known that he doesn't really qualify as "new".
But if you really want to be blown out of your seat, check out Michael Chabon. On the one hand, he doesn't write science fiction, so he doesn't fit into your criteria, but on the other hand, he's quite possibly the best young writer anywhere, in any genre. I've ended up buying a new copy of _The Amazing Adventures of Kavalier and Clay_ every three or four months since it came out, just from people borrowing it and loving it so much they forget to give it back.
I've been using the beta version of 1.2 for a week or so. Here are my thoughts:
1)The searches are considerably faster.
I'm a big pine fan, but evolution won me over on the basis of a single feature: the ability to search large folders quickly. I know it's possible to grep a mail directory, and I've even done so in the past, but the ease (and speed) of searches in evolution is so much greater that it effectively gives you a capability you didn't have before. This is astoundingly useful. For example, if I search my mail folder (28,776 messages) for "crackbaby," it takes 7 seconds to find the single message containing that word (somehow, I'm saddened that it was so few.)
As long as searches keep getting faster, evolution will keep getting better.
2)Bringing up new windows still takes a while, especially when the program has been running for a few days.
3)I'm a little disgusted by the fact that they've changed the key for going to the next unread message from 'n' to '.'. From what I've read on the developer's list, this was a big item of debate, and was ultimately won by the camp that wants the interface to be as natural for Outlook users as possible. It still sucks for us pine guys.
3)Nitpicking, but they need to add a keyboard shortcut for "Reply to List." As I understand Ximian's strategy, a large portion of the audience they target (at least for Connector sales) are the professionals who need to have two computers on their desk -- Linux to do all their work, Microsoft for things like email & word processing. Just my own opinion, but I'd expect such people to be disproportionately subscribed to high-volume lists. (Anybody with better information than my own, please respond).
All in all, I see 1.2 as a nice improvement, except for one or two nitpicks. Keep it up, Ximian!
I've always been attracted to stories like this, in about the same way that people slow down by accidents on the freeway. What gets me is how obvious the scam is when you're not the target, yet how unbelievably effective the scammers can be.
From the site, _Latest Developments at the Nashville Superspeedway Demo,_ we see:
Friday, September 6
* Tilley Foundation is ready to go for the Saturday, Sept. 7 demonstration of their car at the Nashville Super Speedway. They plan to drive 700 miles at highway and race speeds without ever stopping to recharge their batters, though they will change drivers a few times, under tight scrutiny. The batteries are recharged by a proprietary internal process. 10,000 people expected to show up to watch. (Sept. 6, 2002)
A big claim. Watch how the actual test diverges from this description.
Saturday, September 7
*~8:40 am. In the course of 10 minutes, the car began making laps, driven by Carl Tilley. It did about 13 laps at 70 mph, stopped three or four times at pit for some reason. Apparently [retrospect] Tilley could tell there was something not going well with the drive train (related to bearing going out).
Gosh, the car stopped in the pit "for some reason." I wonder what's happening while it's in there. I'm sure they'll be under "tight scrutiny," though.
*~8:50 am. The car is stopped in the pit. Channel 5, local TV affiliate is at the event. The mechanics are checking under the car -- something about an axle bearing going out. They are measuring the voltage on the batteries. The array of 12 batteries charged at 160 volts this morning, and coming back into the pit now, they measure at 139 volts. Tilley said something about six volts were lost during start-up.
*~8:55 am. As the car is sitting there in the pit, with people watching on, the voltage is coming up on the batteries, up to 140.4. "It's like it is recharging from the sky or something!" reports Ken. Tilley reports that the voltages even out on an open stretch according to previous test drives.
Yeah, "or something." Notice how the voltage goes down when the car is making laps, and goes up when it's in the pit. As compwizrd points out, a battery that is under heavy load will recover a little when the load is removed. My own guess was that they'd use the pit time to put a "voltmeter" on the batteries, where the voltmeter had a concealed charger. If they were really planning to go 100 miles, that's one way to pull off the scam.
*9:20 am. Tilley was driving 95 around a corner and he heard something pop. The chasis/bearings are going out. The suspension on the car is not equipped to handle this kind of track. It puts too much pressure on the bearings. "The car is not going to move again today." Race cars have a special kind of suspension to handle tracks at high speed, so it won't put a bind on the bearings. The curves are banked. Regular cars not designed to handle that -- even the DeLorean. "It's a street car, even though it looks like a race car." The car is a 1981. It's just too much for it. The batteries stayed up, kept it charged.
*9:45 am. The demo is temporarily halted due to circumstances unrelated to the technology itself. The wheel bearing is relatively simple to replace, and may take a couple of hours; then the car can be up and running again on the track to continue the demonstration. DeLorean owner at the track said that "this happens all the time" with his car. Once, he had a wheel bearing go out twice in one month.
A nice all-purpose out clause. The test is cancelled "due to circumstances unrelated to the technology itself." If you were a potential backer at this demonstration, you'd be pretty pumped at this point. All you saw was a cool-looking car going really fast for a while, and a mysteriously increasing voltage during the pit stops.
*~10:00 am. Jan Roos, a mechanical engineer who flew down from Massachusetts as a freelance consultant for Channel 5 news in Nashville, comments that even with this demonstration of 19-20 miles, the car has traveled two thirds the distance expected to be achieved by a twelve-12-vold battery array electric vehicle, taking into consideration the 3,000 pound car, with its aerodynamic shape, going at 60 mph.
the truth comes out...
He doesn't see why the car couldn't get going again today (pending repair of the wheel bearing), so the car could surpass the 52 mile "max" point that computer models indicate given current battery and motor capabilities.
And goes right back in again. When you want to believe, there's no convincing you otherwise.
I think this is related to today's other yahoo censorship story. They told the Chinese they were "ridding the world of 'eval'" and somebody got a charachter wrong...
As for your ideas about American history, you really need some schooling. The United States is currently the oldest continuous government in the world. There are older nations to be sure, but none of the governments that were in power in 1788 are still around today in the same form they were then.
Wow. Good thing we've got studmuffins like this modded up to +4. Let's see, in 1783 the US signed a peace treaty recognizing its independence with (Doh!) England.
The last time you could really say England underwent a change in the form of its government (as opposed to a shift in power within the same general outline) would be the Glorious Revolution, which shifted the balance of power from the Court to Parliament for good. The Glorious Revolution, last time I checked (i.e. about 10 seconds ago on Google) occurred in 1688.
This has inspired me to launch an artificial intelligence project of my own. I call it AFF -- Artificial Firefighter.
The AFF consists of a 200 lb. sack of cement that sits on a couch in front of the firehouse television.
This elegant solution occurred to me when I realized that less than 1 percent of a firefighter's time is spent actually fighting fires. Thus, the AFF is just as effective as a fully and rigorously trained human more than 99 percent of the time!
With a whopping 0 lines of code, I believe the AFF to be the ne plus ultra of minimalist robotics.
General Relativity actually does play a role when you're looking at frequencies very, very precisely (i.e. much more precisely than your computer's clock will allow). But you want to be careful -- the folks at NIST are much more on the ball than you give them credit for.
A story...
I work at JILA, which is run jointly by the University of Colorado at Boulder and NIST. A friend of mine works in an ultrafast laser lab, and one day in passing he mentioned that some guys in his lab were seeing general relativistic effects when comparing their laser frequency with NIST.
Now an ultrafast laser is a buff resonator. It's "ultrafast," after all. But NIST is only something like a mile away from the JILA tower. And even in Colorado, a mile's distance just doesn't give you that much change in elevation. So I was pretty skeptical, until my friend pointed out that the official time standard is set at sea level. So even though the US time standard is here in Boulder, the time signals they broadcast are actually adjusted to sync with another clock at sea level (Maryland, I think). So even though the actual difference between the two resonators was probably less than a hundred feet, the laser guys could see general relativistic effects between their laser and the broadcast frequencies.
Moral of the story: "official time" is a little more complex than you think, and the NIST guys are really on the ball.
I haven't read the book, so I'm genuinely curious. Has Wolfram solved any nontrivial outstanding problems with his methods?
Otherwise, it seems like Wolfram is getting praised on the "Galileo's wetnurse" principle -- not for doing anything noteworthy himself, but rather for influencing others who might/might not do something worthwhile.
Speaking as a physics grad student who actually did my undergraduate thesis on the intersection of General Relativity and Quantum Mechanics (a subject cranks love to target), *everybody* has a theory about the fundamental underlying principles of the universe. Very few people have the ability to explain a single phenomenon rigorously and well. Save your praise for the latter.
I think this is a wonderful innovation, and the naysayers are just bitter. I can't think of an invention more useful for the times when I have to park my autogyro more than half a kilometer away from my Esperanto club meetings.
I wouldn't make any long term plans based on this paper. The "one chance in 400" is misleading -- if you look at the paper, what it's really saying is that their experimental result differed from their theoretical result by three standard deviations (three sigma). On the face of it, this
isn't very impressive. The trouble with straightforward statistical analysis in this fasion is that particle physics is hard. Experiments are being done at the limits of detectability, and often in ways that have never been done before. Because of this, it's extremely hard to tell what one sigma is, since it's entirely possible (and somewhat likely) that you just don't understand the pitfalls yet. Particle physicists have a rule of thumb for cases like this: a six sigma effect pans out about half the time. This is only a three sigma result, so adjust your expectations accordingly.
A result like this is worth publishing, but won't persuade many people unless followup experiments get the same results (with *much* better statistics).
Slashdot Mirror
Sorry for all the posts: I now really hate the "HTML formatted" box.
The standard press description is a little confusing. A good way to think about the subject is that atomic clocks are extremely good frequency standards, which incidentally makes them good time standards as well (if I have a pendulum that oscillates once per second, I can measure time by counting the number of oscillations).
The idea behind all atomic clocks is that atoms are very picky about the kinds of light they absorb and emit (that's how astronomers can tell what kinds of atoms make up stars). There are some frequencies of light that interact very strongly with any given kind of atom, and some frequencies where the light barely interacts at all. When the atom absorbs a photon, it jumps to a higher energy state, when it emits a photon, it jumps to a lower energy state.
If you look carefully at the spectrum of light that an atom absorbs or emits, you'll find that the atom isn't equally picky about every kind of transition that it can make. There are some transitions (in cesium, they're called hyperfine transitions) where the atom isn't just picky, it's positively fastidious. What you'll find is that if you want to excite these transitions, you'll have to shine light that is exactly the right frequency, plus or minus a tiny amount (the "linewidth" -- literally, if you plotted absorbtion vs. frequency, the width of the peak you would see on the graph.)
So reasoning backwards, if I'm shining a laser at a cavity of cesium atoms and I measure that they're strongly absorbing the light, then I know the frequency of the laser has to be *exactly* the frequency that excites the atom, plus or minus a tiny linewidth. So I can count the oscillations of my laser and figure out how much time has elapsed. That's basically how an atomic clock works.
But if you wanted to get really anal about it, you could point out that I really don't know the exact frequency of my laser at all -- all I know is that it's the frequency of the atomic transition, plus or minus the linewidth. So if a hypothetical Alice and Bob in adjacent laboratories had lasers locked to the same transition, it's possible that Alice could have her laser locked at the atomic transition frequency minus a linewidth, while Bob has his locked at the atomic transition frequency plus a linewidth. (The linewidth was chosen to be really small, but it's still not zero: also, really narrow lines are hard to lock a laser to, so there's always a tradeoff involved.) So Alice and Bob's clocks will drift a tiny bit relative to each other. But because the linewidth is so small, it will take an insane number of oscillations before Bob measures that one more second has passed than Alice measures. The "1 second in 30 billion years" is just a reflection of this: it measures the linewidth of the transition relative to the frequency of light involved.
The appeal of using Mercury or Strontium atoms (small world: one of my best friends is also working on a Strontium time standard) is that they have a special transition that is even narrower relative to the transition frequency than Cesium's hyperfine transition.
New submission: now with super paragraph breaks! The standard press description is a little confusing. A good way to think about the subject is that atomic clocks are extremely good frequency standards, which incidentally makes them good time standards as well (if I have a pendulum that oscillates once per second, I can measure time by counting the number of oscillations). The idea behind all atomic clocks is that atoms are very picky about the kinds of light they absorb and emit (that's how astronomers can tell what kinds of atoms make up stars). There are some frequencies of light that interact very strongly with any given kind of atom, and some frequencies where the light barely interacts at all. When the atom absorbs a photon, it jumps to a higher energy state, when it emits a photon, it jumps to a lower energy state. If you look carefully at the spectrum of light that an atom absorbs or emits, you'll find that the atom isn't equally picky about every kind of transition that it can make. There are some transitions (in cesium, they're called hyperfine transitions) where the atom isn't just picky, it's positively fastidious. What you'll find is that if you want to excite these transitions, you'll have to shine light that is exactly the right frequency, plus or minus a tiny amount (the "linewidth" -- literally, if you plotted absorbtion vs. frequency, the width of the peak you would see on the graph.) So reasoning backwards, if I'm shining a laser at a cavity of cesium atoms and I measure that they're strongly absorbing the light, then I know the frequency of the laser has to be *exactly* the frequency that excites the atom, plus or minus a tiny linewidth. So I can count the oscillations of my laser and figure out how much time has elapsed. That's basically how an atomic clock works. But if you wanted to get really anal about it, you could point out that I really don't know the exact frequency of my laser at all -- all I know is that it's the frequency of the atomic transition, plus or minus the linewidth. So if a hypothetical Alice and Bob in adjacent laboratories had lasers locked to the same transition, it's possible that Alice could have her laser locked at the atomic transition frequency minus a linewidth, while Bob has his locked at the atomic transition frequency plus a linewidth. (The linewidth was chosen to be really small, but it's still not zero: also, really narrow lines are hard to lock a laser to, so there's always a tradeoff involved.) So Alice and Bob's clocks will drift a tiny bit relative to each other. But because the linewidth is so small, it will take an insane number of oscillations before Bob measures that one more second has passed than Alice measures. The "1 second in 30 billion years" is just a reflection of this: it measures the linewidth of the transition relative to the frequency of light involved. The appeal of using Mercury or Strontium atoms (small world: one of my best friends is also working on a Strontium time standard) is that they have a special transition that is even narrower relative to the transition frequency than Cesium's hyperfine transition.
The standard press description is a little confusing. A good way to think about the subject is that atomic clocks are extremely good frequency standards, which incidentally makes them good time standards as well (if I have a pendulum that oscillates once per second, I can measure time by counting the number of oscillations). The idea behind all atomic clocks is that atoms are very picky about the kinds of light they absorb and emit (that's how astronomers can tell what kinds of atoms make up stars). There are some frequencies of light that interact very strongly with any given kind of atom, and some frequencies where the light barely interacts at all. When the atom absorbs a photon, it jumps to a higher energy state, when it emits a photon, it jumps to a lower energy state. If you look carefully at the spectrum of light that an atom absorbs or emits, you'll find that the atom isn't equally picky about every kind of transition that it can make. There are some transitions (in cesium, they're called hyperfine transitions) where the atom isn't just picky, it's positively fastidious. What you'll find is that if you want to excite these transitions, you'll have to shine light that is exactly the right frequency, plus or minus a tiny amount (the "linewidth" -- literally, if you plotted absorbtion vs. frequency, the width of the peak you would see on the graph.) So reasoning backwards, if I'm shining a laser at a cavity of cesium atoms and I measure that they're strongly absorbing the light, then I know the frequency of the laser has to be *exactly* the frequency that excites the atom, plus or minus a tiny linewidth. So I can count the oscillations of my laser and figure out how much time has elapsed. That's basically how an atomic clock works. But if you wanted to get really anal about it, you could point out that I really don't know the exact frequency of my laser at all -- all I know is that it's the frequency of the atomic transition, plus or minus the linewidth. So if a hypothetical Alice and Bob in adjacent laboratories had lasers locked to the same transition, it's possible that Alice could have her laser locked at the atomic transition frequency minus a linewidth, while Bob has his locked at the atomic transition frequency plus a linewidth. (The linewidth was chosen to be really small, but it's still not zero: also, really narrow lines are hard to lock a laser to, so there's always a tradeoff involved.) So Alice and Bob's clocks will drift a tiny bit relative to each other. But because the linewidth is so small, it will take an insane number of oscillations before Bob measures that one more second has passed than Alice measures. The "1 second in 30 billion years" is just a reflection of this: it measures the linewidth of the transition relative to the frequency of light involved. The appeal of using Mercury or Strontium atoms (small world: one of my best friends is also working on a Strontium time standard) is that they have a special transition that is even narrower relative to the transition frequency than Cesium's hyperfine transition.
Was the disparity between the areas you'd consider important if your only source of information was popular science (ie, most people until their couple years of college) and the areas considered important by scientists themselves. For example, my scientific "grandfather" (advisor's advisor) Ugo Fano wrote a tremendously significant paper that got ranked here at #3. Yet I'd never heard of the man before grad school.
Now, because this has a lot of discussions, when I start typing basebal... I get a lot of urls in the autocompletion field like http://www.baseballthinkfactory.org/files/primer/o racle/
or even unrelated baseball sites. So it's not uncommon for me to have to press downarrow several times. A very useful application of machine learning would be to order the autocompletion possibilities so that my average number of downarrow presses is minimized.
It seems to me that there's another problem with computerized voting machines that hasn't been touched on much.
While it's possible to design a computerized system with a verifiable audit trail, paper records, what have you; the techniques used to ensure that these work properly are not intuitive and are difficult to explain to an interested bystander. Think of trying to explain an MD5 sum to your mother.
In contrast, during the Florida recount *everyone* knew the difference between a hanging chad and a pregnant chad, and most people had passionate opinions about how well they conveyed a voter's intent. Same thing with the infamous "butterfly ballot" in Palm Beach.
I have this nightmare of the next disputed presidential race ending up like the breast implant lawsuits. One side is totally, utterly wrong, but their "experts" sound just as confusing as the other side's experts, and all the gobbledegook provides just enough of a fig leaf for people to believe what they want to believe anyway.
For that reason, I won't support any voting system unless it passes the "USA Today" test -- every important aspect of the system should be completely explainable by one of those infographics that USA Today likes to put on its front page -- ideally, the only words needed would be captions at the bottom.
Punch card systems, despite their problems, pass the USA Today test. So do optical (scantron) systems. Lever voting boxes -- even voting by raised hands. Computerized systems fail miserably.
First, I'd like to note that my group's funding, so far as I know, doesn't involve anything from the DOD. I don't know about any of the other groups at my institute (JILA - see jilawww.colorado.edu), because it basically doesn't matter -- people write their *own* grant applications, after all, including choosing where to apply for them. If you read the article and disregard the rhetoric about "polluted" money, you'll find that it says basically this -- people do neat stuff, and the Pentagon offers to give them money.
I think this process is at least as much a beaurocratic effect as anything else -- the Pentagon has a big budget for research, but there's only so much research you can really do that has direct military applications, so they protect the budget by funding a lot of stuff that might not be militarily important in any obvious way but is pretty neat and could pay off in the long run (again, this is pretty much exactly what the director of DARPA said in the article). The effect is a lot of stuff getting funded that has immediate, concrete civilian benefits (example -- the IR map of the Milky Way, funded by SDI) but military benefits which are less immediate, less concrete, and further off (they never actually built SDI, so the military payoff is a little nebulous, but the papers still got published).
To give an example without the emotional baggage of military funding, consider that I worked last summer on a NASA grant. The area I worked on consisted of field theory applied to ultracold quantum gases. Now, to be perfectly honest, I have no idea how ultracold quantum gases might prove useful to NASA. Going one step further, I have no idea whether there's anything ultracold gases might be useful to which might be useful to NASA. Now, if you were a Village Voice reporter describing my summer, would you say that I was in the gray area of the space-industrial complex, or would you conclude that NASA chose to fund something that didn't directly relate to space, but which is interesting nonetheless?
And I believe that when Joseph Fourier presented his work to the academy of sciences showing that any function could be represented as an infinite sum of sine and cosine functions, the result was a big yawn from everyone.
Actually, Fourier's proof was extremely controversial at the time, and has arguably had a larger impact on the subsequent development of mathematics than anything else in the 19th century not invented by Gauss.
Consider a square wave. It's a discontinuous function that by Fourier's theorem can be represented as an infinite series of continuous functions -- and yet it's trivial to show that any sum of continuous functions must itself be continuous. So which is it -- continuous or discontinuous?
The problem in this specific instance results from a failure to distinguish between pointwise convergence (looks at local behavior -- whether two functions give the same answer at the same point) and functional convergence (loosely, that the functions behave the same over the entire range being considered). But the real problem was that there was enough slop in 19th century definitions and standards of proof that it was possible to "prove" a theorem true or false using equally valid arguments.
There were other problems cropping up at the same time, of course, but the problems of Fourier analysis were a major if not the major cause of the movement for rigor that redefined math in the 20th century.
Connecting all this to things the average Slashdotter will have heard of, the famous Hilbert program was a prominent part of the movement towards rigor -- a series of important questions that had to be answered if rigor were to be possible. Goedel's Incompleteness Theorem and the Turing machine were both answers to Hilbert problems.
My mom publishes a small-town newspaper, and I was surprised to find out how little of the revenue comes from subscriptions -- barely enough to pay for the paper it's printed on(!) Almost all of the take-home money comes from ad sales.
I took a class that focused on the history of newspapers, and found out that there was a big shift in the 1850s toward this model of business -- early newspapers were more or less supported by the money that readers paid the publisher. With the advent of national advertising (which was actually seen as a big democratizing influence at time, although it was very controversial for two or three decades) and cheaper printing technologies, the economies shifted, to where larger circulation (and hence ad revenues) dominated the money that could be made through extracting a profit per physical paper sold.
This is a good reason to be skeptical when people criticise the free content/supported by ads business model -- it's actually much closer to the model embraced by the (successful) newspapers than a subscription based system.
If you're looking for somebody *really* new, check out Ken MacLeod. _The Star Fraction_ and _The Sky Road_ are both different than anything else you're likely to find out there.
If you're looking for somebody newer than Heinlein, Asimov, etc, Kim Stanley Robinson is a good bet. Of course, the Mars series probably makes him so well known that he doesn't really qualify as "new".
But if you really want to be blown out of your seat, check out Michael Chabon. On the one hand, he doesn't write science fiction, so he doesn't fit into your criteria, but on the other hand, he's quite possibly the best young writer anywhere, in any genre. I've ended up buying a new copy of _The Amazing Adventures of Kavalier and Clay_ every three or four months since it came out, just from people borrowing it and loving it so much they forget to give it back.
I've been using the beta version of 1.2 for a week or so. Here are my thoughts:
1)The searches are considerably faster.
I'm a big pine fan, but evolution won me over on the basis of a single feature: the ability to search large folders quickly. I know it's possible to grep a mail directory, and I've even done so in the past, but the ease (and speed) of searches in evolution is so much greater that it effectively gives you a capability you didn't have before. This is astoundingly useful. For example, if I search my mail folder (28,776 messages) for "crackbaby," it takes 7 seconds to find the single message containing that word (somehow, I'm saddened that it was so few.)
As long as searches keep getting faster, evolution will keep getting better.
2)Bringing up new windows still takes a while, especially when the program has been running for a few days.
3)I'm a little disgusted by the fact that they've changed the key for going to the next unread message from 'n' to '.'. From what I've read on the developer's list, this was a big item of debate, and was ultimately won by the camp that wants the interface to be as natural for Outlook users as possible. It still sucks for us pine guys.
3)Nitpicking, but they need to add a keyboard shortcut for "Reply to List." As I understand Ximian's strategy, a large portion of the audience they target (at least for Connector sales) are the professionals who need to have two computers on their desk -- Linux to do all their work, Microsoft for things like email & word processing. Just my own opinion, but I'd expect such people to be disproportionately subscribed to high-volume lists. (Anybody with better information than my own, please respond).
All in all, I see 1.2 as a nice improvement, except for one or two nitpicks. Keep it up, Ximian!
I've always been attracted to stories like this, in about the same way that people slow down by accidents on the freeway. What gets me is how obvious the scam is when you're not the target, yet how unbelievably effective the scammers can be.
From the site, _Latest Developments at the Nashville Superspeedway Demo,_ we see:
Friday, September 6
* Tilley Foundation is ready to go for the Saturday, Sept. 7 demonstration of their car at the Nashville Super Speedway. They plan to drive 700 miles at highway and race speeds without ever stopping to recharge their batters, though they will change drivers a few times, under tight scrutiny. The batteries are recharged by a proprietary internal process. 10,000 people expected to show up to watch. (Sept. 6, 2002)
A big claim. Watch how the actual test diverges from this description.
Saturday, September 7
*~8:40 am. In the course of 10 minutes, the car began making laps, driven by Carl Tilley. It did about 13 laps at 70 mph, stopped three or four times at pit for some reason. Apparently [retrospect] Tilley could tell there was something not going well with the drive train (related to bearing going out).
Gosh, the car stopped in the pit "for some reason." I wonder what's happening while it's in there. I'm sure they'll be under "tight scrutiny," though.
*~8:50 am. The car is stopped in the pit. Channel 5, local TV affiliate is at the event. The mechanics are checking under the car -- something about an axle bearing going out. They are measuring the voltage on the batteries. The array of 12 batteries charged at 160 volts this morning, and coming back into the pit now, they measure at 139 volts. Tilley said something about six volts were lost during start-up.
*~8:55 am. As the car is sitting there in the pit, with people watching on, the voltage is coming up on the batteries, up to 140.4. "It's like it is recharging from the sky or something!" reports Ken. Tilley reports that the voltages even out on an open stretch according to previous test drives.
Yeah, "or something." Notice how the voltage goes down when the car is making laps, and goes up when it's in the pit. As compwizrd points out, a battery that is under heavy load will recover a little when the load is removed. My own guess was that they'd use the pit time to put a "voltmeter" on the batteries, where the voltmeter had a concealed charger. If they were really planning to go 100 miles, that's one way to pull off the scam.
*9:20 am. Tilley was driving 95 around a corner and he heard something pop. The chasis/bearings are going out. The suspension on the car is not equipped to handle this kind of track. It puts too much pressure on the bearings. "The car is not going to move again today." Race cars have a special kind of suspension to handle tracks at high speed, so it won't put a bind on the bearings. The curves are banked. Regular cars not designed to handle that -- even the DeLorean. "It's a street car, even though it looks like a race car." The car is a 1981. It's just too much for it. The batteries stayed up, kept it charged.
*9:45 am. The demo is temporarily halted due to circumstances unrelated to the technology itself. The wheel bearing is relatively simple to replace, and may take a couple of hours; then the car can be up and running again on the track to continue the demonstration. DeLorean owner at the track said that "this happens all the time" with his car. Once, he had a wheel bearing go out twice in one month.
A nice all-purpose out clause. The test is cancelled "due to circumstances unrelated to the technology itself." If you were a potential backer at this demonstration, you'd be pretty pumped at this point. All you saw was a cool-looking car going really fast for a while, and a mysteriously increasing voltage during the pit stops.
*~10:00 am. Jan Roos, a mechanical engineer who flew down from Massachusetts as a freelance consultant for Channel 5 news in Nashville, comments that even with this demonstration of 19-20 miles, the car has traveled two thirds the distance expected to be achieved by a twelve-12-vold battery array electric vehicle, taking into consideration the 3,000 pound car, with its aerodynamic shape, going at 60 mph.
the truth comes out...
He doesn't see why the car couldn't get going again today (pending repair of the wheel bearing), so the car could surpass the 52 mile "max" point that computer models indicate given current battery and motor capabilities.
And goes right back in again. When you want to believe, there's no convincing you otherwise.
I think this is related to today's other yahoo censorship story. They told the Chinese they were "ridding the world of 'eval'" and somebody got a charachter wrong...
Wow. Good thing we've got studmuffins like this modded up to +4. Let's see, in 1783 the US signed a peace treaty recognizing its independence with (Doh!) England.
The last time you could really say England underwent a change in the form of its government (as opposed to a shift in power within the same general outline) would be the Glorious Revolution, which shifted the balance of power from the Court to Parliament for good. The Glorious Revolution, last time I checked (i.e. about 10 seconds ago on Google) occurred in 1688.
This has inspired me to launch an artificial intelligence project of my own. I call it AFF -- Artificial Firefighter.
The AFF consists of a 200 lb. sack of cement that sits on a couch in front of the firehouse television.
This elegant solution occurred to me when I realized that less than 1 percent of a firefighter's time is spent actually fighting fires. Thus, the AFF is just as effective as a fully and rigorously trained human more than 99 percent of the time!
With a whopping 0 lines of code, I believe the AFF to be the ne plus ultra of minimalist robotics.
General Relativity actually does play a role when you're looking at frequencies very, very precisely (i.e. much more precisely than your computer's clock will allow). But you want to be careful -- the folks at NIST are much more on the ball than you give them credit for.
A story...
I work at JILA, which is run jointly by the University of Colorado at Boulder and NIST. A friend of mine works in an ultrafast laser lab, and one day in passing he mentioned that some guys in his lab were seeing general relativistic effects when comparing their laser frequency with NIST.
Now an ultrafast laser is a buff resonator. It's "ultrafast," after all. But NIST is only something like a mile away from the JILA tower. And even in Colorado, a mile's distance just doesn't give you that much change in elevation. So I was pretty skeptical, until my friend pointed out that the official time standard is set at sea level. So even though the US time standard is here in Boulder, the time signals they broadcast are actually adjusted to sync with another clock at sea level (Maryland, I think). So even though the actual difference between the two resonators was probably less than a hundred feet, the laser guys could see general relativistic effects between their laser and the broadcast frequencies.
Moral of the story: "official time" is a little more complex than you think, and the NIST guys are really on the ball.
Otherwise, it seems like Wolfram is getting praised on the "Galileo's wetnurse" principle -- not for doing anything noteworthy himself, but rather for influencing others who might/might not do something worthwhile.
Speaking as a physics grad student who actually did my undergraduate thesis on the intersection of General Relativity and Quantum Mechanics (a subject cranks love to target), *everybody* has a theory about the fundamental underlying principles of the universe. Very few people have the ability to explain a single phenomenon rigorously and well. Save your praise for the latter.
Dude, I must be psychic. I feel like I saw this exact letter on /. two days ago. Oh, wait...
I think this is a wonderful innovation, and the naysayers are just bitter. I can't think of an invention more useful for the times when I have to park my autogyro more than half a kilometer away from my Esperanto club meetings.
I wouldn't make any long term plans based on this paper. The "one chance in 400" is misleading -- if you look at the paper, what it's really saying is that their experimental result differed from their theoretical result by three standard deviations (three sigma). On the face of it, this isn't very impressive. The trouble with straightforward statistical analysis in this fasion is that particle physics is hard. Experiments are being done at the limits of detectability, and often in ways that have never been done before. Because of this, it's extremely hard to tell what one sigma is, since it's entirely possible (and somewhat likely) that you just don't understand the pitfalls yet. Particle physicists have a rule of thumb for cases like this: a six sigma effect pans out about half the time. This is only a three sigma result, so adjust your expectations accordingly. A result like this is worth publishing, but won't persuade many people unless followup experiments get the same results (with *much* better statistics).