When you're too lazy to do your homework, what do you do? You cheat.
As a Genuine University Professor, who has seen a lot of cheating... I can say that this assertion is simply wrong. People cheat when they are desparate; when they don't understand what is going on, when they fear failing. The way you notice cheating is not when people do something right, but rather when they do something wrong. People who are clever enough to cheat well will probably observe that it is less trouble to apply their cleverness to simply doing the work, than to cheat effectively; cheating effectively requires considerable effort.
Petty cheating persists because it is expensive to prosecute. Think about it. In the university setting, for example, faculty are rewarded for bringing dollars and fame --- not for upholding academic standards among their students. If I catch someone cheating on an exam, it is my fervent hope that they will readily admit it, because if they don't, I have to weigh the cost of spending perhaps 40 man-hours (my own time, and others) to deal with an isolated case of petty cheating --- by someone who is almost certainly headed for a dismal grade anyway. People who cheat in class do not get good grades! Seriously! There are steps one can take to make cheating very difficult. For example, if my classes are small enough (under 24 students or so), I try to have an oral final exam. Anyone who can cheat during a one-on-one oral exam, well, they have a very special gift indeed. But I need at least a half hour for each exam, and there is no putting that work off on TA-slaves.
I have run into "malicious" cheaters, but such behavior is very rare --- at least among university students (engineers). Now, it may very well be that CEOs of modern corporations are cut from a different cloth --- Larry Ellison, for example, seems to be the very avatar of acquisitiveness --- but most people are pretty good. And flawed --- sort of like Zoyd Wheeler, in Pynchon's "Vineland."
It's fun to beat up on people who find themselves, through a moment of weakness, in a terrible fix. We have often not bothered to understand their circumstances, nor acknowledge our own role in their predicament. Ronald Reagan, for example, liked to blow hard about the Welfare Queen, a terrible creature which exists in about the same measure as Grendel.
For another example, consider the American Taliban, John Walker. He's a pretty fat target for abuse; but is it so surprising to the nerds of slashdot that someone might do the things that Walker did? And then when someone like Steve Earle writes a sympathetic song about Walker, the derision is turned up to 11. (try this google; most of the entries are either parodies or negative criticism, poisonous "patriotism" or other nonsense.)
If the real case of Walker makes you uneasy, how about the great supernerd John Hackworth in Neal Stephenson's magnificent "The Diamond Age." Hackworth wanted nothing more than to raise his daughter well; he "cheated," got caught in one lie, tried to cover it up, and wound up spectacularly entangled in a series of punishments that lasted over a decade.
So, anyway. It's fun to beat up on SCO, and McBride. One of the differences between most people who read/. and McBride is that very few/. readers would have the spine to stand up and assert something as outlandish as SCO asserts. To/. folk, the SCO business is all very abstract, there's a billion dollars and a corporation at stake... but it's not our money or our corporation. It's more like the WWF, where there is an official Bad Guy who will, at the end of the evening, get stomped by the Good Guy, for the pleasure of the viewing audience.
So, pay attention to the interesting analysis performed by Groklaw-folk, but mod yourself down if you're merely going to hurl abuse at Darl and SCO. This is a tragedy unfolding; a very human tragedy.
Anyways, how do you find a compiler bug, if you can't read the code the compiler generates?
I once found a compiler bug which I was able to demonstrate to the compiler vendor with three lines of C code. The bug was weird. Although simple to state, it took me many months to (a) realize that it was a compiler flaw, and (b) to succinctly identify the flaw. In this case it was remarkably simple: a global float which was initialized when declared was set to the absolute value of the RHS --- it could never be negative.
It was not necessary to understand any assembly to identify this bug. This is not to say that an understanding of assembly is irrelevant to identifying compiler flaws. However, as my example shows, it is not necessary to understand assembly to identify all compiler bugs.
If JimBob SixPack funds research comparing three mechanics, and he's a mechanic,and he's One of the three in the study, and "it just so happens" that the study finds JimBob SixPack is the best place to get your car serviced - that's suspicious (ie we "suspect" that the study results were influenced by the source of funding)
Some of you may recall that, 'way back in 2000, George Bush asked Dick Cheney to perform a study to figure out whom would be the best VP to run with Bush. You'll never guess what happened.
... Because Byron Acohido's stories were themselves bullshit. If you want to see why, read Michael Crichton's novel "Airframe."
Lost in all the handwringing and moralizing of Acohido's articles was the fact that he couldn't be troubled to either
(a) understand statistics, or (b) understand that you don't just change parts willy-nilly on hundred-ton flying machines.
Acohido may have won a Pulitzer, but if he was smart, he'd fly on a 737 to receive it.
And the original parent of this series is just wrong. Seattle has *plenty* of skepticism about Microsoft and Windows. Assertions to the contrary are, simply, false. My own research lab (at the UW in Electrical Engineering) has been linux exclusively since about 1996. We had been running Solaris before that, butl hardware and software cost and instant code portability made linux irresistible. MS found out about our beowulf cluster, and offered to give us enough NT for all of them --- however it was obvious that this would have been a huge setback for us. We said "no thanks" obviously.
The fact that the GPS satellites are in "space" is sort of a NASA technology. However, the other key technologies,
(*) extremely accurate clocks (*) special waveforms with nice correlation properties (*) ionospheric correction (dual frequency L1, L2) (*) tropospheric correction (pressure, water vapor) (*) general relativity correction
... this stuff was all invented either decades before NASA existed, or by different entities entirely (such as DOD or NIST). GPS has proven to be extremely useful, but attributing GPS to NASA is a bit of a stretch.
I have duplicated Pooquey's experience. In the past five years I have purchased three pentium machines (laptop and desktop). My first act upon getting them home was to wipe the Windows off the disk and load up Linux.
Given hoop jumping challenges, I could only contemplate Microsoft's good fortune in extracting nearly a kilobuck from me, for a product that I did not want in the first place.
My newest laptop is... a PowerBook. All of the UI stuff works exactly as I want, and I can still do all my unixy stuff on the command line just as I do at work (on our fleet of linux machines). There are some things missing from the Mac, but on the whole I'm very, very pleased. One of my grad students, and my sysadmin (!) got Mac laptops, too.
You show me the radar that runs at 500 GHz, and I'll eat my hat. You don't know what you're talking about. 500 GHz is *hard*. And I ought to know, I operated a "radar" at 380 GHz once upon a time, and it would fail if you looked at it crosseyed. It would take a whole week to align the optics (yes, optics) and if you bumped the grating wheel,... what's another week! And then there was the fact that the carcinotron only cost about $80k.
And it had the breathtaking output power of about 15 mW.
And oh, it turns out that a few meters of atmosphere is pretty opaque to 380 GHz, so only a jackass would build a radar at that frequency anyway.
I guess I'm crabby tonight. There's an awful lot of dreck sloshing through/. on the topic of these fast transistors. The only people who make sense are the ones who have observed that 500 GHz is the GBP, which means that these things aren't as fast as you think they are in terms of what you can actually use them for.
The weenies who are muttering about how CMOS is low power seem to be forgetting about how much power CMOS uses if you switch them fast. If you doubt me, just run a 2 GHz laptop on your lap for a half hour. Especially if you're a guy.
There are some easy-to-notice differences between meteors and missiles. Low Earth Orbit velocities are about 7.5 km/sec. Meteor velocities can be up to around 70 km/sec for objects gravitationaly bound to this solar system.
(*) escape velocity. Missiles will in general have trajectories which indicate that they must have originated on or near Earth. Meteors, on the other hand, will have trajectories that are consistent with falling from *at*least* the edge of the Earth's gravity well, and perhaps from considerably further out in the Solar System.
(*) radiation signature. Meteors emit light and heat and sound and seismic signals; nukes emit all those, and more --- energetic photons, very large electric fields, etc. "norad"-type satellites have to be able to distinguish between routine energetic events (like lightning) and "interesting" events (like nuke explosions).
Of course "easy to notice" does imply that you have radar instruments and spacecraft which are, in fact, capable of performing trajectory estimates of this rank. Probably a lot of the people who fear such attack have limited abilities in this regard.
Of course, with enough fuel, you could create meteors --- but it would be a lot of fuel to get an Earth bound object to intersect the Earth's surface at meteoric velocity in a short period of time.
My girlfriend laughs and reminds me that it's just a story,
I can't believe it! someone on Slashdot who has an actual girlfriend!
[Okay, I'm being a smartass; I'm married, have a 5 year old child, and a big dog who is snoring on his own couch across the room while I listen to Elvis Costello sing about satellites. But, in my defense, I'm running linux on a laptop that doesn't dual boot! Oops! I meant GNU/Linux!]
Boy: I sort of hate to slashdot the poor National Oceans and Atmospheres Administration, but if you'll look at
http://sec.noaa.gov/ace/MAG_SWEPAM_7d.html
you'll notice the awkward fact that the Earth gets nailed by "acoustic" / "alfvenic" shockwaves every couple of days in the solar wind. Even though there are only about 1-10 particles/cc in the solar wind, the solar wind has a readily definable temperature, speed of sound (and an "alfven speed" for the plasma geeks out there).
-----
And, please: a 100 k-ton nuke detonated a few feet above the surface of the Moon would leave an effect that would be visible to the naked eye while it happened and to big binoculars or small telescopes a year later. Nukes turn out to be sort of big. 100 k-ton is medium sized, and 10 times bigger than what was dropped on Japan.
I'm highly amused by this whole/. thread. However, some of the very finest SciFi results from explicitly violating some basic science principles, such as
"The Gods Themselves" by Asimov, and "The Practice Effect" by Brin,
which explore some rather interesting ideas resulting from weakening the 2nd law of Thermodynamics. It's fun to worship at the altar of Asimov, but the Brin story is better SciFi.
There are other things that Hams can do; back in the 1970s and 1980s there were a couple of spectacular sources of interference on HF: "SSBers" (out of band, over power, illegal CB) and "The Woodpecker" (a Soviet Over-The-Horizon radar).
Hams would routinely police the bottom of the 10 m band to sweep the SSBer's out; and those with automatic morse code keys could very effectively jam the Woodpecker, pushing it out of the 20 and 15 meter bands. I remember the slow-scan TV signals (sort of an early MPEG) as being particularly effective jammers.
Not that I ever engaged in any of that, of course, although I may have been a bit leisurely in tuning up my transmitter from time to time.
A student of mine recently created a rather wonderful PowerPoint presention on her very nice Mac laptop (I'm quite jealous). Upon learning that none shall be permitted to use their own laptops at the conference next week, she transferred her PowerPoint slides to a W2K machine to test drive the talk.
The slides fail utterly to work, despite Microsoft's assurances of complete compatibility between the W2K and Apple versions of PowerPoint.
Thanks, Microsoft! Thanks a whole big bunch!
I think I'm going to eat some asparagus, drive over to Redmond, and pee on their shrubbery.
It's true that Orwell was a communist, and it's true that he was very discouraged by what he observed in the Soviet Union (how could any "small s socialist" not be?). But Orwell's critique of Soviet Communism came in a different book, the marvelous "Animal Farm." ("All animals are equal, but some are more equal than others")
Orwell's original title for "1984" was, in fact, "1948" --- but Orwell's Western editors were too nervous about the bitter criticism of contemporary Western politics, the nascent Cold War, etc. Swapping the digits appeased the editors, and let Orwell have his little joke.
Demanding that science fiction predict the future, and then scoffing when it fails, is really a kind of ad hominem against the genre. Great science fiction need not literally predict as much as it says, "here are some possible implications of X."
One of the most famous "predictions" is that of Orwell's 1984, which (of course) has not exactly come to pass. On the other hand, many concepts of 1984 have proven tremendously robust and recognizable, such as "double speak" and "double think." You can glimpse shadows of the larger issues, such as three major world powers which engage in shifting alliances of 2 vs 1.... On the doublethink front, contemplate the fact that approximately half of US citizens think that Saddam Hussein was heavily involved in the September 11th attacks.
So, read Sterling's "Distraction" and be amazed by an enthusiastic, over-the-top speculation on trends in politics and manipulation of the public, with intriguing little sidetrips on new technology and ancient history (well, not exactly ancient --- but I found the Regulators and Moderators to be truly interesting folk; they don't need to ever come into real existence to be evocative, and to think, "well, really, just what keeps them from existing?") The whole idea of "reputation servers" is coming into existence right now, implemented by Google, blogs, and (yes) Slashdot's cooperative editing and posting system. (Not to mention USNews's annual beauty pageant for universities. The USA has such a tremendous stable of great universities, it is pretty discouraging to see a "top 10" gather so much shallow attention.)
At any rate --- concern about TIA and its kin (which should include Google, you know --- see the interview with Sterling) is perfectly legitimate, and if SciFi isn't perfectly prognostic about what it's going to mean, well, do our leaders really do any better? Does Ashcroft have a conventional understanding of the Bill of Rights?
Any think tank that wouldn't want to have a Bruce Sterling around is a think tank that's too timid to ever say anything truly mind stretching.
Hmmm. It's not obvious to me that the force of gravity will exceed the aero decelleration force on a bullet.
consider a bullet with a front cross section of 1cm by 1cm (10^-4 m^2), travelling at 800 m/s, with a total mass of 5 grams. The gravitational force will be about 9.8*0.005 = 50 mN (milli Newtons).
In one second the bullet will displace a column of air that is 800 m by 10^(-4) m = about 0.1 m^3 or about 100 liters, which is about 4 moles, or about 120 grams. Because the bullet is supersonic it cannot slip perfectly smoothly through the air.
so M(bullet) dv(bullet) = M(air) dv(air)
dv(bullet) = 0.05 * 350m/s
= about 20 m/s
I've assumed that the displaced air can be accelerated up to about the speed of sound 350 m/s...
At any rate, if the bullet is slowed by 20 m/s by air friction, that exceeds the 5 m/s change in velocity due to gravity by a factor of four.
So... anyway. Like I said. It's not obvious to me that the acceleration due to gravity exceeds the acceleration due to air friction.
-----
Try this: drop a bullet out of a car at highway speed (30 m/s). When the bullet hits the ground, how far back has it drifted do to wind pressure? Quite a bit, actually, and that's a the low speed of 30 m/s, not 800 m/s for a high velocity rifle bullet; recall that (roughly speaking) air resistance scales as the cube of the velocity.
But I don't really know anything about guns and bullets.
Now the point I am trying to make is that programmers are always defining new things. Engineers can't responsibly design systems around parts with unknown properties.
Engineering and most programming endeavors are mutually exclusive. A good engineer can't afford to have an unknown in the process.
These assertions are false. Real engineers constantly design systems around parts whose specs are unknown, for systems and circumstances which are not always well defined. The "specs" themselves may explicitly include statements of vagueness for which one can design. For example, "the power supply shall deliver within 10 percent of the nominal specs in the event of every single point failure of any component, including open or short failures between any two terminals of a device."
The gadget that my students and I have built has been designed to operate unattended for months, and gracefully accept network and power outages. We don't know why these outages occur, and we don't even have a very good idea of the statistics of their occurance. But through a combination of "experience" and conscious design we've managed to achieve considerable success, even though we are rather puzzled about the background that our system has to work within.
Sure, complete specs are a great thing to have, when you can have them, but if engineering was limited to meeting the stated specs, we'd be doomed.
Consider: the World Trade Center Towers didn't instantly fall down when they were struck at high speed by large, fully fueled airplanes. In fact, those poor buildings stayed up for about an hour *each* before collapsing. Only about 3000 people were killed. Think about the context, and consider how absolutely robust those 25 year old buildings were when presented with such a terrible insult.
The "failure" of the Towers is awful, but the way the towers failed --- slowly, after an hour --- is in fact a testimonial to some pretty amazing engineering. I think that civil and stuctural engineers should be pumping their fists in the air about the success of the graceful failure of those buildings.
At any rate, the idea that engineers can't design for the unknown is wrong. Graceful failure is something engineers can sure shoot for; graceful failure is not the same thing as "meeting spec", but it does mean that "flat tire" doesn't imply "car bursts into flames." Along those lines, modern automobiles are remarkably tough --- they are designed to accomodate people who can barely drive at all, who will not service them properly... There's a lot of pretty amazing engineering in everyday life which tolerates the unknown routinely and gracefully.
Reed's article is based on the observation that Maxwell's Equations are linear (for most materials) and that, therefore the waves pass through each other without modification (again, unless you're in pretty exotic environments --- early universe, etc.) The problem with interference arises because of imperfect spectral content and non ideal antenna response for both transmitters and receivers. Interference is like being at a party: There are a lot of people talking, and your ears hear in all directions, so you have to be near the person you're trying to talk to.
For a variation on this theme, there's an interesting moment in a movie (Frankie and Johnnie?) where there's a terrific racket in a diner, impossible to understand anything, but a cook and a clerk are communicating easily --- by sign language. Consider also those occasional TV images of the Wall Street pit traders flinging gang signs at each other... the reason that it works is that your eyes have very fine angular sensitivity (high quality antennas) compared to your ears.
Spectral purity and antenna quality limitations can be overcome --- by money. You can build higher quality receivers and transmitters, bigger antenna installations but it costs money and space in fairly unavoidable ways.
Reed is also wrong from a regulatory level. It's not just the FCC that you'd have to work with, but the ITU. Those pesky radio waves have this interesting habit of leaking over borders on the ground, and pretty much everywhere down here from satellites.
There are pretty good reasons to pick on modern broadcasting: crappy content, media concentration --- but "broadcasting" is not one of them. Those great big transmitters permit the use of very dumb receivers with poor sensitivy. The very simplicity and asymmetry of broadcast provides tremendous economic and technical appeal, and I'd be amazed if it ever went away.
Far more interesting is the glacial progress of DTV in broadcast.
Actually, internships are a *cheap* recruiting tool. The company gets to test drive a bunch of greenhorns for approximately 1/2 to 1/3 of what their "real" engineers cost, without any long term commitment to them. If they find a few keepers, then when they finally do convert them to real employees, they're already trained!
Eliminating internship programs is merely a means to save some money right now. It is *not* something that a company with an eye on the long view would do if it was serious about being in the game in the long term.
The really important ways to save money are also really difficult --- they involve layoffs, shedding of unprofitable units (which are supposed to be profitable; internship programs are not supposed to make money, of course).
-----
on a different topic: I've never heard of an unpaid internship for electrical engineers; I had (paid) interships 20 years ago, and my students certainly have been getting paid internships for the past 15 years.
When I started grad school I insisted on programming in C in a FORTRAN shop. This turned out to cause tremendous problems, and I've realized that it was a tremendous error to permit me to use a buggy C compiler to write code that none of my colleagues could understand.
Among other things, the FORTRAN compiler was reentrant, while the C compiler was not. You could have 5 people writing FORTRAN compiling merrily away, and the system would work great. But when I needed to compile, I actually had to tell everyone to take a break --- because it was going to be 5 minutes before they could do anything.
It made me a rather good C programmer, because I had to write code that would actually work without compiling blindly. But I should have used FORTRAN.
Now I'm an Extremely Powerful Professor running my own lab. To avoid being dazzled by heterogeneity, we use
linux C python xml latex
and that's just about it. My students do *not* have permission to use other tools, and they *do* have permission to abandon the legacy perl and C++ code so that we can simplify our overall code base --- at least in terms of minimizing the tools that the next students must be familiar with.
Already been done. Check out the Harris H-series computers, which have a 24 bit single precision int and a 48 bit long int.
It's a fine, fine system, in which single and double precision floats take up 48 bits, but have different precision.
You'll be especially happy to learn that the OS was called VOS for "Vulcan Operating System." The peripheral model was that "everything is a tape drive" which means that you can rewind your files.
Slashdot Mirror
When you're too lazy to do your homework, what do you do? You cheat.
... I can say that this assertion is simply wrong. People cheat when they are desparate; when they don't understand what is going on, when they fear failing. The way you notice cheating is not when people do something right, but rather when they do something wrong. People who are clever enough to cheat well will probably observe that it is less trouble to apply their cleverness to simply doing the work, than to cheat effectively; cheating effectively requires considerable effort.
/. and McBride is that very few /. readers would have the spine to stand up and assert something as outlandish as SCO asserts. To /. folk, the SCO business is all very abstract, there's a billion dollars and a corporation at stake ... but it's not our money or our corporation. It's more like the WWF, where there is an official Bad Guy who will, at the end of the evening, get stomped by the Good Guy, for the pleasure of the viewing audience.
As a Genuine University Professor, who has seen a lot of cheating
Petty cheating persists because it is expensive to prosecute. Think about it. In the university setting, for example, faculty are rewarded for bringing dollars and fame --- not for upholding academic standards among their students. If I catch someone cheating on an exam, it is my fervent hope that they will readily admit it, because if they don't, I have to weigh the cost of spending perhaps 40 man-hours (my own time, and others) to deal with an isolated case of petty cheating --- by someone who is almost certainly headed for a dismal grade anyway. People who cheat in class do not get good grades! Seriously! There are steps one can take to make cheating very difficult. For example, if my classes are small enough (under 24 students or so), I try to have an oral final exam. Anyone who can cheat during a one-on-one oral exam, well, they have a very special gift indeed. But I need at least a half hour for each exam, and there is no putting that work off on TA-slaves.
I have run into "malicious" cheaters, but such behavior is very rare --- at least among university students (engineers). Now, it may very well be that CEOs of modern corporations are cut from a different cloth --- Larry Ellison, for example, seems to be the very avatar of acquisitiveness --- but most people are pretty good. And flawed --- sort of like Zoyd Wheeler, in Pynchon's "Vineland."
It's fun to beat up on people who find themselves, through a moment of weakness, in a terrible fix. We have often not bothered to understand their circumstances, nor acknowledge our own role in their predicament. Ronald Reagan, for example, liked to blow hard about the Welfare Queen, a terrible creature which exists in about the same measure as Grendel.
For another example, consider the American Taliban, John Walker. He's a pretty fat target for abuse; but is it so surprising to the nerds of slashdot that someone might do the things that Walker did? And then when someone like Steve Earle writes a sympathetic song about Walker, the derision is turned up to 11. (try this google; most of the entries are either parodies or negative criticism, poisonous "patriotism" or other nonsense.)
If the real case of Walker makes you uneasy, how about the great supernerd John Hackworth in Neal Stephenson's magnificent "The Diamond Age." Hackworth wanted nothing more than to raise his daughter well; he "cheated," got caught in one lie, tried to cover it up, and wound up spectacularly entangled in a series of punishments that lasted over a decade.
So, anyway. It's fun to beat up on SCO, and McBride. One of the differences between most people who read
So, pay attention to the interesting analysis performed by Groklaw-folk, but mod yourself down if you're merely going to hurl abuse at Darl and SCO. This is a tragedy unfolding; a very human tragedy.
Anyways, how do you find a compiler bug, if you can't read the code the compiler generates?
I once found a compiler bug which I was able to demonstrate to the compiler vendor with three lines of C code. The bug was weird. Although simple to state, it took me many months to (a) realize that it was a compiler flaw, and (b) to succinctly identify the flaw. In this case it was remarkably simple: a global float which was initialized when declared was set to the absolute value of the RHS --- it could never be negative.
It was not necessary to understand any assembly to identify this bug. This is not to say that an understanding of assembly is irrelevant to identifying compiler flaws. However, as my example shows, it is not necessary to understand assembly to identify all compiler bugs.
If JimBob SixPack funds research comparing three mechanics, and he's a mechanic ,and he's One of the three in the study, and "it just so happens" that the study finds JimBob SixPack is the best place to get your car serviced - that's suspicious (ie we "suspect" that the study results were influenced by the source of funding)
Some of you may recall that, 'way back in 2000, George Bush asked Dick Cheney to perform a study to figure out whom would be the best VP to run with Bush. You'll never guess what happened.
... Because Byron Acohido's stories were themselves bullshit. If you want to see why, read Michael Crichton's novel "Airframe."
Lost in all the handwringing and moralizing of Acohido's articles was the fact that he couldn't be troubled to either
(a) understand statistics, or
(b) understand that you don't just change parts willy-nilly on hundred-ton flying machines.
Acohido may have won a Pulitzer, but if he was smart, he'd fly on a 737 to receive it.
And the original parent of this series is just wrong. Seattle has *plenty* of skepticism about Microsoft and Windows. Assertions to the contrary are, simply, false. My own research lab (at the UW in Electrical Engineering) has been linux exclusively since about 1996. We had been running Solaris before that, butl hardware and software cost and instant code portability made linux irresistible. MS found out about our beowulf cluster, and offered to give us enough NT for all of them --- however it was obvious that this would have been a huge setback for us. We said "no thanks" obviously.
For confirmation, check out Howard Roark's approach to sex in The Fountainhead (also by Ayn Rand, of course).
At the start of the book I thought Howard Roark was pretty estimable --- but by the end, I thought "jeez, what a jerk."
Let me offer the "John Hackworth" character of Neal
Stephenson's underappreciated masterpiece,
"The Diamond Age:
or
A Young Lady's Illustrated Primer"
(*) extremely accurate clocks
(*) special waveforms with nice correlation properties
(*) ionospheric correction (dual frequency L1, L2)
(*) tropospheric correction (pressure, water vapor)
(*) general relativity correction
I have duplicated Pooquey's experience. In the past five years I have purchased three pentium machines (laptop and desktop). My first act upon getting them home was to wipe the Windows off the disk and load up Linux.
... a PowerBook. All of the UI stuff works exactly as I want, and I can still do all my unixy stuff on the command line just as I do at work (on our fleet of linux machines). There are some things missing from the Mac, but on the whole I'm very, very pleased. One of my grad students, and my sysadmin (!) got Mac laptops, too.
Given hoop jumping challenges, I could only contemplate Microsoft's good fortune in extracting nearly a kilobuck from me, for a product that I did not want in the first place.
My newest laptop is
You show me the radar that runs at 500 GHz, and I'll eat ... what's another week! And then there was the fact that the carcinotron only cost about $80k.
/. on the topic of these fast transistors. The only people who make sense are the ones who have observed that 500 GHz is the GBP, which means that these things aren't as fast as you think they are in terms of what you can actually use them for.
my hat. You don't know what you're talking about. 500 GHz is *hard*. And I ought to know, I operated a "radar"
at 380 GHz once upon a time, and it would fail if you looked at it crosseyed. It would take a whole week to align the optics (yes, optics) and if you bumped the grating wheel,
And it had the breathtaking output power of about 15 mW.
And oh, it turns out that a few meters of atmosphere is pretty opaque to 380 GHz, so only a jackass would build a radar at that frequency anyway.
I guess I'm crabby tonight. There's an awful lot of dreck sloshing through
The weenies who are muttering about how CMOS is low power seem to be forgetting about how much power CMOS uses if you switch them fast. If you doubt me, just run a 2 GHz laptop on your lap for a half hour. Especially if you're a guy.
There are some easy-to-notice differences between meteors and missiles. Low Earth Orbit velocities are about 7.5 km/sec. Meteor velocities can be up to around 70 km/sec for objects gravitationaly bound to this solar system.
(*) escape velocity. Missiles will in general have trajectories which indicate that they must have originated on or near Earth. Meteors, on the other hand, will have trajectories that are consistent with falling from *at*least* the edge of the Earth's gravity well, and perhaps from considerably further out in the Solar System.
(*) radiation signature. Meteors emit light and heat and sound and seismic signals; nukes emit all those, and more --- energetic photons, very large electric fields, etc. "norad"-type satellites have to be able to distinguish between routine energetic events (like lightning) and "interesting" events (like nuke explosions).
Of course "easy to notice" does imply that you have radar instruments and spacecraft which are, in fact, capable of performing trajectory estimates of this rank. Probably a lot of the people who fear such attack have limited abilities in this regard.
Of course, with enough fuel, you could create meteors --- but it would be a lot of fuel to get an Earth bound object to intersect the Earth's surface at meteoric velocity in a short period of time.
My girlfriend laughs and reminds me that it's just a story,
I can't believe it! someone on Slashdot who has an actual girlfriend!
[Okay, I'm being a smartass; I'm married, have a 5 year old child, and a big dog who is snoring on his own couch across the room while I listen to Elvis Costello sing about satellites. But, in my defense, I'm running linux on a laptop that doesn't dual boot! Oops! I meant GNU/Linux!]
Boy: I sort of hate to slashdot the poor National Oceans and Atmospheres Administration, but if you'll look at
http://sec.noaa.gov/ace/MAG_SWEPAM_7d.html
you'll notice the awkward fact that the Earth gets nailed by "acoustic" / "alfvenic" shockwaves every couple of days in the solar wind. Even though there are only about 1-10 particles/cc in the solar wind, the solar wind has a readily definable temperature, speed of sound (and an "alfven speed" for the plasma geeks out there).
-----
And, please: a 100 k-ton nuke detonated a few feet above the surface of the Moon would leave an effect that would be visible to the naked eye while it happened and to big binoculars or small telescopes a year later. Nukes turn out to be sort of big. 100 k-ton is medium sized, and 10 times bigger than what was dropped on Japan.
I'm highly amused by this whole /. thread. However, some of the very finest SciFi results from explicitly violating some basic science principles, such as
"The Gods Themselves" by Asimov, and
"The Practice Effect" by Brin,
which explore some rather interesting ideas resulting from weakening the 2nd law of Thermodynamics. It's fun to worship at the altar of Asimov, but the Brin story is better SciFi.
Please, don't hurt me.
D = electric flux density, coulombs/m^2
D = epsilon E
ya sure, ya betcha there's flux in capacitors, just happens to be of the electric variety (as opposed to the magnetic variety).
There are other things that Hams can do; back in the 1970s and 1980s there were a couple of spectacular sources of interference on HF: "SSBers" (out of band, over power, illegal CB) and "The Woodpecker" (a Soviet Over-The-Horizon radar).
Hams would routinely police the bottom of the 10 m band to sweep the SSBer's out; and those with automatic morse code keys could very effectively
jam the Woodpecker, pushing it out of the 20 and 15 meter bands. I remember the slow-scan TV signals (sort of an early MPEG) as being particularly effective jammers.
Not that I ever engaged in any of that, of course, although I may have been a bit leisurely in tuning up my transmitter from time to time.
A student of mine recently created a rather wonderful PowerPoint presention on her very nice Mac laptop (I'm quite jealous). Upon learning that none shall be permitted to use their own laptops at the conference next week, she transferred her PowerPoint slides to a W2K machine to test drive the talk.
The slides fail utterly to work, despite Microsoft's assurances of complete compatibility between the W2K and Apple versions of PowerPoint.
Thanks, Microsoft! Thanks a whole big bunch!
I think I'm going to eat some asparagus, drive over to Redmond, and pee on their shrubbery.
It's true that Orwell was a communist, and it's true that he was very discouraged by what he observed in the Soviet Union (how could any "small s socialist" not be?). But Orwell's critique of Soviet Communism came in a different book, the marvelous "Animal Farm." ("All animals are equal, but some are more equal than others")
Orwell's original title for "1984" was, in fact, "1948" --- but Orwell's Western editors were too nervous about the bitter criticism of contemporary Western politics, the nascent Cold War, etc. Swapping the digits appeased the editors, and let Orwell have his little joke.
Demanding that science fiction predict the future, and then scoffing when it fails, is really a kind of ad hominem against the genre. Great science fiction need not literally predict as much as it says, "here are some possible implications of X."
... On the doublethink front, contemplate the fact that approximately half of US citizens think that Saddam Hussein was heavily involved in the September 11th attacks.
One of the most famous "predictions" is that of Orwell's 1984, which (of course) has not exactly come to pass. On the other hand, many concepts of 1984 have proven tremendously robust and recognizable, such as "double speak" and "double think." You can glimpse shadows of the larger issues, such as three major world powers which engage in shifting alliances of 2 vs 1.
So, read Sterling's "Distraction" and be amazed by an enthusiastic, over-the-top speculation on trends in politics and manipulation of the public, with intriguing little sidetrips on new technology and ancient history (well, not exactly ancient --- but I found the Regulators and Moderators to be truly interesting folk; they don't need to ever come into real existence to be evocative, and to think, "well, really, just what keeps them from existing?") The whole idea of "reputation servers" is coming into existence right now, implemented by Google, blogs, and (yes) Slashdot's
cooperative editing and posting system. (Not to mention USNews's annual beauty pageant for universities. The USA has such a tremendous stable of great universities, it is pretty discouraging to see a "top 10" gather so much shallow attention.)
At any rate --- concern about TIA and its kin (which should include Google, you know --- see the interview with Sterling) is perfectly legitimate, and if SciFi isn't perfectly prognostic about what it's going to mean, well, do our leaders really do any better? Does Ashcroft have a conventional understanding of the Bill of Rights?
Any think tank that wouldn't want to have a Bruce Sterling around is a think tank that's too timid to ever say anything truly mind stretching.
Hmmm. It's not obvious to me that the force of gravity will exceed the aero decelleration force on a bullet.
... anyway. Like I said. It's not obvious to me that the acceleration due to gravity exceeds the acceleration due to air friction.
consider a bullet with a front cross section of 1cm by 1cm (10^-4 m^2), travelling at 800 m/s, with a total mass of 5 grams. The gravitational force will be about 9.8*0.005 = 50 mN (milli Newtons).
In one second the bullet will displace a column of air that is 800 m by 10^(-4) m = about 0.1 m^3 or
about 100 liters, which is about 4 moles, or about 120 grams. Because the bullet is supersonic it
cannot slip perfectly smoothly through the air.
so M(bullet) dv(bullet) = M(air) dv(air)
dv(bullet) = 0.05 * 350m/s
= about 20 m/s
I've assumed that the displaced air can be accelerated up to about the speed of sound 350 m/s...
At any rate, if the bullet is slowed by 20 m/s by air friction, that exceeds the 5 m/s change in velocity due to gravity by a factor of four.
So
-----
Try this: drop a bullet out of a car at highway speed (30 m/s). When the bullet hits the ground, how far back has it drifted do to wind pressure? Quite a bit, actually, and that's a the low speed of 30 m/s, not 800 m/s for a high velocity rifle bullet; recall that (roughly speaking) air resistance scales as the cube of the velocity.
But I don't really know anything about guns and bullets.
Now the point I am trying to make is that programmers are always defining new things. Engineers can't responsibly design systems around parts with unknown properties.
Engineering and most programming endeavors are mutually exclusive. A good engineer can't afford to have an unknown in the process.
These assertions are false. Real engineers constantly design systems around parts whose specs are unknown, for systems and circumstances which are not always well defined. The "specs" themselves may explicitly include statements of vagueness for which one can design. For example,
"the power supply shall deliver within 10 percent of the nominal specs in the event of every single point failure of any component, including open or short failures between any two terminals of a device."
The gadget that my students and I have built has been designed to operate unattended for months, and gracefully accept network and power outages. We don't know why these outages occur, and we don't even have a very good idea of the statistics of their occurance. But through a combination of "experience" and conscious design we've managed to achieve considerable success, even though we are rather puzzled about the background that our system has to work within.
Sure, complete specs are a great thing to have, when you can have them, but if engineering was limited to meeting the stated specs, we'd be doomed.
Consider: the World Trade Center Towers didn't instantly fall down when they were struck at high speed by large, fully fueled airplanes. In fact, those poor buildings stayed up for about an hour *each* before collapsing. Only about 3000 people were killed. Think about the context, and consider how absolutely robust those 25 year old buildings were when presented with such a terrible insult.
The "failure" of the Towers is awful, but the way the towers failed --- slowly, after an hour --- is in fact a testimonial to some pretty amazing engineering. I think that civil and stuctural engineers should be pumping their fists in the air about the success of the graceful failure of those buildings.
At any rate, the idea that engineers can't design for the unknown is wrong. Graceful failure is something engineers can sure shoot for; graceful failure is not the same thing as "meeting spec", but it does mean that "flat tire" doesn't imply "car bursts into flames." Along those lines, modern automobiles are remarkably tough --- they are designed to accomodate people who can barely drive at all, who will not service them properly... There's a lot of pretty amazing engineering in everyday life which tolerates the unknown routinely and gracefully.
Reed's article is based on the observation that Maxwell's Equations are linear (for most materials) and that, therefore the waves pass through each other without modification (again, unless you're in pretty exotic environments --- early universe, etc.) The problem with interference arises because of imperfect spectral content and non ideal antenna response for both transmitters and receivers. Interference is like being at a party: There are a lot of people talking, and your ears hear in all directions, so you have to be near the person you're trying to talk to.
... the reason that it works is that your eyes have very fine angular sensitivity (high quality antennas) compared to your ears.
For a variation on this theme, there's an interesting moment in a movie (Frankie and Johnnie?) where there's a terrific racket in a diner, impossible to understand anything, but a cook and a clerk are communicating easily --- by sign language. Consider also those occasional TV images of the Wall Street pit traders flinging gang signs at each other
Spectral purity and antenna quality limitations can be overcome --- by money. You can build higher quality receivers and transmitters, bigger antenna installations but it costs money and space in fairly unavoidable ways.
Reed is also wrong from a regulatory level. It's not just the FCC that you'd have to work with, but the ITU. Those pesky radio waves have this interesting habit of leaking over borders on the ground, and pretty much everywhere down here from satellites.
There are pretty good reasons to pick on modern broadcasting: crappy content, media concentration --- but "broadcasting" is not one of them. Those great big transmitters permit the use of very dumb receivers with poor sensitivy. The very simplicity and asymmetry of broadcast provides tremendous economic and technical appeal, and I'd be amazed if it ever went away.
Far more interesting is the glacial progress of DTV in broadcast.
Actually, internships are a *cheap* recruiting tool. The company gets to test drive a bunch of greenhorns for approximately 1/2 to 1/3 of what their "real" engineers cost, without any long term commitment to them. If they find a few keepers, then when they finally do convert them to real employees, they're already trained!
Eliminating internship programs is merely a means to save some money right now. It is *not* something that a company with an eye on the long view would do if it was serious about being in the game in the long term.
The really important ways to save money are also really difficult --- they involve layoffs, shedding of unprofitable units (which are supposed to be profitable; internship programs are not supposed to make money, of course).
-----
on a different topic: I've never heard of an unpaid internship for electrical engineers; I had (paid) interships 20 years ago, and my students certainly have been getting paid internships for the past 15 years.
When I started grad school I insisted on programming in C in a FORTRAN shop. This turned out to cause tremendous problems, and I've realized that it was a tremendous error to permit me to use a buggy C compiler to write code that none of my colleagues could understand.
Among other things, the FORTRAN compiler was reentrant, while the C compiler was not. You could have 5 people writing FORTRAN compiling merrily away, and the system would work great. But when I needed to compile, I actually had to tell everyone to take a break --- because it was going to be 5 minutes before they could do anything.
It made me a rather good C programmer, because I had to write code that would actually work without compiling blindly. But I should have used FORTRAN.
Now I'm an Extremely Powerful Professor running my own lab. To avoid being dazzled by heterogeneity, we use
linux
C
python
xml
latex
and that's just about it. My students do *not* have permission to use other tools, and they *do* have permission to abandon the legacy perl and C++ code so that we can simplify our overall code base --- at least in terms of minimizing the tools that the next students must be familiar with.
Already been done. Check out the Harris H-series computers, which have a 24 bit single precision int and a 48 bit long int.
It's a fine, fine system, in which single and double precision floats take up 48 bits, but have different precision.
You'll be especially happy to learn that the OS was called VOS for "Vulcan Operating System." The peripheral model was that "everything is a tape drive" which means that you can rewind your files.