What you've written is absolutely true -- deserts are the ideal terrain for a first deployment of autonomous vehicles. But I suspect these may fall prey to the same fate as other autonomous technology we've seen deployed in a hostile environment, namely unattended web servers on the internet. How long will it take the bad-guys (hackers) to figure out that a row of inflated garbage bags might appear as impassable boulders to these things? Or that a couple of mannequins may act like scarecrows, making the machines falsely report "hostile activity"? Or that they can perhaps "paint a hole" (a la Wile E. Coyote) that the robot will think it has to dodge? (Not to mention a little pile of "FREE BIRD SEED" on a cardboard train track!)
Yes, I know that various radars and sensors can detect those specific ruses, but I'm suggesting that a determined opponent would have lots more incentive to try lots of tricks to defeat these things. If a $10 trick can keep the U.S. Army from sending well-equipped forces down your valley, that's a helluva good investment.
Autonomy is a nice idea, but it works best in a nice environment. These things might be best suited as "slave units", loosely tethered to a lead vehicle operated (locally or remotely) by an actual human.
Well, he can hope that an appellate court will strike down the verdict since it certainly doesn't pass the "smell test" amongst the experts here on Slashdot. However, I'm not sure how British courts handle appeals. And I'm not sure he can subpoena Slashdot to testify on his behalf.
Except the nature of the charge against him was such that "intent" was a major component of it. The judge was using the evidence available to determine intent.
Look at the points the judge had to go on:
Cuthbert is a "security professional," who presumably knows both how to "hack" and how to "test".
Both hackers and testers access sites in non-standard ways.
Cuthbert accessed the site in a non-standard way. (I've read he tried a directory search.)
The charge requires "intent to defraud".
Cuthbert lied to police, then changed his story.
To be a crime, the statute says 'unauthorized access with intent to defraud.' It's obvious to me that the judge is not a technologist, and so he probably relied on expert testimony to determine if the behavior was "hacking" or merely "investigating". He probably got conflicting testimony from both sides, and so internally labeled the behavior as "ambiguous". And that's not wrong, both hackers and testers would start with the same techniques. So from there, he went to look at the rest of the charge, which says "intent".
At this point, the judge actually has to make a "judgement" -- what was Cuthbert's intent? The judge chose to use the "lying to police" when evaluating his intentions.
So, he's got ambiguous evidence of a crime, but a proven ethically-challenged defendant. He made a "judgement". That's his job.
Now we on Slashdot are all so smart that we know a directory search is a harmless way to view the contents of a site. Or is it? If you go sniffing around some site that takes credit cards, why would you do it? Could it possibly be that he WAS trying some initial penetration sniffing prior to committing an attack? We don't know, we're not Cuthbert. He claimed it was an innocent browse, but he also claimed something else to the cops. I don't know if the judge made a mistake in this case, but I do know we've only looked at the pro-Cuthbert side of things. The judge was there for the whole proceeding.
TFA quite clearly states that he was convicted because he lied to the police about his activities. Here's the quote:
"Instead, Judge Purdy found Cuthbert guilty, because he had initially lied to the police about what he had done; Cuthbert originally told the police one story and later changed it.
Judge Purdy said that Cuthbert was "deliberately trying to throw the police off the trail", by saying one thing and then another.
The fact that Cuthbert had changed his story on how and why he had originally accessed the site was the crucial factor in reaching a conviction, the judge said. "
The article above also says "The defence also pointed out that Cuthbert had not attempted to defraud the site." What it should have said is that Cuthbert DID attempt to defraud the police. Very unprofessional behavior from a supposed "security professional."
Moral of the story: don't lie to the cops about security testing. Take them seriously. Had he been honest, this wouldn't even have been prosecuted.
The Vulture Capitalists aren't looking for "The Next Big Thing", which was the failing of the 1990s bubble. It says they're looking for "The Current Big Thing" and investing in that.
You made your situation sound like "I have a Next Big Thing" idea, but don't have the finished product or the rabid fan-boi base like MySQL. That's what failed to make them money in the past.
But you're still spending money on maintenance on things with a finite lifetime. For example, buying $360 worth of tires once every four years is an ongoing expense. Just because you might make it 5 years due to lower mileage doesn't change the equation, only the numbers.
You're mistakenly stating that just because you don't put $7.50 a month in the bank to cover future tires means you aren't incurring an expense by driving your vehicle. Using the vehicle incurs SOME expense; you're simply deferring the payment. And that's my point regarding the "subscription" for these PCs -- these systems WILL require maintenance upgrades. They'll require OS patches, application patches, failed hard drive replacements, chipsets that blow out because a kid poured milk in the vents, all that stuff.
When you run a business, you either plan for these things or you go out of business quickly. That means you estimate usage, you estimate ongoing costs, and you then charge a fee to cover your expected costs and turn a profit. If it's a car business, you charge per month (with penalties for excessive mileage.) If it's a computer business, you charge per month and hope like hell that your initial estimates will cover your costs and turn a profit.
Do you pay a monthly fee for your car to keep working? Or do you consider oil changes, brakes, shocks not to be part of the cost of owning a car? Just because you don't spend the money in a "periodic" fashion on your car doesn't make it any less of an ongoing expense. Complex systems, be they automobiles or computers, will require ongoing maintenance costs. (And yes, I pay a monthly fee for my ReplayTV to get data; you may pay a similar fee for your DVR cable box.)
So now they'll offer these PICs, and they've got them locked down so Joe Sixpack theoretically won't be able to install a trojan or get a virus. But any operating system is going to require constant security patches. They don't "just keep working", because the attackers don't stop trying to hack them. And to justify the ongoing fee to the users, the ISPs will likely continue offering "new content" so Joe thinks he's getting something for his money. That will incur ongoing expenses.
Personally, I think the ISPs should offer these with a Linux based distro. They'd have far more control over content, far better remote diagnostic capabilities and much lower per-seat prices; not to mention the risk-avoidance gained by not relying on IE for security. The first time someone comes out with a trojan, worm or virus that attacks these things, the ISPs are going to have to spend a fortune reimaging their customer base.
It's a good approach: make a bone simple box that's ISP- (not user-) managed. Want new software? Go to the ISPs web site and request they install version 2.0 (or whatever). At least Joe Sixpack won't be able to install his own trojans. If the ISP is smart, they won't charge for a per-install basis, but include it as a part of the monthly rental cost.
Anyone else think it's an ugly lump-shape? It's pretty obvious to me that they didn't hire any Apple beauticians to work on this one.
We have an original Reserator, and the instructions only say "add water." We added filtered water, and then an amount of bacteriacide/algaecide designed for humidifiers as protection against stuff growing in the hoses and blocks. We did this mostly to keep the internals from being coated by slime that could reduce the cooling capacity of the system.
However, the currently-slimy interior of the tank leads me to believe the water may still harbor some unhealthy life forms. I'm thinking an actual ethylene glycol anti-freze would prove more toxic to the bad things. The hose, gasket and block materials are all ordinary compounds that are unaffected by ethylene glycol in a typical car engine, so it should be safe to try without buying the Zalman bottle.
Leaks haven't been a big problem. There is a small pressure relief hole in the top center of the cap plug, which does leak if you tip the device on its side. (This fix is to not tip it on its side.) And we did have to send the flow indicator in for replacement -- while the tube ends are solid metal threaded hose connetors, the flow indicator body is just an acrylic pipe that cracked invisibly under the label when tightened, and had an incredibly slow leak that only manifested itself when the hoses were moved.
May I ask why you would ever shut off your reserator? My son shuts off his PC every night (the hard drives are noisy) but never the Reserator. At night he can't even hear the pump that's just 12 feet from his bed. The only time he ever shuts it down is to carry the thing to a LAN party (which is still a two-person job until he buys some of Zalman's new quick-disconnects.)
The pump draws about 5W. At my local electric rate that's under $4.00 per year. Cheap insurance when compared to what might happen if he forgot to turn it on. Thermal overload shutdowns are still too hot for the chips, and are no guarantee of damage prevention. Plus, you may (or may not) have thermal monitoring of your northbridge or your graphics chipsets -- they may get hotter than the CPU's threshhold.
If there's some other reason where you absolutely need to power down the Reserator when the PC is off, just buy a 120VAC relay to switch the pump on and off, and control it with 12VDC from the PC's power supply -- it then turns the pump on and off with your system's power. They're available on all the do-it-yourself watercooling sites.
Actually, the manufacture of CFCs and HCFCs was only restricted by the Montreal Protocol, and not banned outright. There will always remain a few legal and legitimate uses of them (asthma inhalers, for example,) although they are few and far between. There are also exemptions that were put in place to allow for servicing some really big (and expensive) existing air conditioning units (the U.S. insisted.) These uses include large office tower style air conditioners that have an expected 40 year service life, and the treaty has rules phasing these compounds out for a long time yet. The manufacture of CFC based refrigerants won't be completely banned until sometime in the 2030s!
However, because manufacture was so severely curtailed and new consumption was prohibited, most manufacturers simply stopped producing it completely - there wasn't enough economic incentive to continue production at the restricted levels.
There are still existing stockpiles of CFC 12. It's funny, but the first effect the Montreal Protocol had was to cause HVAC companies across the nation to stockpile thousands of pounds of the stuff, and the net result of that was an order of magnitude price increase in the early 90s, from about $4/pound to about $60/pound. If you need Freon(TM) brand refrigerant, you can certainly still get it. Go Froogle for R-12, and I bet you'll find some.
Do you remember the guy who tried immersion cooling his PC in fluorinert a few years back? His plan was to filled a cooler with the stuff, submerge his motherboard, and then add dry ice to keep it all about -60. After adding about $800 worth of the stuff he discovered that it gelled somewhere above the freezing point of carbon dioxide, making it useless for the extreme cooling he had been hoping for.
Definitely the profession was watered down by all the moneyseekers in the 90s. And those people are already moving on to other fields. Everyone's discovered that a rote-learned task is the easiest to outsource.
And yes, abstractions will hide knowledge, keeping people from needing to learn what's in the box. But a lot of boxes in computer science are still troublesome enough that there is still be a need for the deeper understanding.
The key to survival is to know when a particular box can be ignored. For example, I've stopped worrying about drive controllers, and now trust someone else to determine that SATA is "faster" than IDE. Since I've recognized that I've become a generic consumer of drives, I no longer have to keep current on the technology, how commands are transmitted, cylinders, heads, encodings, etc. And it's a good thing: the field is now so broad and so deep that keeping up across the board is too tough. I still retain my old working knowledge of drives: magnetic coated platters are spun fast, heads go back and forth, etc., but unless I have a reason to dive into a particular technology, it's adequate for me just to know the basics. I can mostly ignore drives and focus instead on my specialty.
I killed two birds with one stone, and bought an APC battery backup UPS. The software that came with the UPS (which mostly sucks, by the way) does allow it to show battery status. At any time it can report how much power is being drawn, in addition to battery voltage, temperature, etc.
I bought a fairly big UPS for the capacity to hold up not only the PC and its thirsty innards, but to power the other semi-essential crap I have in the computer room. I've got the router and cable modem (almost required if you have VOIP), the LCD monitor, and a couple of USB hubs all plugged into the UPS. The USB hubs are required for not only the USB keyboard and mouse, but the battery uses USB to send the "power fail" warnings to the PC. All in all it draws 213 watts at pretty slack usage of the CPU.
Certainly I could unplug a lot of the other crap and just have the PC on the UPS and measure it, but I wasn't looking for that number.
With no fan, it should be absolutely inaudible from a centimeter away. I would expect an expensive power supply to not suffer from a cheap whiny transformer.
Except it has a fan. According to TFA, it doesn't kick on until the PSU's temp reaches 40, 47.5 or 55 degrees (switch selectable.) Theoretically, the temperature of the PSU will rise as the power load increases, so unless you're slinging lots of polygons through your video card and lots of equations through your CPU, it shouldn't come on very often. But it will -- if you need 500W of power, you're going to be generating lots of heat.
The bigger "silent" issue I have is that it's only an 80mm fan. Small diameter fans are by necessity louder than large diameter fans due to simple physics: a smaller fan must turn faster to move the same volume of air as a larger fan. The faster they turn, the more noise they produce. And as you point out, 90W is a lot of heat to dissipate.
If you reread it, I originally wrote "witnessed by man so far." That pretty much means (to me, anyway) it had to have occurred in written or oral history, which restricts it to about the last 8,000 years (in China anyway, or 6,000 years if you're a fundamentalist christian and believe that the earth was created in 4004 BC.)
If you, or any of your forebears were alive in human form to witness the Chicxulub impact 65 million years ago, then I apologize. Otherwise, I believe the ribbon probably would have more mass than the meteor believed responsible for the Tunguska explosion, which I believe to be the largest in recent history.
But whatever, you goaded me into it, time to do some simple math.
According to wikipedia, the Tunguska meteor was among the larger human-timeframe impact events. At least we have some estimates to work with. It destroyed virtually everything in a 50km radius. According to wiki, they estimated it was a stoney meteor about 60 meters in diameter; using your 1500kg/m3 formula, that's about 170,000,000kg. That's still a lot of weight, but closer to reality. I'll try to compare that to the ribbon.
The LiftPort site says "the ribbon will be about the size of a sheet of paper". (A4 has an area of 625 cm2.) And they say the ribbon needs to be 36,000 km long (I thought geosynch orbits were 40,077km away, but I must have been mistaken.) According to CalTech, carbon nanotubes in an armchair configuration have a density of 1.33 g/cm3. Simple math says that's 2,992,500,000 kg (but please recheck my math, I may have missed a g to kg or cm to km somewhere.)
The Tunguska explosion destroyed everything within a 50km radius. The elevator ribbon appears to be roughly an order of magnitude more mass, but also won't be falling at the same rate. It will probably strike the earth with a much smaller effect radius, but over an incredible distance.
I think we're both right. It's nowhere near the size of the meteor that caused the extinction level event that felled the dinosaurs. But I can pretty safely state that you'd want to be a long way from any tsunami-prone Pacific coast lines if this puppy decided to detach from its skyhook.
(And no, I would run screaming from any association with Day After Tomorrow.)
Well, it's not like you can "patch" pieces of ribbon together to make one big long ribbon. The strength comes from the uniformity of the fibers, and patches would quickly disintegrate under the strain. It needs to be one continuous finished piece, which means it would almost certainly need to be fully manufactured from raw materials brought into space. It would be virtually impossible to produce a single spool with 40000km of rope like that here on earth, and then lift it up all at once into space. The factory would be placed at the farthest point from earth. As it produced the fiber, it would drop the ribbon out the "bottom" of the factory, toward the earth.
The factory itself might have to start out located much further out than its final geosynchronous orbit. As the ribbon lengthens, its CG will shift downward. It would be advantageous to not have to move the ribbon much once it's assembled, so that means you'd want the last bit of ribbon done just as the CG approached its final geosynchronous orbit.
I thought about that, but carbon fiber is already used commercially in many heat resistant applications. While I doubt that carbon fiber's characteristics confer identical properties to carbon nanotubes, I imagine they share some similarities. But this ribbon is going to require some remarkable toughness properties to survive. It needs to withstand the Van Allen radiation. It needs to hang the last few hundred miles in the atmosphere, where it will be subject to rain, storms, ice, lightning, etc. It will accumulate lots of dust, and quite possibly debris clouds near some of the messier LEOs. It needs to withstand the thermal shock of the daily transition from sunlight to shadow and vice versa.
But what might really do it in, though, is this ribbon is going to be a giant electrical generator. As it orbits through the earth's magnetic field, it's going to develop a tremendous charge along its length. What will dissipate that charge? Is the fiber going to be adequate to carry that charge? Do we ground the terminus or use the current? And what happens to the fibers as the ribbon expands, contracts and flexes throughout the day? Will microscopic voids appear in the fiber, which may in turn cause tiny arcs? And might these arcs eventually burn through the fibers, causing catastrophic failure?
Incorporating 22,000 miles of 0000 gauge copper welding cable is not likely to be a good answer, at least not from a weight / strength / cost perspective. But something will need to deal with the charge, and I've seen nothing so far that does.
So, throw in several dozen masters' theses worth of materials science, and you want it to magically incinerate as it falls, too? Well, why not?
For the cable to work (if the tip is to hover about a "terminal station" here on earth), its center of gravity (CG) needs to be in a geosynchronous orbit, which is about 22,000 miles up. Let's also assume that the cable is tapered. The strongest (thickest) part needs to be at the CG, where the load is heaviest (it's holding up 22,000 miles of cable after all.) Because the distant ends don't have to hold up as much cable weight, they can be much thinner. Let's also assume that the United States built the cable, and is hosting the terminal station somewhere in Central America.
If the cable were to break, the part "below" the break would fall to earth, simply because it would have to, and the part "above" the break would be flung out into space. For fun, lets assume the cable breaks near the CG, leaving 22,000 miles of cable to fall to earth, which has an equatorial circumference of 24,000 miles.
As the 22,000 mile-long-strand of cable begins to fall, it picks up speed at 9.8 meters per second per second. And as it falls, the earth continues to spin beneath it, causing the falling cable to head west. As the speed of descent picks up, it will pull the heavy broken end down at a supersonic rate. The world's longest, loudest sonic boom will wrap around the earth's equator as the cable winds around it as a string winds around a yo-yo. At first, the sound will be merely deafeningly loud to the equatorial residents, but at some point that shock wave may prove fatally loud for any victims within close range. The far end will be pulled down faster and faster, like a bullwhip. Without doing the math, after several hours of falling the massive broken tip will finally be pulled to earth with energy likely exceeding that of any meteoric collision witnessed by man so far. It will probably trigger a host of phenomena around the equator to the final impact site, including tornados, hurricanes, earthquakes, tsunamis and possibly even volcanos.
Other than that, it's perfectly safe. Nothing to worry about.
Rather than USENET, I believe the O.P. was referring to gmail. Web based or not, people (perhaps wrongly) have an expectation of privacy in their personal communications. For Google to be snarfing this out of gmailed data would be extremely evil (violating one of their core principles.)
While the suggestion that Google is monitoring gmail for stock activity is really funny, in real life it wouldn't happen for several reasons. First, they'd be violating their user's privacy. Next, they might be seen as taking advantage of insider information. While companies may have a policy in place about not discussing stock information over email, and most people won't violate their employers rules, human behavior does suggest they'll still talk ABOUT their company.
Finally, if it was revealed that Google ever tried this, you can bet the scammers would immediately start exchanging gmails about some company as part of a pump'n'dump scheme!
BTW, traffic analysis is amazing stuff -- it's been suggested that the pizza delivery businesses in Washington D.C. know when something is up before the rest of us because they deliver more midnight pizzas than usual to the Pentagon.
Slashdot Mirror
Yes, I know that various radars and sensors can detect those specific ruses, but I'm suggesting that a determined opponent would have lots more incentive to try lots of tricks to defeat these things. If a $10 trick can keep the U.S. Army from sending well-equipped forces down your valley, that's a helluva good investment.
Autonomy is a nice idea, but it works best in a nice environment. These things might be best suited as "slave units", loosely tethered to a lead vehicle operated (locally or remotely) by an actual human.
Well, he can hope that an appellate court will strike down the verdict since it certainly doesn't pass the "smell test" amongst the experts here on Slashdot. However, I'm not sure how British courts handle appeals. And I'm not sure he can subpoena Slashdot to testify on his behalf.
Look at the points the judge had to go on:
- Cuthbert is a "security professional," who presumably knows both how to "hack" and how to "test".
- Both hackers and testers access sites in non-standard ways.
- Cuthbert accessed the site in a non-standard way. (I've read he tried a directory search.)
- The charge requires "intent to defraud".
- Cuthbert lied to police, then changed his story.
To be a crime, the statute says 'unauthorized access with intent to defraud.' It's obvious to me that the judge is not a technologist, and so he probably relied on expert testimony to determine if the behavior was "hacking" or merely "investigating". He probably got conflicting testimony from both sides, and so internally labeled the behavior as "ambiguous". And that's not wrong, both hackers and testers would start with the same techniques. So from there, he went to look at the rest of the charge, which says "intent".At this point, the judge actually has to make a "judgement" -- what was Cuthbert's intent? The judge chose to use the "lying to police" when evaluating his intentions.
So, he's got ambiguous evidence of a crime, but a proven ethically-challenged defendant. He made a "judgement". That's his job.
Now we on Slashdot are all so smart that we know a directory search is a harmless way to view the contents of a site. Or is it? If you go sniffing around some site that takes credit cards, why would you do it? Could it possibly be that he WAS trying some initial penetration sniffing prior to committing an attack? We don't know, we're not Cuthbert. He claimed it was an innocent browse, but he also claimed something else to the cops. I don't know if the judge made a mistake in this case, but I do know we've only looked at the pro-Cuthbert side of things. The judge was there for the whole proceeding.
"Instead, Judge Purdy found Cuthbert guilty, because he had initially lied to the police about what he had done; Cuthbert originally told the police one story and later changed it.
Judge Purdy said that Cuthbert was "deliberately trying to throw the police off the trail", by saying one thing and then another.
The fact that Cuthbert had changed his story on how and why he had originally accessed the site was the crucial factor in reaching a conviction, the judge said. "
The article above also says "The defence also pointed out that Cuthbert had not attempted to defraud the site." What it should have said is that Cuthbert DID attempt to defraud the police. Very unprofessional behavior from a supposed "security professional."
Moral of the story: don't lie to the cops about security testing. Take them seriously. Had he been honest, this wouldn't even have been prosecuted.
You made your situation sound like "I have a Next Big Thing" idea, but don't have the finished product or the rabid fan-boi base like MySQL. That's what failed to make them money in the past.
I like the way you put that. I read it as: "They're not nobodies, they're the 19,000th most popular rock band in the world!"
You're mistakenly stating that just because you don't put $7.50 a month in the bank to cover future tires means you aren't incurring an expense by driving your vehicle. Using the vehicle incurs SOME expense; you're simply deferring the payment. And that's my point regarding the "subscription" for these PCs -- these systems WILL require maintenance upgrades. They'll require OS patches, application patches, failed hard drive replacements, chipsets that blow out because a kid poured milk in the vents, all that stuff.
When you run a business, you either plan for these things or you go out of business quickly. That means you estimate usage, you estimate ongoing costs, and you then charge a fee to cover your expected costs and turn a profit. If it's a car business, you charge per month (with penalties for excessive mileage.) If it's a computer business, you charge per month and hope like hell that your initial estimates will cover your costs and turn a profit.
So now they'll offer these PICs, and they've got them locked down so Joe Sixpack theoretically won't be able to install a trojan or get a virus. But any operating system is going to require constant security patches. They don't "just keep working", because the attackers don't stop trying to hack them. And to justify the ongoing fee to the users, the ISPs will likely continue offering "new content" so Joe thinks he's getting something for his money. That will incur ongoing expenses.
Personally, I think the ISPs should offer these with a Linux based distro. They'd have far more control over content, far better remote diagnostic capabilities and much lower per-seat prices; not to mention the risk-avoidance gained by not relying on IE for security. The first time someone comes out with a trojan, worm or virus that attacks these things, the ISPs are going to have to spend a fortune reimaging their customer base.
Anyone else think it's an ugly lump-shape? It's pretty obvious to me that they didn't hire any Apple beauticians to work on this one.
However, the currently-slimy interior of the tank leads me to believe the water may still harbor some unhealthy life forms. I'm thinking an actual ethylene glycol anti-freze would prove more toxic to the bad things. The hose, gasket and block materials are all ordinary compounds that are unaffected by ethylene glycol in a typical car engine, so it should be safe to try without buying the Zalman bottle.
Leaks haven't been a big problem. There is a small pressure relief hole in the top center of the cap plug, which does leak if you tip the device on its side. (This fix is to not tip it on its side.) And we did have to send the flow indicator in for replacement -- while the tube ends are solid metal threaded hose connetors, the flow indicator body is just an acrylic pipe that cracked invisibly under the label when tightened, and had an incredibly slow leak that only manifested itself when the hoses were moved.
The pump draws about 5W. At my local electric rate that's under $4.00 per year. Cheap insurance when compared to what might happen if he forgot to turn it on. Thermal overload shutdowns are still too hot for the chips, and are no guarantee of damage prevention. Plus, you may (or may not) have thermal monitoring of your northbridge or your graphics chipsets -- they may get hotter than the CPU's threshhold.
If there's some other reason where you absolutely need to power down the Reserator when the PC is off, just buy a 120VAC relay to switch the pump on and off, and control it with 12VDC from the PC's power supply -- it then turns the pump on and off with your system's power. They're available on all the do-it-yourself watercooling sites.
However, because manufacture was so severely curtailed and new consumption was prohibited, most manufacturers simply stopped producing it completely - there wasn't enough economic incentive to continue production at the restricted levels.
There are still existing stockpiles of CFC 12. It's funny, but the first effect the Montreal Protocol had was to cause HVAC companies across the nation to stockpile thousands of pounds of the stuff, and the net result of that was an order of magnitude price increase in the early 90s, from about $4/pound to about $60/pound. If you need Freon(TM) brand refrigerant, you can certainly still get it. Go Froogle for R-12, and I bet you'll find some.
Do you remember the guy who tried immersion cooling his PC in fluorinert a few years back? His plan was to filled a cooler with the stuff, submerge his motherboard, and then add dry ice to keep it all about -60. After adding about $800 worth of the stuff he discovered that it gelled somewhere above the freezing point of carbon dioxide, making it useless for the extreme cooling he had been hoping for.
And yes, abstractions will hide knowledge, keeping people from needing to learn what's in the box. But a lot of boxes in computer science are still troublesome enough that there is still be a need for the deeper understanding.
The key to survival is to know when a particular box can be ignored. For example, I've stopped worrying about drive controllers, and now trust someone else to determine that SATA is "faster" than IDE. Since I've recognized that I've become a generic consumer of drives, I no longer have to keep current on the technology, how commands are transmitted, cylinders, heads, encodings, etc. And it's a good thing: the field is now so broad and so deep that keeping up across the board is too tough. I still retain my old working knowledge of drives: magnetic coated platters are spun fast, heads go back and forth, etc., but unless I have a reason to dive into a particular technology, it's adequate for me just to know the basics. I can mostly ignore drives and focus instead on my specialty.
I bought a fairly big UPS for the capacity to hold up not only the PC and its thirsty innards, but to power the other semi-essential crap I have in the computer room. I've got the router and cable modem (almost required if you have VOIP), the LCD monitor, and a couple of USB hubs all plugged into the UPS. The USB hubs are required for not only the USB keyboard and mouse, but the battery uses USB to send the "power fail" warnings to the PC. All in all it draws 213 watts at pretty slack usage of the CPU.
Certainly I could unplug a lot of the other crap and just have the PC on the UPS and measure it, but I wasn't looking for that number.
Except it has a fan. According to TFA, it doesn't kick on until the PSU's temp reaches 40, 47.5 or 55 degrees (switch selectable.) Theoretically, the temperature of the PSU will rise as the power load increases, so unless you're slinging lots of polygons through your video card and lots of equations through your CPU, it shouldn't come on very often. But it will -- if you need 500W of power, you're going to be generating lots of heat.
The bigger "silent" issue I have is that it's only an 80mm fan. Small diameter fans are by necessity louder than large diameter fans due to simple physics: a smaller fan must turn faster to move the same volume of air as a larger fan. The faster they turn, the more noise they produce. And as you point out, 90W is a lot of heat to dissipate.
If you, or any of your forebears were alive in human form to witness the Chicxulub impact 65 million years ago, then I apologize. Otherwise, I believe the ribbon probably would have more mass than the meteor believed responsible for the Tunguska explosion, which I believe to be the largest in recent history.
But whatever, you goaded me into it, time to do some simple math.
According to wikipedia, the Tunguska meteor was among the larger human-timeframe impact events. At least we have some estimates to work with. It destroyed virtually everything in a 50km radius. According to wiki, they estimated it was a stoney meteor about 60 meters in diameter; using your 1500kg/m3 formula, that's about 170,000,000kg. That's still a lot of weight, but closer to reality. I'll try to compare that to the ribbon.
The LiftPort site says "the ribbon will be about the size of a sheet of paper". (A4 has an area of 625 cm2.) And they say the ribbon needs to be 36,000 km long (I thought geosynch orbits were 40,077km away, but I must have been mistaken.) According to CalTech, carbon nanotubes in an armchair configuration have a density of 1.33 g/cm3. Simple math says that's 2,992,500,000 kg (but please recheck my math, I may have missed a g to kg or cm to km somewhere.)
The Tunguska explosion destroyed everything within a 50km radius. The elevator ribbon appears to be roughly an order of magnitude more mass, but also won't be falling at the same rate. It will probably strike the earth with a much smaller effect radius, but over an incredible distance.
I think we're both right. It's nowhere near the size of the meteor that caused the extinction level event that felled the dinosaurs. But I can pretty safely state that you'd want to be a long way from any tsunami-prone Pacific coast lines if this puppy decided to detach from its skyhook.
(And no, I would run screaming from any association with Day After Tomorrow.)
The factory itself might have to start out located much further out than its final geosynchronous orbit. As the ribbon lengthens, its CG will shift downward. It would be advantageous to not have to move the ribbon much once it's assembled, so that means you'd want the last bit of ribbon done just as the CG approached its final geosynchronous orbit.
But what might really do it in, though, is this ribbon is going to be a giant electrical generator. As it orbits through the earth's magnetic field, it's going to develop a tremendous charge along its length. What will dissipate that charge? Is the fiber going to be adequate to carry that charge? Do we ground the terminus or use the current? And what happens to the fibers as the ribbon expands, contracts and flexes throughout the day? Will microscopic voids appear in the fiber, which may in turn cause tiny arcs? And might these arcs eventually burn through the fibers, causing catastrophic failure?
Incorporating 22,000 miles of 0000 gauge copper welding cable is not likely to be a good answer, at least not from a weight / strength / cost perspective. But something will need to deal with the charge, and I've seen nothing so far that does.
So, throw in several dozen masters' theses worth of materials science, and you want it to magically incinerate as it falls, too? Well, why not?
If the cable were to break, the part "below" the break would fall to earth, simply because it would have to, and the part "above" the break would be flung out into space. For fun, lets assume the cable breaks near the CG, leaving 22,000 miles of cable to fall to earth, which has an equatorial circumference of 24,000 miles.
As the 22,000 mile-long-strand of cable begins to fall, it picks up speed at 9.8 meters per second per second. And as it falls, the earth continues to spin beneath it, causing the falling cable to head west. As the speed of descent picks up, it will pull the heavy broken end down at a supersonic rate. The world's longest, loudest sonic boom will wrap around the earth's equator as the cable winds around it as a string winds around a yo-yo. At first, the sound will be merely deafeningly loud to the equatorial residents, but at some point that shock wave may prove fatally loud for any victims within close range. The far end will be pulled down faster and faster, like a bullwhip. Without doing the math, after several hours of falling the massive broken tip will finally be pulled to earth with energy likely exceeding that of any meteoric collision witnessed by man so far. It will probably trigger a host of phenomena around the equator to the final impact site, including tornados, hurricanes, earthquakes, tsunamis and possibly even volcanos.
Other than that, it's perfectly safe. Nothing to worry about.
Right! Uhh... so can I have your PC then?
No, wait. I'm really sure I don't want to read Friday's entry.
Whew. I breathed half a sigh of relief when I read that.
Ooo, ooh, my turn to hurl the trite cliche!
[ check, check, one, two. OK, got it. ]
*ahem*
"You must be new here."
Thank you, thank you, I'll be here all the week! Tip your moderators, they work hard for you! Thanks, everyone. Good night!
Rather than USENET, I believe the O.P. was referring to gmail. Web based or not, people (perhaps wrongly) have an expectation of privacy in their personal communications. For Google to be snarfing this out of gmailed data would be extremely evil (violating one of their core principles.)
While the suggestion that Google is monitoring gmail for stock activity is really funny, in real life it wouldn't happen for several reasons. First, they'd be violating their user's privacy. Next, they might be seen as taking advantage of insider information. While companies may have a policy in place about not discussing stock information over email, and most people won't violate their employers rules, human behavior does suggest they'll still talk ABOUT their company.
Finally, if it was revealed that Google ever tried this, you can bet the scammers would immediately start exchanging gmails about some company as part of a pump'n'dump scheme!
BTW, traffic analysis is amazing stuff -- it's been suggested that the pizza delivery businesses in Washington D.C. know when something is up before the rest of us because they deliver more midnight pizzas than usual to the Pentagon.