Is this a new record for number of misstatements? Let's count:
"a chemical plant run by Union Carbide". Whoa there. India is a semi-socialist country; it has some very strict laws limiting foreign ownership and control. It would be fairer to say "a company majority owned by the Indian Government, with Union Carbide as a minor partner". I realize this doesnt fit in with the "bad multinational american companies" bogey-man, but tough.
"released about 40 tons of a toxic gas". Maybe it would be fair to state that this was an UNINTENTONAL release, probably caused by various factors including tired workers and labor unrest.
"...plant was in the business of creating chemicals deadly to life." Well, if you call locusts, nematodes, chiggers, termites, ants, earwigs, wasps, et al, "life", I guess thats right. But a few people think it's a good idea in a country with a marginal food supply to save a few million human lives from death by starvation by producting chemicals to kill the above-listed lifeforms, so that humans might have a chance of eating some food.
"Safety at the plant had not been a concern of management". Hmmm, let me think, let's assume the management has no concern for human life, including their own lives or those of their workers-- they just care about making money. In order to make money in that business, you have to be able to control dozens of chemical reactions, temperatures, pressures, flow rates, valves, pumps, manifolds, etc.. If you don't have the right control systems in place everywhere, things can quickly go out of control, and your many billion dollar plant is not making pesticides, it's making a mixture of 31% water, 38% salt, and 21% brown sludge. There HAVE to be extensive mechanisms in place to prevent chemical releases, at the very least those chemicals are expensive!
"Today, the site remains a contaminated wasteland, unusable and never cleaned up." It comes down to economics: is the cost of cleanup less than the value of the land? India is not that small that it will miss a few acres.
"The survivors have been minimally compensated". There's a fund of $328 million dollars. If that were spread evenly over the 3,000-some families of survivors, each family would get about $100,000 each. That's about 35 years of average earnings. I wouldnt call that "minimal".
Yes, it was a bad thing that happened, but it doesnt help to misdirect blame and energy.
CS is one of the few professional fields left where a college degree is not all that necessary. Sure, it's "nice" to know the intricacies of B-trees, finite-state automata and the like, but these things rarely show up in their pure theoretical form in most everyday work.
What IS important is to be able to absorb info, make design tradeoffs, and apply your posterior to the chair for extended periods of time until you get the dang program working!
I've seen CS PHd's that can't figure out the * operator in C. I've also seen high-school dropouts that can do most anything with a computer. You decide which is more marketable.
I just looked it up, it's the subjunctive tense. It's used in indirect discourse, or when there's doubt, or when there's no doubt what's quoted is counter-factual: (Example of all three: Newscaster reporting "Jacko claims he's normal and innocent"),
I think German has the same thing. A whole bunch of word suffixes get added if the subject is somewhat in doubt. It makes a newscaster's job kinda hard if they're trying to track a developing story on the air.
Hmmm, this sounds like one of those ideas that's bound to lose:
When you shrink a generator and turbine to one tenth its former size:
The power output goes down by a factor of 1,000. (power out is proportional to L x W x H )
The friction in the bearings goes down by a factor of 100 (proportional to surface area of bearings)
The windage losses due to air friction between the generator rotor and stator stay about the same.
The air friction losses in the turbine may go waay up (as the ratio of turbulent flow to mainline flow goes way up).
The thermal input from the burning gases goes down by a factor of 1,000.
The thermal losses only drop by a factor of 100. Eventually the losses become greater than the thermal input, making it impossible to sustain burning.
So every time you shrink these things, the power out goes way down, the efficiency goes waaay down.
Do this a few times and you'll have a turbine that can't even overcome its internal friction and a generator that, even if you could turn it, would be way down on the efficiency scale.
Shrinking these things is a very very very *losing* thing to do.
This seems like a really poor application for computers.
I can see the computer's analysis:
Brushstrokes: 97% Leonardo Da Vinci
Contrast: 98% Leonardo Da Vinci
Demarcation: 96% Leonardo Da Vinci
Fletangio: 97% Leonardo Da Vinci
Impressimi: 98% Leonardo Da Vinci
Random kid passing by: "Mom, look at that neat picture of dogs playing poker!"
" Thin CRTs offer the best of both worlds -- superior picture quality with a slim size. "
I wouldnt call a 20% reduction, from 51 to 41cm deep , a "slim" CRT, nor worthy of Slashdot coverage. And they're probably compromising on something-- I'd guess they're going to lose a bit of convergence near the edges.
IMHO there's not a chance in a million this is doable anytime in the next 50 years.
Computers and sensors and "AI" are nowhere near capable of detecting, identifying, and processing everything the human eye and mind can do.
Just think of what kind of judgements you make every second-- driver has grey hair, and is wearing a 1940's hat? Is ther FUR on the hat? Double your caution.
Obstacle in the road? Is it a wrinkled beige plastic bag, inflated by the breeze, or is it a one-ton boulder that rolled down the canyon? You can tell, I doubt if a computer can make that judgement.
Car beside you throws off a spray from going thru a mud puddle. You know it's not a hazard. Are your radar or ultrasonic sensors going to identify this as harmless, or as an solid wall you're approaching at a 45 degree angle?
A little fog this morning. You can maneuver because you can identify the cars by their tailight configurations. Wanna try this with a computer?
There's a foil candy-wrapper on the road. it happens to have landed with a 90 degree bend in the middle. Said wrapper is an excellent corner-reflector, reflecting radar signals with the apparent cross section of a large semi. You wouldnt give this a second thought. A computer would SLAM ON the brakes to avoid hitting the semi.
All these and probably hundreds more everyday situations your brain can analyze and discard the harmless ones. It seems mighty unlikey that a computer could do 10% as well, which is pretty bad as the whole point is to do better than humans do.
Hmmm, only glitches:
* Usually the task is loosely defined
* Usually the environment is also.
Example: Boss says "write a program that sends me a weekly summary of the file system backup status by email". I don't think a computer is going to be able to make much out of this. It takes a heck of a lot of human judgement to figure out (1) What level of detail the boss wants. (2) Do we format it as tables, line graphs, bar charts, ?? (3) Can his computer handle text only, or.pdf, or HTML, or FLASH, or 3D?? (4) Do I have the political "juice" to get access to all the sysadmin data? (5) Are the backup statistics reliable? (6) Is the boss's dumb son-in-law going to block my access unless I throw him a bone? (7) How am I doing to seamlessly do this for both Unix and NT? (8) Are the Windows API's really usable for this, or are they going to be hopeless as in the last 4 projects?
I don't think any computer parser is going to be able to make these judgements.
One might suspect Intel realized a while ago that Itanic was never going to make it. Now they're generating various scenarios, trying to figure out how to write off the investment at the least painful moment. One good instant might have been when Andy Grove left. Or maybe the next guy is revving up for this. Anyway, the shoe is going to drop sometime in the next 12 months. Don't expect much sudden drop in their stock as the market willprobably anticipate this.
IIRC when that Cd full of Windows source code got out, somebody did a "grep strcat" on it. Something like 13,000 hits.
Now some of those are going to be safe, but how many of those are capable of overflowing under the right conditions? 1% ? 10%?
The mind reels.
Thank you Guardian! Last time I looked, the US military space budget was bigger than NASA's. What do you think they do with all that money, play Space Invaders?
The US military has been developing anti-satellite weapons in plain sight for, oh, 30 years. IIRC we have a little missle, launched from a F-15, that can knock down satellites. We've had it for 30-some years. It doesnt violate any treaty as it's not based in space. Prolly been other more secret developments, not anything to get too worked up about.
Ok Class, I want a paragraph from each of you, starting out: The motivation to adhere to a treaty is... Single-spaced, by the end of the hour....
( 60 mins later )
Ok, I see nobody has written anything. Let me explain:
Sovreign nations sign treaties when it is in their interest to do so. The interest might be political, economic, social, historical, cultural, or otherwise. The participants may be acting in good faith, or being devious, or seeing a short-term advantage, or a strategic opportunity.
In any case, nations adhere to a treaty as long as it is in their interest, and not much longer than that. That's why most treaties have a boilerplate paragraph along the lines of:: And any participant can back out of the treaty with 30 days notice.
So it's not too helpful to yell BUT THAT WOULD VIOLATE THE TREATY! A treaty is not an iron-clad dissuader, just a way to identify common interests, if any.
See: German non-agression treaty, et. al....
I'd reserve the phrase "totally refutes" for occasions where.... this actually happens.
What I saw of the "refutation" was a few bits of unconvincing excuses and loose logic. The similarity in headers and the number and length of exact code matches is compelling and proabably irrefutable evidence.
The original WW2 VT fuses used silver conductive ink, circa 1942. This isnt new.
IIRC the old circa 1977 Vax cards were near 30 layers.
More layers isnt necessarily better-- it imples a lot of crossing wires, which is often the result of bad design.
Silver nanospheres are not going to carry much current.... You probably couldnt make a PC motherboard or high-speed bus this way.
Insulating ink between layers is going to give whoppingly large capacitance and crosstalk between layers. Not good.
Most buses have to be designed to a certain surge impedance, which is hard enough to do with current technology. Imagine trying to maintain 120 ohms across multiple sprayed-on layers... A deginer's nightmare.
Printing by injet is a SERIAL process, one stripe at a time, one layer at a time. Regular PCB processes are multi-parallel-- exposing many boards at a flash, etching many boards in one tank. Old way is intrinsically much faster.
"Ground based".... so it can only jam satellites above it's horizon. Not terribly useful.
How can this be new or newsworthy? All it takes to jam a satellite is a few watts of power. Any TV station uplink truck could do the same, given the coordinates and frequency. I suspect the military has had this capability for oh, 40 yrs or more.
"without burning it up". The range between blotting out a receiver and burning it up is probably at least a factor of 1,000,000, or much more if the satellite designers took any EMP precautions. Shame on them if they didnt.
All the coal and oil on the planet (about 3 teratons) is only about 8% of the carbon dissolved in the oceans. Which seems to imply two things:
(1) We need to stir up the oceans a bit to get some of that CO2-poor deep water to the surface.
(2) If we got desperate we could mine the waters for carbon.
Just a FEW quibbles with this nuclear gas rocket design:
* 25,000 degrees C hot uranium hexafloride is going to melt and react with the the quartz toute-suite.
* You'd need hundreds of pounds of 100% enriched UF6 to get a critical mass. Even under pressure, that's a lot of volume.
* Reactors are controllable due to the 1 to 2 percent of fissions that result in delayed neutron emission. But this gas is going to have a lot more than 1 or 2 percent variations in density. Ergo you're going to have a really hard time (~impossible) controlling the reaction.
* You're still going to need reaction mass to shove out the back. Just try to find a compound that is (1) Liquid, (2) Not too toxic (3) Doesnt disassociate at 25,000K
Otherwise OK!
Enough holes in this story to drive several Beowulf clusters thru it:
* They switch at 27Mhz do they? How many times can you flip a mechanical switch before it breaks? (About 0.1 seconds worth for most switches)
* Ok, it takes no power somehow to hold the info. But what about reading/writing it? It's going to take not only power, but several transistors per bit. Good old DRAM nowdays is down to 1 well and one MOSFET per bit. ( plus row and column drivers). Can't even approach that with anything mechanical.
* 27MHz is right in the middle of the old CB band! Watch out for truckers on your tail.
* Volume goes down as the cube, surface area as the square of the linear dimensions. So these thingies are really at the whim of surface tension and surface electrostatic effects. The good part is they should be rather insensitive to mechanical shock. The bad part: watch out for static electricity, Hall Effect, and dust!
* Sounds about as practical as bubble memory, string floppies, and 90-column oval-holed cards.
As a last resort, we could look at the science behind ELF before we worry too much about the "damage":
(1) ELF transmitters are only a megawatt or so. The ELF waves are sooo long (many thousands of miles), that a little 50 mile antenna only radiates oh, maybe 5 watts of effective radiated power. (the rest just heats up the wires). Those 5 watts get spread more-or-less evenly all around the earth.
(2) Your tpical large marine creature is maybe one billionth the size of the earth, so we're down to maybe 5-billionths of a watt hitting the beast.
(3) Your typical animal is an even smaller fraction of the thousand-mile ELF wavelength. So about 99.99999% of the energy incident on say a giant squid goes right through it. We're now down to 5 quadrillionths of a watt.
(4) A typical nerve discharge is around a THOUSAND to a MILLION times that amount of energy, so the ELF signal is that much weaker than the thousands of nerve impluses going off right inside the squid's body every second.
(5) So I would not worry too much about ELF harming anything.
(6) And, oh, as other have mentioned, the energy from power lines is many orders of magnitude stronger than ELF (and even that is hard to pick up any distance from power lines).
I don't think 332 angles gives you ~16 bits/pit. At best you get ~8 bits/pit. And a lousy error rate. Certainly not usable for storage of your typical MPEG data, where a one-bit error can cause major visual damage. To get data reliability you'd have to use a very aggressive error-correcting code which would undo a lot of the gains in density. There aint no free lunch. You can get more density, but you have to give up some signal/noise ratio. Maybe the Simpson's reference is apropos-- cartoons can be rendered with fewer bits.
OH, I forgot, most places won't loan you money for 30 yrs on something like solar cells, which will probably die within half that time, so you would have to pay about double the interest rate for the $17K. And it may be hard to get any kind of mortgage on a house that is so unusual with questionable market value. So your house may end up costing MUCH MUCH more due to the extra interest costs. And I forgot to add the cost of maintenance and eventual replacement, which makes the whole concept even sillier.
Slashdot Mirror
- "a chemical plant run by Union Carbide". Whoa there. India is a semi-socialist country; it has some very strict laws limiting foreign ownership and control. It would be fairer to say "a company majority owned by the Indian Government, with Union Carbide as a minor partner". I realize this doesnt fit in with the "bad multinational american companies" bogey-man, but tough.
- "released about 40 tons of a toxic gas". Maybe it would be fair to state that this was an UNINTENTONAL release, probably caused by various factors including tired workers and labor unrest.
- "...plant was in the business of creating chemicals deadly to life." Well, if you call locusts, nematodes, chiggers, termites, ants, earwigs, wasps, et al, "life", I guess thats right. But a few people think it's a good idea in a country with a marginal food supply to save a few million human lives from death by starvation by producting chemicals to kill the above-listed lifeforms, so that humans might have a chance of eating some food.
- "Safety at the plant had not been a concern of management". Hmmm, let me think, let's assume the management has no concern for human life, including their own lives or those of their workers-- they just care about making money. In order to make money in that business, you have to be able to control dozens of chemical reactions, temperatures, pressures, flow rates, valves, pumps, manifolds, etc.. If you don't have the right control systems in place everywhere, things can quickly go out of control, and your many billion dollar plant is not making pesticides, it's making a mixture of 31% water, 38% salt, and 21% brown sludge. There HAVE to be extensive mechanisms in place to prevent chemical releases, at the very least those chemicals are expensive!
- "Today, the site remains a contaminated wasteland, unusable and never cleaned up." It comes down to economics: is the cost of cleanup less than the value of the land? India is not that small that it will miss a few acres.
- "The survivors have been minimally compensated". There's a fund of $328 million dollars. If that were spread evenly over the 3,000-some families of survivors, each family would get about $100,000 each. That's about 35 years of average earnings. I wouldnt call that "minimal".
Yes, it was a bad thing that happened, but it doesnt help to misdirect blame and energy.CS is one of the few professional fields left where a college degree is not all that necessary. Sure, it's "nice" to know the intricacies of B-trees, finite-state automata and the like, but these things rarely show up in their pure theoretical form in most everyday work. What IS important is to be able to absorb info, make design tradeoffs, and apply your posterior to the chair for extended periods of time until you get the dang program working! I've seen CS PHd's that can't figure out the * operator in C. I've also seen high-school dropouts that can do most anything with a computer. You decide which is more marketable.
I just looked it up, it's the subjunctive tense. It's used in indirect discourse, or when there's doubt, or when there's no doubt what's quoted is counter-factual: (Example of all three: Newscaster reporting "Jacko claims he's normal and innocent"),
I think German has the same thing. A whole bunch of word suffixes get added if the subject is somewhat in doubt. It makes a newscaster's job kinda hard if they're trying to track a developing story on the air.
- The power output goes down by a factor of 1,000. (power out is proportional to L x W x H )
- The friction in the bearings goes down by a factor of 100 (proportional to surface area of bearings)
- The windage losses due to air friction between the generator rotor and stator stay about the same.
- The air friction losses in the turbine may go waay up (as the ratio of turbulent flow to mainline flow goes way up).
- The thermal input from the burning gases goes down by a factor of 1,000.
- The thermal losses only drop by a factor of 100. Eventually the losses become greater than the thermal input, making it impossible to sustain burning.
So every time you shrink these things, the power out goes way down, the efficiency goes waaay down.Do this a few times and you'll have a turbine that can't even overcome its internal friction and a generator that, even if you could turn it, would be way down on the efficiency scale. Shrinking these things is a very very very *losing* thing to do.
This seems like a really poor application for computers. I can see the computer's analysis: Brushstrokes: 97% Leonardo Da Vinci Contrast: 98% Leonardo Da Vinci Demarcation: 96% Leonardo Da Vinci Fletangio: 97% Leonardo Da Vinci Impressimi: 98% Leonardo Da Vinci Random kid passing by: "Mom, look at that neat picture of dogs playing poker!"
" Thin CRTs offer the best of both worlds -- superior picture quality with a slim size. " I wouldnt call a 20% reduction, from 51 to 41cm deep , a "slim" CRT, nor worthy of Slashdot coverage. And they're probably compromising on something-- I'd guess they're going to lose a bit of convergence near the edges.
- They make a carbon nanotube that's at least a mile long.
- And that can can take 3x the stress required.
- And can withstand a direct hit by a 747.
- And can withstand a typical lightning-bolt.
- And can take a direct hit by a pea-sized piece of orbiting space-junk.
- And you find an insurance company that will write a policy on it. (Both liability and collision).
Until then I'll ust yawn at these semi-weekly pop-ups of this subject.Hmmm, only glitches: * Usually the task is loosely defined * Usually the environment is also. Example: Boss says "write a program that sends me a weekly summary of the file system backup status by email". I don't think a computer is going to be able to make much out of this. It takes a heck of a lot of human judgement to figure out (1) What level of detail the boss wants. (2) Do we format it as tables, line graphs, bar charts, ?? (3) Can his computer handle text only, or .pdf, or HTML, or FLASH, or 3D?? (4) Do I have the political "juice" to get access to all the sysadmin data? (5) Are the backup statistics reliable? (6) Is the boss's dumb son-in-law going to block my access unless I throw him a bone? (7) How am I doing to seamlessly do this for both Unix and NT? (8) Are the Windows API's really usable for this, or are they going to be hopeless as in the last 4 projects?
I don't think any computer parser is going to be able to make these judgements.
One might suspect Intel realized a while ago that Itanic was never going to make it. Now they're generating various scenarios, trying to figure out how to write off the investment at the least painful moment. One good instant might have been when Andy Grove left. Or maybe the next guy is revving up for this. Anyway, the shoe is going to drop sometime in the next 12 months. Don't expect much sudden drop in their stock as the market willprobably anticipate this.
IIRC when that Cd full of Windows source code got out, somebody did a "grep strcat" on it. Something like 13,000 hits. Now some of those are going to be safe, but how many of those are capable of overflowing under the right conditions? 1% ? 10%? The mind reels.
The US military has been developing anti-satellite weapons in plain sight for, oh, 30 years. IIRC we have a little missle, launched from a F-15, that can knock down satellites. We've had it for 30-some years. It doesnt violate any treaty as it's not based in space. Prolly been other more secret developments, not anything to get too worked up about.
Ok Class, I want a paragraph from each of you, starting out: The motivation to adhere to a treaty is... Single-spaced, by the end of the hour.... ( 60 mins later ) Ok, I see nobody has written anything. Let me explain: Sovreign nations sign treaties when it is in their interest to do so. The interest might be political, economic, social, historical, cultural, or otherwise. The participants may be acting in good faith, or being devious, or seeing a short-term advantage, or a strategic opportunity. In any case, nations adhere to a treaty as long as it is in their interest, and not much longer than that. That's why most treaties have a boilerplate paragraph along the lines of :: And any participant can back out of the treaty with 30 days notice.
So it's not too helpful to yell BUT THAT WOULD VIOLATE THE TREATY! A treaty is not an iron-clad dissuader, just a way to identify common interests, if any.
See: German non-agression treaty, et. al....
Try reading the Constitution, bit-breath.
The loser calling the winner and "conceding" is a political ritual. It has absolutely nothing to do with who gets elected president.
I'd reserve the phrase "totally refutes" for occasions where.... this actually happens. What I saw of the "refutation" was a few bits of unconvincing excuses and loose logic. The similarity in headers and the number and length of exact code matches is compelling and proabably irrefutable evidence.
A very unbalanced comparison:
/dev/sda0
Format c: is more analogous to mkfs
rm -Rf / is more like deltree c:
and IIRC the Windows del command waits 5 seconds on each busy file before giving up the delete, making NTFS deletes on busy files seem very slow.
Let's at least do our meaningless comparisions correctly!
All the coal and oil on the planet (about 3 teratons) is only about 8% of the carbon dissolved in the oceans. Which seems to imply two things: (1) We need to stir up the oceans a bit to get some of that CO2-poor deep water to the surface. (2) If we got desperate we could mine the waters for carbon.
Just a FEW quibbles with this nuclear gas rocket design: * 25,000 degrees C hot uranium hexafloride is going to melt and react with the the quartz toute-suite. * You'd need hundreds of pounds of 100% enriched UF6 to get a critical mass. Even under pressure, that's a lot of volume. * Reactors are controllable due to the 1 to 2 percent of fissions that result in delayed neutron emission. But this gas is going to have a lot more than 1 or 2 percent variations in density. Ergo you're going to have a really hard time (~impossible) controlling the reaction. * You're still going to need reaction mass to shove out the back. Just try to find a compound that is (1) Liquid, (2) Not too toxic (3) Doesnt disassociate at 25,000K Otherwise OK!
Enough holes in this story to drive several Beowulf clusters thru it: * They switch at 27Mhz do they? How many times can you flip a mechanical switch before it breaks? (About 0.1 seconds worth for most switches) * Ok, it takes no power somehow to hold the info. But what about reading/writing it? It's going to take not only power, but several transistors per bit. Good old DRAM nowdays is down to 1 well and one MOSFET per bit. ( plus row and column drivers). Can't even approach that with anything mechanical. * 27MHz is right in the middle of the old CB band! Watch out for truckers on your tail. * Volume goes down as the cube, surface area as the square of the linear dimensions. So these thingies are really at the whim of surface tension and surface electrostatic effects. The good part is they should be rather insensitive to mechanical shock. The bad part: watch out for static electricity, Hall Effect, and dust! * Sounds about as practical as bubble memory, string floppies, and 90-column oval-holed cards.
As a last resort, we could look at the science behind ELF before we worry too much about the "damage": (1) ELF transmitters are only a megawatt or so. The ELF waves are sooo long (many thousands of miles), that a little 50 mile antenna only radiates oh, maybe 5 watts of effective radiated power. (the rest just heats up the wires). Those 5 watts get spread more-or-less evenly all around the earth. (2) Your tpical large marine creature is maybe one billionth the size of the earth, so we're down to maybe 5-billionths of a watt hitting the beast. (3) Your typical animal is an even smaller fraction of the thousand-mile ELF wavelength. So about 99.99999% of the energy incident on say a giant squid goes right through it. We're now down to 5 quadrillionths of a watt. (4) A typical nerve discharge is around a THOUSAND to a MILLION times that amount of energy, so the ELF signal is that much weaker than the thousands of nerve impluses going off right inside the squid's body every second. (5) So I would not worry too much about ELF harming anything. (6) And, oh, as other have mentioned, the energy from power lines is many orders of magnitude stronger than ELF (and even that is hard to pick up any distance from power lines).
I don't think 332 angles gives you ~16 bits/pit. At best you get ~8 bits/pit. And a lousy error rate. Certainly not usable for storage of your typical MPEG data, where a one-bit error can cause major visual damage. To get data reliability you'd have to use a very aggressive error-correcting code which would undo a lot of the gains in density. There aint no free lunch. You can get more density, but you have to give up some signal/noise ratio. Maybe the Simpson's reference is apropos-- cartoons can be rendered with fewer bits.
OH, I forgot, most places won't loan you money for 30 yrs on something like solar cells, which will probably die within half that time, so you would have to pay about double the interest rate for the $17K. And it may be hard to get any kind of mortgage on a house that is so unusual with questionable market value. So your house may end up costing MUCH MUCH more due to the extra interest costs. And I forgot to add the cost of maintenance and eventual replacement, which makes the whole concept even sillier.