...it's an inference engine with fuzzy logic rather than discrete logic, such that if you represented the inference training set in an N-ary tree, the fuzzy value is proportional to the fraction of branches in the tree that match a given inference. (They'd be better off with an S-curve, as that seems to be a better model for modeling real-world situations than a linear system.)
Yes, but modelling slashdotters is an almost impossible task. Surely they should start with something simpler. (Actually, what they've invented is a rather nice inference engine. Not completely convinced of the originality of that, but I'll accept that existing inference engines tended to be limited to basic inferences, linear separation of problem-spaces, etc, whereas this one seems on the surface to be a bit more powerful.)
According to Plato's "Republic", democracy is only sustainable if the masses are, in fact, educated. I'm paraphrasing here, but he essentially predicts that all other democratic systems will revert to what is basically a totalitarian state by any other name. The only difference is that coups are by ballot and therefore much cheaper. The obvious solution to this is not to add further totalitarianism to the mix, but to improve the education of the masses. Given the complexity of modern life, I personally hold that we need to evolve towards Homo Universalis if we're to achieve this. We'll never reach that state, except in extraordinary individuals, even if it were taken as the ideal. However, until the average person actually comprehends the notion that cause will always have effect and that an unintended consequence is a consequence nonetheless, society cannot solve anything. That includes Lovelock's non-democratic solution. (See: Fred Hoyle's Molecule Men, Ossian's Ride and A For Andromeda.)
The big problem with my proposal is... well, ok, there are lots of big problems. Expense, the fact that teachers are rarely the ones who understand the subjects, the dumbness of humanity, social inertia, the amazing lack of understanding of how to educate, and the fact that it'd take 2-3 generations minimum to clear out the ignorance -- way too long a timeframe to be useful here.
He's ruined the credibility of a geek lawyer firm, he's possibly bankrupted the financiers of both sides, he's crippled the firms that ponied up the Linux license fee, he's turned many IT managers insane and/or paranoid delusional, and he's ruined the credibility of some sectors of the IT press.
Doesn't matter that he lost the case, the collateral damage is stupendous. If you assume he had demonic intent, rather than merely evil intent, this is the perfect outcome.
Winning might actually have been less demonic, as it would have meant he could pay people back. As it stands, there's not a chance in hell he'll even be able to pay the court costs, never mind anyone else. Thus, enormous sums of money have simply vanished, to all intents and purposes.
This means those companies involved in the case have suffered Seven years of famine. Literally, a disaster of Biblical proportions.
Oh, c'mon. You think Toyota is going to write their embedded controller software in C? (Or Algol, that uses semicolons too.) It'll be in COBOL, surely.
You're assuming said vocal chords are in the physical plane of existence. I see no evidence to support this theory. Darl is clearly a Greater Demon from the ninth plane of hell.
SCO will survive for as long as Hollywood scriptwriters need material for zombie and vampire movies. Besides, you need holy water to kill the undead. In this case, I suggest filling a tanker with the stuff and using a water cannon.
I'd go for a tetrahedral organization myself. You'd need more of these clocks, sure, but you've a better chance of filtering out the noise that is still going to be present.
Noise source 1: Space expands, but it doesn't do so uniformly on a scale that would be practical for your experiment. If it was uniform, you could eliminate the relativistic effect easily.
Noise source 2: Space isn't of uniform density. The biggest problem will be localized variations in the solar wind, but the gas streaming off planets, space dust, and other such irritants, will give you gravitational fluctuations. For that matter, so will the orbits of planets. If you can plot these, you can computationally eliminate them.
Noise source 3: Chaos Theory. Can't avoid this one. James Gleik's book "Chaos" demonstrates the problem when he talks about a billiard ball bouncing off the cushions on the table. If the only perturbations to the trajectory are the bounces and gravitational fields, if a ball travels far enough to bounce seven times then the orbits of asteroids around Alpha Centuri would need to be considered to determine the final resting point.
The idea of collecting far more data than you strictly need is that you can apply error-correction methods to reduce or eliminate the errors that have been introduced by the noise.
For that matter, you could exploit relativistic effects. With a geophysical map of the area, you should be able to determine your position from the nature of the drift in the clock (so long as you have a second clock source of equal accuracy that is fixed relative to the geography you can measure the drift with).
Likewise, two aircraft could measure their relative velocity with incredible accuracy.
In terms of physics experiments, gravitational wave detectors need extremely accurate measurements. Any experiment that will (once and for all) determine if Relativistic or QM gravity is the correct model will also require this kind of accuracy to tell the difference.
Same for superstrings. You aren't going to be able to find them if you rely on a self-winding wristwatch for timings.
The world's most accurate clock means the world's most accurate frames-per-second measure. How can anyone not want to measure the FPS on their games to a thousand decimal places?
Ogg -could- potentially take off, I'm sure. You'd need a small portable app that'll run across a wide range of mobile phones (ideally including ones that don't support streaming video normally) AND you'd need to anticipate where the next big user of video is coming from.
No good just doing the phones, because by the time you can port a viewer across a large enough selection of phones, the market will be somewhere else. The only solution to forever chasing after the field is to second-guess where the field will end up. Cut directly there, rather than wait for others to figure out what the goal "should be".
But even that is not enough, since the market uses a well-defined set of codecs already. You need to have a way of viewing any legacy codec without the user realizing that that is what they are doing.
Ok, so now we know what Phineas and Ferb are going to do today.:)
Seriously, I'm going to generalize what I believe to be your key point -- it's no good running after where others have been. You need to get there first or (at worst) shortly behind. Waiting until the market (ANY market) stabilizes and then copying it will never work.
In other words, get something that fundamentally works and can be extended/ported so incredibly easily that interoperability can be tagged on quickly at the end. That way, you either create the standard OR can adapt to the standard when there is one, without faffing around.
(eg: NV and other protocols came out long before h264, but died because they were fundamentally wrong, hard to translate, non-portable and a bloody pain. Adequate for video streaming over the Internet, barely acceptable for storage, but definitely the work of a Great Old One. Nothing else is that insane.)
Slashdot Mirror
...it's an inference engine with fuzzy logic rather than discrete logic, such that if you represented the inference training set in an N-ary tree, the fuzzy value is proportional to the fraction of branches in the tree that match a given inference. (They'd be better off with an S-curve, as that seems to be a better model for modeling real-world situations than a linear system.)
Yes, but modelling slashdotters is an almost impossible task. Surely they should start with something simpler. (Actually, what they've invented is a rather nice inference engine. Not completely convinced of the originality of that, but I'll accept that existing inference engines tended to be limited to basic inferences, linear separation of problem-spaces, etc, whereas this one seems on the surface to be a bit more powerful.)
There are different sizes of infinity, and therefore it is entirely possible for an infinite task to grow into a larger infinite task.
It's human intelligence than I'm unsure about.
According to Plato's "Republic", democracy is only sustainable if the masses are, in fact, educated. I'm paraphrasing here, but he essentially predicts that all other democratic systems will revert to what is basically a totalitarian state by any other name. The only difference is that coups are by ballot and therefore much cheaper. The obvious solution to this is not to add further totalitarianism to the mix, but to improve the education of the masses. Given the complexity of modern life, I personally hold that we need to evolve towards Homo Universalis if we're to achieve this. We'll never reach that state, except in extraordinary individuals, even if it were taken as the ideal. However, until the average person actually comprehends the notion that cause will always have effect and that an unintended consequence is a consequence nonetheless, society cannot solve anything. That includes Lovelock's non-democratic solution. (See: Fred Hoyle's Molecule Men, Ossian's Ride and A For Andromeda.)
The big problem with my proposal is... well, ok, there are lots of big problems. Expense, the fact that teachers are rarely the ones who understand the subjects, the dumbness of humanity, social inertia, the amazing lack of understanding of how to educate, and the fact that it'd take 2-3 generations minimum to clear out the ignorance -- way too long a timeframe to be useful here.
Is it technically torture if the creature is one of the living dead or from some other demonic plane of existence?
*watches as the whoosh flies past*
What you posted was fairly funny -- especially as it might very well turn out to be the right answer.
It's artificial Life, Jim, but not as we know it.
He's ruined the credibility of a geek lawyer firm, he's possibly bankrupted the financiers of both sides, he's crippled the firms that ponied up the Linux license fee, he's turned many IT managers insane and/or paranoid delusional, and he's ruined the credibility of some sectors of the IT press.
Doesn't matter that he lost the case, the collateral damage is stupendous. If you assume he had demonic intent, rather than merely evil intent, this is the perfect outcome.
Winning might actually have been less demonic, as it would have meant he could pay people back. As it stands, there's not a chance in hell he'll even be able to pay the court costs, never mind anyone else. Thus, enormous sums of money have simply vanished, to all intents and purposes.
This means those companies involved in the case have suffered Seven years of famine. Literally, a disaster of Biblical proportions.
Superconductors repel magnetic fields, so all you really need to do is stick your head in a bucket of liquid nitrogen.
Magnets don't smell. Unless they're on fire.
There isn't any space on their refrigerator to keep them, though.
The CERN physicists play with 100T magnets. Which seem to be located in Switzerland. Under the bank. Hmmm.
Oh, c'mon. You think Toyota is going to write their embedded controller software in C? (Or Algol, that uses semicolons too.) It'll be in COBOL, surely.
So the PR folk are protons?
Oh yeah, that had runaway acceleration too.
To boldly. Fragment. What no man. Has fragmented before.
You're assuming said vocal chords are in the physical plane of existence. I see no evidence to support this theory. Darl is clearly a Greater Demon from the ninth plane of hell.
SCO will survive for as long as Hollywood scriptwriters need material for zombie and vampire movies. Besides, you need holy water to kill the undead. In this case, I suggest filling a tanker with the stuff and using a water cannon.
You mean Linux isn't written in ADA?
*ducks*
I'd go for a tetrahedral organization myself. You'd need more of these clocks, sure, but you've a better chance of filtering out the noise that is still going to be present.
Noise source 1: Space expands, but it doesn't do so uniformly on a scale that would be practical for your experiment. If it was uniform, you could eliminate the relativistic effect easily.
Noise source 2: Space isn't of uniform density. The biggest problem will be localized variations in the solar wind, but the gas streaming off planets, space dust, and other such irritants, will give you gravitational fluctuations. For that matter, so will the orbits of planets. If you can plot these, you can computationally eliminate them.
Noise source 3: Chaos Theory. Can't avoid this one. James Gleik's book "Chaos" demonstrates the problem when he talks about a billiard ball bouncing off the cushions on the table. If the only perturbations to the trajectory are the bounces and gravitational fields, if a ball travels far enough to bounce seven times then the orbits of asteroids around Alpha Centuri would need to be considered to determine the final resting point.
The idea of collecting far more data than you strictly need is that you can apply error-correction methods to reduce or eliminate the errors that have been introduced by the noise.
For that matter, you could exploit relativistic effects. With a geophysical map of the area, you should be able to determine your position from the nature of the drift in the clock (so long as you have a second clock source of equal accuracy that is fixed relative to the geography you can measure the drift with).
Likewise, two aircraft could measure their relative velocity with incredible accuracy.
In terms of physics experiments, gravitational wave detectors need extremely accurate measurements. Any experiment that will (once and for all) determine if Relativistic or QM gravity is the correct model will also require this kind of accuracy to tell the difference.
Same for superstrings. You aren't going to be able to find them if you rely on a self-winding wristwatch for timings.
The world's most accurate clock means the world's most accurate frames-per-second measure. How can anyone not want to measure the FPS on their games to a thousand decimal places?
Ogg -could- potentially take off, I'm sure. You'd need a small portable app that'll run across a wide range of mobile phones (ideally including ones that don't support streaming video normally) AND you'd need to anticipate where the next big user of video is coming from.
No good just doing the phones, because by the time you can port a viewer across a large enough selection of phones, the market will be somewhere else. The only solution to forever chasing after the field is to second-guess where the field will end up. Cut directly there, rather than wait for others to figure out what the goal "should be".
But even that is not enough, since the market uses a well-defined set of codecs already. You need to have a way of viewing any legacy codec without the user realizing that that is what they are doing.
Ok, so now we know what Phineas and Ferb are going to do today. :)
Seriously, I'm going to generalize what I believe to be your key point -- it's no good running after where others have been. You need to get there first or (at worst) shortly behind. Waiting until the market (ANY market) stabilizes and then copying it will never work.
In other words, get something that fundamentally works and can be extended/ported so incredibly easily that interoperability can be tagged on quickly at the end. That way, you either create the standard OR can adapt to the standard when there is one, without faffing around.
(eg: NV and other protocols came out long before h264, but died because they were fundamentally wrong, hard to translate, non-portable and a bloody pain. Adequate for video streaming over the Internet, barely acceptable for storage, but definitely the work of a Great Old One. Nothing else is that insane.)