What about black holes, then? It's a singularity that looks absurd on paper, but turned out to be the way reality works with a fair degree of certainty.
The "singularity" at the event horizon is just an artefact of the chosen coordinate system (just like the coordinate singularity at the north pole). Note that the event horizon exists anyway (i.e. you still don't get any information out of the black hole), it's just not singular. The singuarity in the center of the black hole has never been seen (nor will it ever be seen, because it's shielded by the event horizon), and it's practically certain that quantum gravity will kick in before you reach it (meaning that it's prediction through classical general relativity is bogus because GR is simply not applicable where it predicts the singularity).
Yes, physical reality and the complexities of human society are drastically different fields, but the nature of both technological progress and societal advancement are both notoriously hard to predict.
However it's easy to predict that there are limits. Any gain of knowledge is, at last, based on physical processes, and physical processes are, again, limited in speed and size. Where the limits lie is surely harder to predict, but that they are there is virtually certain.
More encouragingly, there are a lot of less dramatic outcomes that might make slow colonization of the galaxy possible.
A singularity, of whatever kind, is always a sign that one has overstepped the validity range of the current theory. Therefore the Singularity will never hit us.
The Mormons perhaps have enough money to do something like that out of religious conviction. They trekked across the US into Utah despite great odds and losses in order to find a place that wouldn't keep outlawing their practices and pestering them.
So in order to get interstellar space travel, we should put all our energies into outlawing the practices of the Mormons and pestering them?
Well, a thousand AIs working on science problems could easily fully test the robustness of any given theory, but even if they had "real" intelligence, there's no guarantee that a thousand AIs would formulate a new theory that would help us get along into space, only make it more probable that if such a theory were to exist, it would be thought of.
The AIs will solve the problems by putting us into a matrix where they can change the laws of physics however needed, and thus enable space travel. Sure, it will be simulated space travel, but then, it doesn't matter since no one will know that anyway.:-)
"Tests don't prove you know anything; they only prove you know how to take tests!"
Of course this validates them as conditions for school entry: Since successfully finishing school is based mainly on passing tests, your test passing ability is exactly the right thing to test for.:-)
His argument is that the harder the test the less relevant knowledge of the actual answers to the questions posed on the test are to determining your relative score.
However he measures the hardness of the test just by the number of people who pass or fail it, but implicitly assumes that a harder test implies harder questions. That's in itself a fallacy. I can make a test harder by simply raising the number of questions you have to answer right. For example, let's take an extreme case: Say I raise the bar so hight that you'll only pass if you answer all questions correctly. Now if you know everything, you'll of course pass, but if you have to guess at just one question, you'll only have a 50% chance to pass. And your chances will go down exponentially with the number of questions you have to guess. Unless the questions themselves are extremely easy, that hypothetical test would be a very hard one (that it, very few would pass it), but at the same time very efficient to let only those with actual knowledge pass: If you just guess, you've got nearly no chance to pass.
In short, the number of people who actually pass/fail a test in isolation simply can't be used to decide if a test was meaningful.
Well, they could introduce a "don't skip this" flag, so that you couldn't fast-forward across ads (similar to what DVDs already do). OTOH the answer would probably be machines which do exactly the opposite: Automatically skip everything with the "don't skip this" flag, thus automatically removing ads.
This is what I got from the article: the mass of a kilogram is not based on anything fundamental (like a number of atoms), it is based on the mass of a prototype.
What's more important: There's evidence that despite all care, the mass of that prototype might be changing (see the first paragraph after the image). Which, of course, is the worst thing which can happen to a prototype you base your measurements on. (The page linked is BTW another experiment intended to provide a new kg definition).
Strictly speaking, a balance also compares weights. However due to the fact that fluctuations of the gravitational field are usually negligible over the extension of a balance, the weight is to very good approximation proportional to the mass (it would be exactly proportional if both objects would be at the very same place, but that's obviously not possible). Therefore the comparison of weight results, to a very good approximation, in a comparison of mass (because the local gravitational acceleration cancels out). To see that it is really a comparison of weights, think of the following two experiments:
Think of a balance in gravity-free space. It won't work (unless you accelerate it, but acceleration and gravitation are equivalent).
Think of a very big balance, so that one side is above earth, and the other side is above moon. Then the same mass on both sides won't balance out.
That's exactly what he meant with "tamper resistance". Note that as long as your device allows modified software to be installed, this clause doesn't affect you.
BTW, an interesting question would be if it would violate the GPLv3 if a networked box would run modified software quite fine, but always transmit a checksum of the running code to the server it connects to (done directly by the hardware, so no modification of the software can change that), and that server would then deny to serve the box if it doesn't run the original software. Technically, the box wouldn't be tamper-resistant (after all, the modified code works just fine, and if you'd connect to a different server which doesn't care about the checksum, there would be no problem), but it might make the box useless anyway (because the service you get from that server might not be available otherwise).
Well, given that changing the license of the already received code has to be a willful act by the receiving party, I don't think there should be a problem. Basically the upgrade clause says "if we later publish different terms, and you like those terms better then the current one, you are allowed (not required) to switch to those new license terms. There's nothing forced upon you (indeed, AFAIU you are even allowed to redistribute the software with the upgrade clause removed, if you prefer). IANAL however, therefore I might miss something here.
how would you qualify what is proper in many cases, specifically the legalities of licensing?
Proper research about the legalities of licensing is generally to ask a lawyer. Unless you are a lawyer, in that case proper research probably means reading the license(s) in question, the applicable law, and any prior court rulings relating to the issue (IANAL however, therefore there might be something about proper research by a lawyer which I'm missing).
That's the problem with geeks: They speak a language where I love you is not well-formed unless you redefine the words, and then they wonder why they don't get a girlfriend.:-)
Latin hasn't been a "home language" anywhere for hundreds of years
Well, Latin is still the "home language" of the Vatican. However, it's not the native language of the inhabitants, because the birth rate there should be fairly low.:-)
Maybe they plan to use a more active policy: Examine those products for not yet known vulnerabilitites, find out the ways those might get fixed, patent all viable methods, then tell the company about the vulnerabilities.
Slashdot Mirror
The "singularity" at the event horizon is just an artefact of the chosen coordinate system (just like the coordinate singularity at the north pole). Note that the event horizon exists anyway (i.e. you still don't get any information out of the black hole), it's just not singular. The singuarity in the center of the black hole has never been seen (nor will it ever be seen, because it's shielded by the event horizon), and it's practically certain that quantum gravity will kick in before you reach it (meaning that it's prediction through classical general relativity is bogus because GR is simply not applicable where it predicts the singularity).
However it's easy to predict that there are limits. Any gain of knowledge is, at last, based on physical processes, and physical processes are, again, limited in speed and size. Where the limits lie is surely harder to predict, but that they are there is virtually certain.
No argument with that.
A singularity, of whatever kind, is always a sign that one has overstepped the validity range of the current theory. Therefore the Singularity will never hit us.
So in order to get interstellar space travel, we should put all our energies into outlawing the practices of the Mormons and pestering them?
The AIs will solve the problems by putting us into a matrix where they can change the laws of physics however needed, and thus enable space travel. Sure, it will be simulated space travel, but then, it doesn't matter since no one will know that anyway.
Of course this validates them as conditions for school entry: Since successfully finishing school is based mainly on passing tests, your test passing ability is exactly the right thing to test for.
However he measures the hardness of the test just by the number of people who pass or fail it, but implicitly assumes that a harder test implies harder questions. That's in itself a fallacy. I can make a test harder by simply raising the number of questions you have to answer right. For example, let's take an extreme case: Say I raise the bar so hight that you'll only pass if you answer all questions correctly. Now if you know everything, you'll of course pass, but if you have to guess at just one question, you'll only have a 50% chance to pass. And your chances will go down exponentially with the number of questions you have to guess. Unless the questions themselves are extremely easy, that hypothetical test would be a very hard one (that it, very few would pass it), but at the same time very efficient to let only those with actual knowledge pass: If you just guess, you've got nearly no chance to pass.
In short, the number of people who actually pass/fail a test in isolation simply can't be used to decide if a test was meaningful.
Of course the number you're thinking of is 42, which would be option b, because the number is found in the Hitch Hiker's Guide to the Galaxy.
Well, they could introduce a "don't skip this" flag, so that you couldn't fast-forward across ads (similar to what DVDs already do). OTOH the answer would probably be machines which do exactly the opposite: Automatically skip everything with the "don't skip this" flag, thus automatically removing ads.
What's more important: There's evidence that despite all care, the mass of that prototype might be changing (see the first paragraph after the image). Which, of course, is the worst thing which can happen to a prototype you base your measurements on. (The page linked is BTW another experiment intended to provide a new kg definition).
Well, a cable clearly isn't a digital medium. However, for some strange reason ethernet cards are not called ethernet modems :-)
Well, of course the problem is that RMS wanted to get a high grade from Tanenbaum and therefore used a microkernel design.
That's exactly what he meant with "tamper resistance". Note that as long as your device allows modified software to be installed, this clause doesn't affect you.
BTW, an interesting question would be if it would violate the GPLv3 if a networked box would run modified software quite fine, but always transmit a checksum of the running code to the server it connects to (done directly by the hardware, so no modification of the software can change that), and that server would then deny to serve the box if it doesn't run the original software. Technically, the box wouldn't be tamper-resistant (after all, the modified code works just fine, and if you'd connect to a different server which doesn't care about the checksum, there would be no problem), but it might make the box useless anyway (because the service you get from that server might not be available otherwise).
What if I prefer to modify obfuscated code?
Well, given that changing the license of the already received code has to be a willful act by the receiving party, I don't think there should be a problem. Basically the upgrade clause says "if we later publish different terms, and you like those terms better then the current one, you are allowed (not required) to switch to those new license terms. There's nothing forced upon you (indeed, AFAIU you are even allowed to redistribute the software with the upgrade clause removed, if you prefer). IANAL however, therefore I might miss something here.
Proper research about the legalities of licensing is generally to ask a lawyer. Unless you are a lawyer, in that case proper research probably means reading the license(s) in question, the applicable law, and any prior court rulings relating to the issue (IANAL however, therefore there might be something about proper research by a lawyer which I'm missing).
That's the problem with geeks: They speak a language where I love you is not well-formed unless you redefine the words, and then they wonder why they don't get a girlfriend. :-)
In Russia consiliae, Lingua Latina TE loquitur!
Well, Latin is still the "home language" of the Vatican. However, it's not the native language of the inhabitants, because the birth rate there should be fairly low.
Maybe they plan to use a more active policy: Examine those products for not yet known vulnerabilitites, find out the ways those might get fixed, patent all viable methods, then tell the company about the vulnerabilities.
What about pacemakers? Generally places with high magnetic fields carry special warning signs that people with pacemakers should keep off.
I don't think anyone puts power lines under high tension. Maybe you meant a high voltage power line?
No, quite the opposite: It delivered too high voltage, so the lightbulb burned out.