Of course, in 120 years you'll get implants which make you always simply know what time it is, without having to look at some device. Looking at some devices on your hands would only distract you when operating your flying cars.:-)
So if two of them worked together, they'd get the complete movie, and there would be no way to know who released it (assuming they are careful at cutting, so it's not possible to identify the edited section).
Or they even remove a third scene and some unrelated person gets the blame.
Of course the only one you should sue to live forever is the death. He killed so many people, that he is certainly guilty of mass murder. With death removed, there's no limit to your life.
That's nothing. I've got a device which gets 0% false positives. Now, that device has a rather high rate of false negatives, so I augmented it with another device which gives you 0% false negatives, but some false positives. Together they give you the perfect diagnosis tool: No false positives (device 1) plus no false negatives (device 2).
Here's how it works: The first device is just a piece of paper with the text "you are healthy" written on it. It's obvious that you won't get false positives from that. The second device is almost the same, except that it contains the text "you are ill."
I'm planning to also release more specific tools do decide e.g. between "you have cancer" and "you have no cancer."
Unless "approved" means "gets through the lameness filter," that's nonsense. Anyone can post. It's just that if people don't consider what you post acceptable, they'll mod you down. That's life, live with it. I've gotten things moderated down which I considered good jokes. So what, maybe the jokes were not really so good, maybe the moderators just didn't get them, whatever. The worst which could happen would be that I lose my karma bonus (which I don't normally use anyway). Oh, and Adblock might get some work if I lose my good karma.
Now, it happens that I've written enough posts which moderators considered interesting, informative or insightful that I'm probably not in danger of losing my good karma. Now you may think that my posts didn't actually deserve those moderations. So what. The moderators thought they deserved them, so they got them. And I got Karma for writing posts that obviously the people with moderation points liked.
I don't know what will happen with this posting. Maybe it will be moderated to -1 Offtopic (because it is offtopic for the story), maybe some moderators will consider the post interesting or even insightful. Of course I'd prefer the latter, but I don't care too much. I didn't write this post to get Karma, I wrote it in the hope that you (or other people who think like you) get some insight from it.
We should avoid "nonstandard posting" so that we don't get our "posting privileges" revoked?
I never heard of anyone getting their posting privilege revoked. The worst which will happen is that your posts will get low score, which means that some people will not see your posts (I personally usually read at -1, so I will see your posts even then).
Perhaps the brilliance of the analogy flew over your head like a rock.
Maybe. Or maybe the analogy just wasn't that brilliant.
Well, the true reason why aliens are scared is that lasers rays of exactly that type are a side effect of their most effective weapon: The blackhole gun. It works by producing black holes, which then are fired onto the enemy ship. The production process causes laser light to be emitted in some direction; usually that laser is directed into space, exactly to scare the enemy away (actually using that weapon is very expensive, therefore it's a big advantage if the enemy just goes away in fear). Well, actually they have lost the knowledge how to make true blackhole guns long ago, but that's kept top secret, so every alien war party expects that the other party has the real thing, which is why the laser trick works.
Indeed, it's time that someone invents the carbon nanotube unlimited current room temperature superconductor, so that we can have cables carrying electricity from space without loss, and without collapsing under their own weight.
The center of mass not only depends on the positions of the parts, but also on, well, their masses. So if you make your satellite heavy enough, it can be even just one meter above the geosynchronous orbit. Of course there's a tradeoff involved: The heavier the satellite gets, the more expensive it gets to put it there. OTOH, having it at greater distance will also increase the cost. I guess there would be an ideal distance where the cost is minimized.
Assuming the whole thing can actually be built, of course.
First, I explicitly qualified "array new" - non-array new is of course essential. And yes, you cannot put auto_ptr in a container (you can put shared_ptr in a container, but that's not always desirable). So insofar as you are advocating that simple new/delete has its place in well-designed programs, I totally agree.
OTOH, std::vector is a replacement for newed arrays. There's no law that you have to change your collection's size just because you can. And apart from construction, the performance of vector and newed array should be exactly the same on any self-respecting implementation. And the performance differences on construction rarely matter.
Now, there are scenarios where that difference matters, and there are even scenarios where malloc is the right tool (interfacing with libraries written in C). That's why I wrote "high chances that you've made a design error", not "a sure sign that you've made a design error).
I didn't advocate removing pointers and arrays, or even just one of them. I was advocating removing the decay of arrays to pointers. That is, keep pointers as they are (well, there could be improvements in that area, too, but that's a separate concern), keep arrays, but make arrays first-class objects which don't decay to a pointer to the first element. If you need a pointer to the first element, you can always get it explicitly: &a[0]
And yes, I do know that this isn't really possible, because it would break B compatibility (no, that's no typo; the way C arrays work is originally due to B compatibility, since C++ inherits them from C, it's restrained by B compatibility as well).
C++ does have better memory safety, unless you insist in using the old C stuff still included, instead of using the superior C++ alternatives. For example, if you have a single malloc or array new in your program, there are high chances that you've made a design error.
And of course you can easily have buffer overflows in Fortran as well.
Templates are useful, but "generic programming", doing arbitrary computation at compile time with templates, was a terrible idea.
Doing arbitrary computation at compile time with templates is called "template metaprogramming." The "generic programming" is what the STL does: Provide a single templated implementation for an algorithm working out of the box on vastly different data structures.
Depends on what you consider "ruining" it. Things I'd like to be removed from C++ (but backwards compatibility disallows it): * writing reinterpret_casts as constructor-style casts (constructor-style casts should be restricted to static/const cast) * implicit decay of arrays to pointers (yes, that's a C feature inherited by C++; arrays should be first-class objects) * implicit conversion from bool to int (explicit conversion should stay) * C++ declaration syntax (of course by being replaced with a more sane syntax; this would also save us from writing "typename" and "template" inside templates just to disambiguate the grammar; but just not having C++'s most vexing parse would be a big advantage) * local declaration of global functions
It takes 20 minutes to pick a random number between 1 and 8?!
Yes. After all, if you just "randomly" say a number, it's not realloy very random. Therefore you'll have to resort to a truly random process. Since few people carry perfect octaeders with them (and AFAIK there are none supplied by McDonalds), the best choice it to toss coins.
Now you might think that tossing coins three times should be enough, because every toss gives you one bit. But that's not true: There may be a bias in your tossing, and who tells you that the coin you chose really is fair? Therefore you should make a protocol which as far as possible eliminates any non-randomness.
The first step is to remove a possible simple bias by tossing the coins two times, discarding the tosses if both results are equal, and taking the second one otherwise. This alone increases the average number of tosses by a factor of 4 (a factor 2 because you have two tosses for one result, and another factor of two because half of your tosses are invalid).
However, that's not enough: There might e.g. be a systematic drift in the probability (for whatever reason). Therefore you should repeat each toss series, but this time choose the first result instead of the second if the two tosses disagree. Since now you may have conflicting results, you repeat that procedure for the conflicting bits to decide which one to choose. This gives another factor of 4.
All together you now have 48 tosses. But that's still not enough: To make completely sure that your results are random, you should also determine the order of the bits. There are 6 different permutations; since 6 is not a power of 2, in order to chose the right permutation, you apply the following protocol: Double-toss the coin four times. If the result is not 2 bits each repeated two times, discard the result and start over. Otherwise, there are exactly 6 possible combinations, thus determining the required permutation of the bits previously obtained. Note that on average, you'll need 32/3, i.e. almost 11 tosses to determine the permutation. Together that's almost 59 tosses to randomly select one number. Assuming you need about 20 seconds per toss, you indeed get an average of about 20 minutes for your choice.
Not really. We just want things for free. The GPL gives us things for free.
No it doesn't. Microsoft thought it did. Which led to their embarrassing about face. "Yay, we're doing something different, because we're Microsoft and we just like acting crazy every now and then... *ahem, cough* (also PNAMBIC)"
If you want stuff for free you need the BSD license. Whether it's what's best for you at time T is another question.
Well, if the GPL gives you something for free depends on if you are a developer or an user. If you are just an user, you indeed get the programs for free. If you are a developer basing your code on the GPLed code and distributing it, you pay with your code. If you like that price (i.e. the value you get by using the GPLed code is larger than the value of your own contribution), then go with the GPLed code. If you don't like the price, look elsewhere for code to use, or write it yourself.
Yeah, figuring out the effect hormones exert on the interactions of dendrites is such a trivial thing.
I'd consider that part of simulating the brain, not part of simulating the body. Therefore it's an argument which supports my claim instead of contradicting it.
The unabomber wasn't rational but fortunately he was a mathematician
An irrational matematician may sound like an oxymoron, but really, there are uncountably many of them. Rational matematicians are the exception, and even they are dense.:-)
I think simulating the body reactions is a few orders of magnitude simpler than simulating the brain. Especially since it only needs to simulate the experience (e.g. the simulated stomach doesn't need to simulate the digestion of simulated food, it's enough if it emulates the filling state (and probably a few other simple data points).
Slashdot Mirror
Of course, in 120 years you'll get implants which make you always simply know what time it is, without having to look at some device. Looking at some devices on your hands would only distract you when operating your flying cars. :-)
So if two of them worked together, they'd get the complete movie, and there would be no way to know who released it (assuming they are careful at cutting, so it's not possible to identify the edited section).
Or they even remove a third scene and some unrelated person gets the blame.
Well, hopefully the police doesn't think so when you take it out to check your blood ...
Of course the only one you should sue to live forever is the death. He killed so many people, that he is certainly guilty of mass murder. With death removed, there's no limit to your life.
That's nothing. I've got a device which gets 0% false positives. Now, that device has a rather high rate of false negatives, so I augmented it with another device which gives you 0% false negatives, but some false positives. Together they give you the perfect diagnosis tool: No false positives (device 1) plus no false negatives (device 2).
Here's how it works:
The first device is just a piece of paper with the text "you are healthy" written on it. It's obvious that you won't get false positives from that.
The second device is almost the same, except that it contains the text "you are ill."
I'm planning to also release more specific tools do decide e.g. between "you have cancer" and "you have no cancer."
Unless "approved" means "gets through the lameness filter," that's nonsense. Anyone can post. It's just that if people don't consider what you post acceptable, they'll mod you down. That's life, live with it. I've gotten things moderated down which I considered good jokes. So what, maybe the jokes were not really so good, maybe the moderators just didn't get them, whatever. The worst which could happen would be that I lose my karma bonus (which I don't normally use anyway). Oh, and Adblock might get some work if I lose my good karma.
Now, it happens that I've written enough posts which moderators considered interesting, informative or insightful that I'm probably not in danger of losing my good karma. Now you may think that my posts didn't actually deserve those moderations. So what. The moderators thought they deserved them, so they got them. And I got Karma for writing posts that obviously the people with moderation points liked.
I don't know what will happen with this posting. Maybe it will be moderated to -1 Offtopic (because it is offtopic for the story), maybe some moderators will consider the post interesting or even insightful. Of course I'd prefer the latter, but I don't care too much. I didn't write this post to get Karma, I wrote it in the hope that you (or other people who think like you) get some insight from it.
I never heard of anyone getting their posting privilege revoked. The worst which will happen is that your posts will get low score, which means that some people will not see your posts (I personally usually read at -1, so I will see your posts even then).
Maybe. Or maybe the analogy just wasn't that brilliant.
Well, the true reason why aliens are scared is that lasers rays of exactly that type are a side effect of their most effective weapon: The blackhole gun. It works by producing black holes, which then are fired onto the enemy ship. The production process causes laser light to be emitted in some direction; usually that laser is directed into space, exactly to scare the enemy away (actually using that weapon is very expensive, therefore it's a big advantage if the enemy just goes away in fear). Well, actually they have lost the knowledge how to make true blackhole guns long ago, but that's kept top secret, so every alien war party expects that the other party has the real thing, which is why the laser trick works.
Indeed, it's time that someone invents the carbon nanotube unlimited current room temperature superconductor, so that we can have cables carrying electricity from space without loss, and without collapsing under their own weight.
The center of mass not only depends on the positions of the parts, but also on, well, their masses. So if you make your satellite heavy enough, it can be even just one meter above the geosynchronous orbit. Of course there's a tradeoff involved: The heavier the satellite gets, the more expensive it gets to put it there. OTOH, having it at greater distance will also increase the cost. I guess there would be an ideal distance where the cost is minimized.
Assuming the whole thing can actually be built, of course.
Given that they have Arnie as governor, it's no surprise that they push for their legislation to be more violent. Games are just the start!
First, I explicitly qualified "array new" - non-array new is of course essential. And yes, you cannot put auto_ptr in a container (you can put shared_ptr in a container, but that's not always desirable). So insofar as you are advocating that simple new/delete has its place in well-designed programs, I totally agree.
OTOH, std::vector is a replacement for newed arrays. There's no law that you have to change your collection's size just because you can. And apart from construction, the performance of vector and newed array should be exactly the same on any self-respecting implementation. And the performance differences on construction rarely matter.
Now, there are scenarios where that difference matters, and there are even scenarios where malloc is the right tool (interfacing with libraries written in C). That's why I wrote "high chances that you've made a design error", not "a sure sign that you've made a design error).
I didn't advocate removing pointers and arrays, or even just one of them. I was advocating removing the decay of arrays to pointers. That is, keep pointers as they are (well, there could be improvements in that area, too, but that's a separate concern), keep arrays, but make arrays first-class objects which don't decay to a pointer to the first element. If you need a pointer to the first element, you can always get it explicitly: &a[0]
And yes, I do know that this isn't really possible, because it would break B compatibility (no, that's no typo; the way C arrays work is originally due to B compatibility, since C++ inherits them from C, it's restrained by B compatibility as well).
Given that C++ inherited that ugly syntax from C: no, it isn't.
C++ does have better memory safety, unless you insist in using the old C stuff still included, instead of using the superior C++ alternatives. For example, if you have a single malloc or array new in your program, there are high chances that you've made a design error.
And of course you can easily have buffer overflows in Fortran as well.
Doing arbitrary computation at compile time with templates is called "template metaprogramming." The "generic programming" is what the STL does: Provide a single templated implementation for an algorithm working out of the box on vastly different data structures.
This is why I use languages that don't change every few years.
What language would that be? I can't think of a single major programming language that hasn't been updated at least twice.
INTERCAL?
Depends on what you consider "ruining" it. Things I'd like to be removed from C++ (but backwards compatibility disallows it):
* writing reinterpret_casts as constructor-style casts (constructor-style casts should be restricted to static/const cast)
* implicit decay of arrays to pointers (yes, that's a C feature inherited by C++; arrays should be first-class objects)
* implicit conversion from bool to int (explicit conversion should stay)
* C++ declaration syntax (of course by being replaced with a more sane syntax; this would also save us from writing "typename" and "template" inside templates just to disambiguate the grammar; but just not having C++'s most vexing parse would be a big advantage)
* local declaration of global functions
It takes 20 minutes to pick a random number between 1 and 8?!
Yes. After all, if you just "randomly" say a number, it's not realloy very random. Therefore you'll have to resort to a truly random process. Since few people carry perfect octaeders with them (and AFAIK there are none supplied by McDonalds), the best choice it to toss coins.
Now you might think that tossing coins three times should be enough, because every toss gives you one bit. But that's not true: There may be a bias in your tossing, and who tells you that the coin you chose really is fair? Therefore you should make a protocol which as far as possible eliminates any non-randomness.
The first step is to remove a possible simple bias by tossing the coins two times, discarding the tosses if both results are equal, and taking the second one otherwise. This alone increases the average number of tosses by a factor of 4 (a factor 2 because you have two tosses for one result, and another factor of two because half of your tosses are invalid).
However, that's not enough: There might e.g. be a systematic drift in the probability (for whatever reason). Therefore you should repeat each toss series, but this time choose the first result instead of the second if the two tosses disagree. Since now you may have conflicting results, you repeat that procedure for the conflicting bits to decide which one to choose. This gives another factor of 4.
All together you now have 48 tosses. But that's still not enough: To make completely sure that your results are random, you should also determine the order of the bits. There are 6 different permutations; since 6 is not a power of 2, in order to chose the right permutation, you apply the following protocol: Double-toss the coin four times. If the result is not 2 bits each repeated two times, discard the result and start over. Otherwise, there are exactly 6 possible combinations, thus determining the required permutation of the bits previously obtained. Note that on average, you'll need 32/3, i.e. almost 11 tosses to determine the permutation. Together that's almost 59 tosses to randomly select one number. Assuming you need about 20 seconds per toss, you indeed get an average of about 20 minutes for your choice.
SCNR :-)
Not really. We just want things for free. The GPL gives us things for free.
No it doesn't. Microsoft thought it did. Which led to their embarrassing about face. "Yay, we're doing something different, because we're Microsoft and we just like acting crazy every now and then... *ahem, cough* (also PNAMBIC)"
If you want stuff for free you need the BSD license. Whether it's what's best for you at time T is another question.
Well, if the GPL gives you something for free depends on if you are a developer or an user. If you are just an user, you indeed get the programs for free. If you are a developer basing your code on the GPLed code and distributing it, you pay with your code. If you like that price (i.e. the value you get by using the GPLed code is larger than the value of your own contribution), then go with the GPLed code. If you don't like the price, look elsewhere for code to use, or write it yourself.
Yeah, figuring out the effect hormones exert on the interactions of dendrites is such a trivial thing.
I'd consider that part of simulating the brain, not part of simulating the body. Therefore it's an argument which supports my claim instead of contradicting it.
An irrational matematician may sound like an oxymoron, but really, there are uncountably many of them. Rational matematicians are the exception, and even they are dense. :-)
I don't think they want to run that simulation on laptops. How many laptops do you need to get the computing power of a current supercomputer?
It has. You didn't really think that basement you're living in is real, did you?
Of course. Just add a simulated anaesthticum to your simulated brain.
I think simulating the body reactions is a few orders of magnitude simpler than simulating the brain. Especially since it only needs to simulate the experience (e.g. the simulated stomach doesn't need to simulate the digestion of simulated food, it's enough if it emulates the filling state (and probably a few other simple data points).