Heck, to defeat this you could just use AES with a default key. Everyone can use the same key, and have it be publicly known. It's fine because this thing doesn't have the compute power to decrypt in real time, even if it knows what it needs to be decrypting and what the key is. Screw handshaking, key management, etc -- just make the CPU cost nonzero and you're done.
On the contrary. Go create a single-color or grayscale smooth one-dimensional gradient on a large-ish image (1024x1024 or so). It will show clear evidence of banding at 8 bits per channel, since there are only 256 color levels available.
This will be substantially reduced if everything were properly dithered, but in normal software and normal displays it is not.
How worth it is I don't know, but there is absolutely an easily detectable difference. How about testing your hypothesis before claiming you know what you're talking about, hmm? It's not exactly a difficult experiment to carry out.
Yup, the aircraft have an easier time of it. And even so, the autopilots can't eliminate pilot error as a source of crashes. You won't eliminate either driver or pilot error as a source of crashes in the forseeable future, even with plausible advancements in technology. You can *certainly* improve on the current state of affairs, though. I was just saying that OP was being excessively optimistic.
If you dithered that gradient properly it would be much harder to tell. You've essentially created a signal with a period of 8 pixels and overlaid it on top of the smooth gradient as a result of the quantization errors. It's that signal that's easy to see. If you dithered it properly, the noise would be shaped so that it didn't show up at one specific frequency band (ie 8 pixels and its harmonics), it would be much less noticeable. Note that eg digital photography processes do this inherently to some degree in the analog sensor.
Of course, very little software will do that dithering properly for you, so it's kind of a moot point. But then, not much will make proper use of 10 bit per channel color, either. I'd love to see one of these things up close...
Your eye only has 3 color sensors. Therefore you can reproduce any spectrum in a way that your eye will see as equivalent with only 3 color elements. That said, RGB doesn't do a perfect job of this -- there are some colors at the edges of the color space that your eye can see that RGB can't produce.
Now, it's entirely possible that the easiest way to produce the full spectrum when it comes to actually building a display is with more than 3 different color elements, but 3 is sufficient if they're the right 3 (RGB isn't quite).
Commercial airplanes basically fly on autopilot. Accident rates are low. The most common cause of accidents is still pilot error.
I certainly agree, once the technology gets there, automated driving will vastly improve safety. But to assume that driver error will be *eliminated* as a cause of accidents, at least for the forseeable future, is wrong.
Do you have a theoretical basis for that statement? If you don't like the number, it wouldn't be hard to change -- the current version won't let you configure it over 20, but hacking the source wouldn't be hard at all. I recommend raising it slowly, though -- you may experience performance problems. I don't know; I haven't tried this.
Actually, Wine is an alternate implementation of the APIs, not an emulation of them. There's a difference, at least if you're using the words in a technical sense rather than a regular English sense. Which, when being pedantic about a technical matter, is the correct sense to use them in.
Yup. Got that part. But what are the actual operations being performed, in more detail than "professional work"? Yeah, Maya, 3DS -- but how are those different from other programs that render 3D images?
Because anything capable of applying 1.4km/s delta-v to a 70kt asteroid is a larger propulsion system than has ever been built. Even if it's a mass driver, rather than a rocket engine, that still implies a very large engineering project -- larger than has been done in space before. Once we have the experience doing that sort of thing, it will be easier, but doing it on the second manned mission to any asteroid, ever, is likely to be a tall order.
You don't. But it's a convenient and legal way to test large file transfers. None of the devs and major testers actually need freenet for anything in particular that I'm aware of. Does that mean they shouldn't work on it or test it?
No, the energy to stop it is far less, since it's on an orbit very similar to ours. Most of the impact energy comes from it falling down Earth's gravity well. The number you want is the hyperbolic excess velocity, ie the speed it would have at impact if the Earth's gravity had no effect (roughly). See my other post for numbers.
Yes, you have to secure the machine freenet is running on. Unfortunately, if you actually want to run a node, the livecd doesn't work well, since it won't maintain a store long-term. If all you want to do is insert something and run, though, it'll work great.
I don't know much about MUTE, but probably because Freenet has a scalable routing protocol, possibly better security, and possibly a more active development community.
That's the thing, this one isn't hard to capture. It's in about as good an orbit for the purpose as you could ask for. There are probably some that would be easier, but what you really want for a first try is something that's smaller. Of course, I have no clue what it's made of -- you also want to check that it would be useful before capturing it;)
It's got 1.37 km/s hyperbolic excess velocity, and on an orbit that damn near intersects ours. That means it takes a little more than 1370 m/s of delta-v to perform the capture. At 7.1E7 kg, that's about 6.6E13 joules -- approximately 15kt TNT equivalent worth of energy.
Assuming a high performance LOX/Methane engine, it would need about 34kt of propellant (rockets are inefficient for delta-v low relative to exhaust velocity). Note that this is a significant proportion of the asteroid mass. To make it economical, you'd need something more exotic -- a mass drive throwing bits of asteroid, or a high performance solar-electric ion drive, for example.
AnotherIndex is no longer in the default bookmark set. It wasn't updated all that often, it provided no site descriptions for the vast majority of sites, and it did a crappy job categorizing things (ie, putting things in completely absurd categories, not just failing to categorize them).
If you're interested in privacy, censorship resistant networks, and related topics, you're missing out on a fun place to hang out and discuss things. Fire up a node, load up FMS (the Freenet Messaging System; think usenet over Freenet -- there's a link in the default bookmark set), and subscribe to some boards that look interesting. The signal to noise ratio is quite good, though some of the posters are predictably extreme in any conceivable fashion. There are a couple trolls, but probably fewer than/. Start up a Freenet blog if you're so inclined, or go read other people's. Much like the regular internet blagosphere, most of it is completely boring, and some of it is really interesting.
Anyway, I encourage you to download it, fire up a node, and take a look around. You might be pleasantly surprised. Of course, you might decide that there's not much interesting and uninstall it, but if so, you haven't really lost anything more than a little bit of time. So come try it out!
P.S. FMS can be annoying to get set up, as it's still in an early alpha state and the docs suck. But join #freenet on irc.freenode.net and you'll find people who can answer questions about it, or the node itself if you're having trouble with that.
Yes, I agree, it's definitely good to hear. One of the medium-term goals for Freenet (0.8.0, hopefully) is steganographic transports that will hopefully make it much harder to see that you're running a (darknet) node.
The same laws that give ISPs protection from their users' downloading child porn should be applicable to individuals running freenet. This has not specifically been tested, however the legal framework for it appears fairly solid. Also, because of the fact that's it not stored as.jpgs on your disk (even encrypted) but rather scattered pieces of files of which you'll only have parts, in a form that can't be used to reconstruct the whole file, it would be rather difficult to convince a court that it was *you* downloading the files as opposed to your node propagating others' requests. And, it's not proof, though it should be reassuring -- no one has gotten in trouble for running freenet, even though the authorities are certainly paying attention (as far as we know, anyway).
Slashdot Mirror
Heck, to defeat this you could just use AES with a default key. Everyone can use the same key, and have it be publicly known. It's fine because this thing doesn't have the compute power to decrypt in real time, even if it knows what it needs to be decrypting and what the key is. Screw handshaking, key management, etc -- just make the CPU cost nonzero and you're done.
On the contrary. Go create a single-color or grayscale smooth one-dimensional gradient on a large-ish image (1024x1024 or so). It will show clear evidence of banding at 8 bits per channel, since there are only 256 color levels available.
This will be substantially reduced if everything were properly dithered, but in normal software and normal displays it is not.
How worth it is I don't know, but there is absolutely an easily detectable difference. How about testing your hypothesis before claiming you know what you're talking about, hmm? It's not exactly a difficult experiment to carry out.
Yup, the aircraft have an easier time of it. And even so, the autopilots can't eliminate pilot error as a source of crashes. You won't eliminate either driver or pilot error as a source of crashes in the forseeable future, even with plausible advancements in technology. You can *certainly* improve on the current state of affairs, though. I was just saying that OP was being excessively optimistic.
If you dithered that gradient properly it would be much harder to tell. You've essentially created a signal with a period of 8 pixels and overlaid it on top of the smooth gradient as a result of the quantization errors. It's that signal that's easy to see. If you dithered it properly, the noise would be shaped so that it didn't show up at one specific frequency band (ie 8 pixels and its harmonics), it would be much less noticeable. Note that eg digital photography processes do this inherently to some degree in the analog sensor.
Of course, very little software will do that dithering properly for you, so it's kind of a moot point. But then, not much will make proper use of 10 bit per channel color, either. I'd love to see one of these things up close...
Your eye only has 3 color sensors. Therefore you can reproduce any spectrum in a way that your eye will see as equivalent with only 3 color elements. That said, RGB doesn't do a perfect job of this -- there are some colors at the edges of the color space that your eye can see that RGB can't produce.
Now, it's entirely possible that the easiest way to produce the full spectrum when it comes to actually building a display is with more than 3 different color elements, but 3 is sufficient if they're the right 3 (RGB isn't quite).
It displays targetted ads on billboards *and* blocks spam? Aren't those mutually exclusive?
Commercial airplanes basically fly on autopilot. Accident rates are low. The most common cause of accidents is still pilot error.
I certainly agree, once the technology gets there, automated driving will vastly improve safety. But to assume that driver error will be *eliminated* as a cause of accidents, at least for the forseeable future, is wrong.
Sig seen on /.: "Instead of post anonymously, try the post humously chekbox."
I'd always wondered how to do that! Thanks, /.!
Do you have a theoretical basis for that statement? If you don't like the number, it wouldn't be hard to change -- the current version won't let you configure it over 20, but hacking the source wouldn't be hard at all. I recommend raising it slowly, though -- you may experience performance problems. I don't know; I haven't tried this.
Actually, Wine is an alternate implementation of the APIs, not an emulation of them. There's a difference, at least if you're using the words in a technical sense rather than a regular English sense. Which, when being pedantic about a technical matter, is the correct sense to use them in.
I'm assuming the Quaddro drivers don't require it for no reason, though -- ie they actually use some features in it. What are those?
Yup. Got that part. But what are the actual operations being performed, in more detail than "professional work"? Yeah, Maya, 3DS -- but how are those different from other programs that render 3D images?
What exactly is this enabling? I get that it's for "professional" applications, but what features do those use that aren't turned on normally?
Because anything capable of applying 1.4km/s delta-v to a 70kt asteroid is a larger propulsion system than has ever been built. Even if it's a mass driver, rather than a rocket engine, that still implies a very large engineering project -- larger than has been done in space before. Once we have the experience doing that sort of thing, it will be easier, but doing it on the second manned mission to any asteroid, ever, is likely to be a tall order.
You don't. But it's a convenient and legal way to test large file transfers. None of the devs and major testers actually need freenet for anything in particular that I'm aware of. Does that mean they shouldn't work on it or test it?
No, the energy to stop it is far less, since it's on an orbit very similar to ours. Most of the impact energy comes from it falling down Earth's gravity well. The number you want is the hyperbolic excess velocity, ie the speed it would have at impact if the Earth's gravity had no effect (roughly). See my other post for numbers.
Yes, you have to secure the machine freenet is running on. Unfortunately, if you actually want to run a node, the livecd doesn't work well, since it won't maintain a store long-term. If all you want to do is insert something and run, though, it'll work great.
I don't know much about MUTE, but probably because Freenet has a scalable routing protocol, possibly better security, and possibly a more active development community.
That's the thing, this one isn't hard to capture. It's in about as good an orbit for the purpose as you could ask for. There are probably some that would be easier, but what you really want for a first try is something that's smaller. Of course, I have no clue what it's made of -- you also want to check that it would be useful before capturing it ;)
I think doing it on the second manned mission to any asteroid would be optimistic...
It's got 1.37 km/s hyperbolic excess velocity, and on an orbit that damn near intersects ours. That means it takes a little more than 1370 m/s of delta-v to perform the capture. At 7.1E7 kg, that's about 6.6E13 joules -- approximately 15kt TNT equivalent worth of energy.
Assuming a high performance LOX/Methane engine, it would need about 34kt of propellant (rockets are inefficient for delta-v low relative to exhaust velocity). Note that this is a significant proportion of the asteroid mass. To make it economical, you'd need something more exotic -- a mass drive throwing bits of asteroid, or a high performance solar-electric ion drive, for example.
AnotherIndex is no longer in the default bookmark set. It wasn't updated all that often, it provided no site descriptions for the vast majority of sites, and it did a crappy job categorizing things (ie, putting things in completely absurd categories, not just failing to categorize them).
If you're interested in privacy, censorship resistant networks, and related topics, you're missing out on a fun place to hang out and discuss things. Fire up a node, load up FMS (the Freenet Messaging System; think usenet over Freenet -- there's a link in the default bookmark set), and subscribe to some boards that look interesting. The signal to noise ratio is quite good, though some of the posters are predictably extreme in any conceivable fashion. There are a couple trolls, but probably fewer than /. Start up a Freenet blog if you're so inclined, or go read other people's. Much like the regular internet blagosphere, most of it is completely boring, and some of it is really interesting.
Anyway, I encourage you to download it, fire up a node, and take a look around. You might be pleasantly surprised. Of course, you might decide that there's not much interesting and uninstall it, but if so, you haven't really lost anything more than a little bit of time. So come try it out!
P.S. FMS can be annoying to get set up, as it's still in an early alpha state and the docs suck. But join #freenet on irc.freenode.net and you'll find people who can answer questions about it, or the node itself if you're having trouble with that.
Yes, I agree, it's definitely good to hear. One of the medium-term goals for Freenet (0.8.0, hopefully) is steganographic transports that will hopefully make it much harder to see that you're running a (darknet) node.
The same laws that give ISPs protection from their users' downloading child porn should be applicable to individuals running freenet. This has not specifically been tested, however the legal framework for it appears fairly solid. Also, because of the fact that's it not stored as .jpgs on your disk (even encrypted) but rather scattered pieces of files of which you'll only have parts, in a form that can't be used to reconstruct the whole file, it would be rather difficult to convince a court that it was *you* downloading the files as opposed to your node propagating others' requests. And, it's not proof, though it should be reassuring -- no one has gotten in trouble for running freenet, even though the authorities are certainly paying attention (as far as we know, anyway).