Taking another user's example, look what happens if you don't send the noncompliant information: A bad implementation won't keep the connection alive, which means you'll be stuck reconnecting repeatedly, with all the performance degradation that entails.
As another comment pointed out, Linux will have DRM eventually. I find it ironic, however, that DRM may help keep Linux out of political hot water.
Current DRM schemes like Palladium are designed in such a way that if you don't use the system, you can't use the content. This means you have a choice.
Without that choice, we'd be stuck with some method of certifying each kernel and media app build, which would grind OSS development to a halt.
Cache is only useful when you're working on a small data set. Multimedia applications tend to constantly move through a large data set, which makes the cache all but useless. Full details here.
You have to look at it from the perspective of an eleven-year-old, which is how old I was when he told me.:) I've since learned enough to understand why it's unlikely, but what the hell..It's still got a funny theme.
You don't tell someone they're full of shit when they tell "The Three Billy Goats 'Gruff'" do you? It's kindof a tech-support fairy tale.
Who can you trust if you can't trust the government?
Seriously, though, I think you're right. If the transfer process isn't completely automated, then the occasionaly black-hat officer is going to erase his mess-ups.
Even if it is automated, the officer can do things like stand with his back to the camera (if circumstances allow) or obstruct the microphone. That's Darwin's evolution for you.
If, however, the hard drive is destroyed under mysterious circumstances, then the arrested individual has an excellent oppertunity to place doubt on the reputation of the officer who arrested him.
A friend of mine saw a cop car on an offramp from the freeway, with his hazard lights on. He pulled over behind him and asked if he could help. The cop had pulled over because his in-car computer wasn't working.
My friend got it working, and bid the cop farewell. The cop wrote him a ticket for parking on an offramp.
My friend, being a tolerant guy, explained to the cop that there was one last thing he had to do to keep the laptop from breaking again...so he set it to self-analyze for 120 million hours--roughly four years. That model couldn't be easily rebooted...
Why not? That's a lot of video. It'll eventually have to be lossless, thanks to the lawsuit-happy nature of the legal system, which means it'll be harder to compress. There'd also be too much labor involved in cutting out all the irrelevant video data.
One of the first courses in all college business curriculums I've seen is "Business Statistics" (BA154 here at GRCC.).
The course focuses on making decisions based on statistics. In the second week of class, we learned what a standard deviation was, and we never stopped using it throughout the semester.
But perhaps ignorance would explain business tactics of the 90's.
Hence the original assertion that you can't have AI because it's just 1s and 0s is false.
Maybe I should have been clearer earlier...I don't disagree with you on this point. I merely disagree on your equating, in a sense of technique, digital electronics with the brain's synapses.
There are a couple of stages you're missing. First, you have to deal with the amount of neurotransmitter that is received by a synapse. There is an amount of leakage during the transmission process that varies based on a few environmental characteristics, like how far the firing synapse is from the receiving synapse, and how many transmitters and receptors there are.
All of these things can be affected by natural(genetic) or artifical(medicine) factors.
There's a great deal of low-level analog behavior in the brain.
Those are just state machiens.
on
AI Going Nowhere?
·
· Score: 2, Interesting
All of the spiking networks I've seen were nothing more than state machines that depend on numeric comparisons.
But I'm not an expert, and that's just my personal opinion.
I agree that AI is somewhat neglected. However, anything more automated than state machines and fuzzy logic is notoriously difficult to write properly.
If you have the (human) time, use a state machine. You can put all the intelligence and complexity of human behaviour you want into it.
No, just that there will be a disproportionately larger number of tech people to boring people showing up here on Slashdot.
Put another way, people on Slashdot are more likely to be tech people, who are more likely to be unemployed than people of other fields.
Granted, that doesn't justify his use of the phrase "most of us," but it does give you an idea of part of the formula that could be used to reach that conclusion.
AMD released a marketing schedule back just before they changed their labeling system. AFAIK, they haven't had any slips in their release dates. They may just be sacrificing performance for market presence, which is a Really Good Idea.
If you lose market presence, you lose the game. Permanently. The only way back in is to come up with something so revolutionary, you'd attract VC in this economy.
I think you're mistaking peak throughput with average throughput. Let me know the first time you read enough physically consecutive data to reach the peak throughput of IDE or any SCSI variant.
Solid-state mass storage would all but eliminate latency, but it's just not cheap enough for consumer use yet.
Well, that may not be practical, but consider this: Game interactivity could gain massive boost of any sort of fluid model was used; Watch all the windows in a building get blown out when some idiot throws a grenade. Another example would be mass-spring systems for solids. Ever want to take out a load-bearing wall in a building map? Or cause a cave-in?
Don't forget that there are a lot of potential future kernel hackers still in diapers right now, and a greater Linux marketshare will result in them being exposed to Linux sooner.
Slashdot Mirror
The blink tag works great for papers on quantum physics.
(Credit to UserFriendly goes here)
Taking another user's example, look what happens if you don't send the noncompliant information: A bad implementation won't keep the connection alive, which means you'll be stuck reconnecting repeatedly, with all the performance degradation that entails.
As another comment pointed out, Linux will have DRM eventually. I find it ironic, however, that DRM may help keep Linux out of political hot water.
Current DRM schemes like Palladium are designed in such a way that if you don't use the system, you can't use the content. This means you have a choice.
Without that choice, we'd be stuck with some method of certifying each kernel and media app build, which would grind OSS development to a halt.
Cache is only useful when you're working on a small data set. Multimedia applications tend to constantly move through a large data set, which makes the cache all but useless. Full details here.
Licensing the code also allows Microsoft to do with it what others can't: Make changes and keep them secret.
It let's them take a piece of Linux and incorporate it into their own products, without releasing their source code.
I can't imagine it as being a very large piece of Linux, though, given that the kernel has so many contributors.
You have to look at it from the perspective of an eleven-year-old, which is how old I was when he told me. :) I've since learned enough to understand why it's unlikely, but what the hell..It's still got a funny theme.
You don't tell someone they're full of shit when they tell "The Three Billy Goats 'Gruff'" do you? It's kindof a tech-support fairy tale.
Who can you trust if you can't trust the government?
Seriously, though, I think you're right. If the transfer process isn't completely automated, then the occasionaly black-hat officer is going to erase his mess-ups.
Even if it is automated, the officer can do things like stand with his back to the camera (if circumstances allow) or obstruct the microphone. That's Darwin's evolution for you.
If, however, the hard drive is destroyed under mysterious circumstances, then the arrested individual has an excellent oppertunity to place doubt on the reputation of the officer who arrested him.
I've got a story:
A friend of mine saw a cop car on an offramp from the freeway, with his hazard lights on. He pulled over behind him and asked if he could help. The cop had pulled over because his in-car computer wasn't working.
My friend got it working, and bid the cop farewell. The cop wrote him a ticket for parking on an offramp.
My friend, being a tolerant guy, explained to the cop that there was one last thing he had to do to keep the laptop from breaking again...so he set it to self-analyze for 120 million hours--roughly four years. That model couldn't be easily rebooted...
It's going to lead to interesting legal battles. Sure you can circumstantial video to bolster your case, but can you use it as your primary weapon?
I only say circumstantial because the recording was made while the officer didn't consider it important/wasn't paying attention.
Why not? That's a lot of video. It'll eventually have to be lossless, thanks to the lawsuit-happy nature of the legal system, which means it'll be harder to compress. There'd also be too much labor involved in cutting out all the irrelevant video data.
One of the first courses in all college business curriculums I've seen is "Business Statistics" (BA154 here at GRCC.).
The course focuses on making decisions based on statistics. In the second week of class, we learned what a standard deviation was, and we never stopped using it throughout the semester.
But perhaps ignorance would explain business tactics of the 90's.
Hence the original assertion that you can't have AI because it's just 1s and 0s is false.
Maybe I should have been clearer earlier...I don't disagree with you on this point. I merely disagree on your equating, in a sense of technique, digital electronics with the brain's synapses.
Let me reword what I said, then.
Sure, transmitters either fire, or they don't. The amount of signal communicated by that firing isn't gauranteed.
There are a couple of stages you're missing. First, you have to deal with the amount of neurotransmitter that is received by a synapse. There is an amount of leakage during the transmission process that varies based on a few environmental characteristics, like how far the firing synapse is from the receiving synapse, and how many transmitters and receptors there are.
All of these things can be affected by natural(genetic) or artifical(medicine) factors.
There's a great deal of low-level analog behavior in the brain.
All of the spiking networks I've seen were nothing more than state machines that depend on numeric comparisons.
But I'm not an expert, and that's just my personal opinion.
I agree that AI is somewhat neglected. However, anything more automated than state machines and fuzzy logic is notoriously difficult to write properly.
If you have the (human) time, use a state machine. You can put all the intelligence and complexity of human behaviour you want into it.
No, just that there will be a disproportionately larger number of tech people to boring people showing up here on Slashdot.
Put another way, people on Slashdot are more likely to be tech people, who are more likely to be unemployed than people of other fields.
Granted, that doesn't justify his use of the phrase "most of us," but it does give you an idea of part of the formula that could be used to reach that conclusion.
AMD released a marketing schedule back just before they changed their labeling system. AFAIK, they haven't had any slips in their release dates. They may just be sacrificing performance for market presence, which is a Really Good Idea.
If you lose market presence, you lose the game. Permanently. The only way back in is to come up with something so revolutionary, you'd attract VC in this economy.
Software-mode Mesa, baby! :P
It's the same reasoning that helps me choose OSS software over non-OSS. Are you going to call that silly, here on Slashdot?
I think you're mistaking peak throughput with average throughput. Let me know the first time you read enough physically consecutive data to reach the peak throughput of IDE or any SCSI variant.
Solid-state mass storage would all but eliminate latency, but it's just not cheap enough for consumer use yet.
Considering the tech sector was the hardest hit? And tech people without jobs are more likely to spend time on Slashdot?
One game: Software rendered Doom III. :)
Well, that may not be practical, but consider this: Game interactivity could gain massive boost of any sort of fluid model was used; Watch all the windows in a building get blown out when some idiot throws a grenade. Another example would be mass-spring systems for solids. Ever want to take out a load-bearing wall in a building map? Or cause a cave-in?
Don't forget that there are a lot of potential future kernel hackers still in diapers right now, and a greater Linux marketshare will result in them being exposed to Linux sooner.
Was I the only one who immediately suspected he forged documents claiming prior art? People in that business aren't known for their creativity...