In what way would my life be different if my brain were (or weren't) in a vat? If it's untestable then there isn't any useful difference. Any distinction at all some might say.
It depends. I think it comes down to redistribution rights. You can license some rights to someone, but not redistribution. Some shareware likely fits into that category. So people who got licenses directly from the author would have permanent rights for his own use, but not to redistribute. The GPL explicitly grants redistribution rights.
Blurring from saccades is uniform in angular degrees. Blurring from lens effects isn't. The difference is measurable. In a picture it'd translate to less blurring near the center of the painting as opposed to the closest to the painter.
No, they're referring to the way we perceive exponential differences in light linearly. An example of this is the way digital cameras adjust for this behavior when concerting from raw.
You may be right. But one objection to the camera obscura theory was that such a technique had never been demonstrated. Now it has. It's far from proof, but it sure is interesting.
Actually, getting tone right requires an awful lot... people need theory and practice. The fact is, we perceive color relative to surrounding color. The light in the local area will affect the actual color value, but affect our perception of that color much less. Getting actual tone rather than a good relative tone is probably impossible for the unaided artist.
What you're saying is true. What the article is referring to is actually the nature of perception. When we see light, we perceive differences in brightness linearly when the actual energy difference is exponential. This is something that cameras take into account when converting from raw. With paintings and photographs you're taking a huge range of brightnesses and compressing it into a very small range that depends on the lighting at the time of viewing. I don't doubt that an optical aid would give you a much more accurate brightness curve.
The unaided eye does actually have limitations. No amount of skill or intelligence can overcome that.
Also, one objection had been that no such method had ever been demonstrated. Now it has.
It's not proof, but it does diminish the opposing arguments.
Whole classes of simulation demand exact reproducibility, and at the same time "local" results can vary wildly with the smallest difference in input. And they need the speed that modern floating point hw provides. Global results shouldn't be so dependent, but that's not the point. And yes, scientists want to understand such these details. But that's also besides the point.
Many people are ok with the collection of metadata, but wouldn't be if they knew the kinds of details that may be learned from metadata. E.g. political leanings, sexual orientation, degrees of separation from, well anyone. I'm actually curious how much they can figure out.
That said, I don't believe it's just metadata. The size of datacenter the NSA just built tells me they're collecting whatever flows thru their capture devices. Wholesale.
Slashdot Mirror
In what way would my life be different if my brain were (or weren't) in a vat? If it's untestable then there isn't any useful difference. Any distinction at all some might say.
Cite some evidence.
It depends. I think it comes down to redistribution rights. You can license some rights to someone, but not redistribution. Some shareware likely fits into that category. So people who got licenses directly from the author would have permanent rights for his own use, but not to redistribute. The GPL explicitly grants redistribution rights.
I agree. You should stay away from that particular scam - the dollar. Let me have any that come your way: I'm a sucker.
What's wrong with sodomizing your neighbor's wife, so long as she consents? Are you worried about her dog? I doubt he'll be too jealous.
Lose != loose
Grammar-nazi! Wait... oops!
Racial profiling is morally wrong, ineffective and easy for a criminal to defeat. It's simply not justifiable.
I wonder why he was trying to avoid being caught...
You must be new here
Wish I had some mod points. Parent is exactly right.
So fucking what? They're different issues.
Blurring from saccades is uniform in angular degrees. Blurring from lens effects isn't. The difference is measurable. In a picture it'd translate to less blurring near the center of the painting as opposed to the closest to the painter.
No, they're referring to the way we perceive exponential differences in light linearly. An example of this is the way digital cameras adjust for this behavior when concerting from raw.
You may be right. But one objection to the camera obscura theory was that such a technique had never been demonstrated. Now it has. It's far from proof, but it sure is interesting.
No. People have pointed out details that are practically impossible for a master to acheive, let alone a 20 something with no training.
Actually, getting tone right requires an awful lot... people need theory and practice. The fact is, we perceive color relative to surrounding color. The light in the local area will affect the actual color value, but affect our perception of that color much less. Getting actual tone rather than a good relative tone is probably impossible for the unaided artist.
What you're saying is true. What the article is referring to is actually the nature of perception. When we see light, we perceive differences in brightness linearly when the actual energy difference is exponential. This is something that cameras take into account when converting from raw. With paintings and photographs you're taking a huge range of brightnesses and compressing it into a very small range that depends on the lighting at the time of viewing. I don't doubt that an optical aid would give you a much more accurate brightness curve.
The unaided eye does actually have limitations. No amount of skill or intelligence can overcome that. Also, one objection had been that no such method had ever been demonstrated. Now it has. It's not proof, but it does diminish the opposing arguments.
You must be new here
Whole classes of simulation demand exact reproducibility, and at the same time "local" results can vary wildly with the smallest difference in input. And they need the speed that modern floating point hw provides. Global results shouldn't be so dependent, but that's not the point. And yes, scientists want to understand such these details. But that's also besides the point.
It could be a huge problem. See butterfly effect. Even 1ulp difference can compound in large classes of simulation.
IEEE math is deterministic, but implementations may not be. There have been many hardware bugs. In a simulation reproducibility may be critical.
That said, I don't believe it's just metadata. The size of datacenter the NSA just built tells me they're collecting whatever flows thru their capture devices. Wholesale.
I wholeheartedly support this project and already signed up. Thank you