My impression is that it *does* act like a lens, based on references to similar optical techniques. Just an imperfect lens.
I'm sorry, I'm so used to people not reading the articles here that it didn't occur to me that you might just have a different take on it than I do.:-)
Do you mean the Java compiler or the JIT compiler? If you mean the JIT compiler, keep in mind that JIT compilation occurs at runtime, so this kind of analysis would be prohibitively expensive.
In either case, I think they analysis you're asking for is the kind that's easy for a human to do, but harder for an algorithm to do automatically. I don't know what's involved, and it could probably be done, but to prove that these transformations are safe is no mean feat.
Man, I'm so sad I read Slashdot for months before I bothered to get an ID. By the time I finally got one, they had already moved from punched cards to reel-to-reel magnetic tapes, so I don't have any good stories.
Technically, I don't think your content-addressable memory gives O(1) because "O" itself is defined as asymptotic behaviour; that is, the resources required by an algorithm as the size of the input goes to infinity.
No matter how big your content-addressable memory is, it is finite, and so there is some input size for which you can't use your contant-time algorithm. When you try to scale your search algorithm to suit these cases, it will become O(log n).
Having said that, my very limited understanding is that this is exactly the case with quantum computing, as you say. Namely, they give some great speedups for problem sizes that fit into a given number of qbits, but they don't give any benefit in asymptotic behaviour.
Yes, you're missing the point. To put it in your terms, these guys are inventing the language that could be used to write your C compiler, or perhaps the code your C compiler generates.
Well, I disagree, but gine. You can define "information" and "meaningful" and "patterns" however you want, but your definition has no bearing on data compression, which is what we were discussing.
No, you still have it wrong. Information is entropy. More information is more entropy. However, imagine the amount of information in a JPEG of your face, compared with a JPEG of bits from/dev/random. The latter will have more information and thus more entropy. That shouldn't give you an inferiority complex.:-)
You remind me of a quote I think came from Gustav Eiffel. He said (in French) that anyone can make a bridge that stands; but it takes an engineer to make a bridge that just barely stands.
If you get a degree in architecture, does that make you an architect? No. If you get a law degree, does that make you a lawyer? No.
Likewise, just because you have an engineering degree doesn't automatically make you an engineer. I sympathize, because I have an MASc in Computer Engineering, but I am not a PEng, so I am very much undecided as to whether I deserve the title.
Real engineers work in a manner and with accountability that the software field lacks. Engineering means having reduced a craft to practice. That has not yet happened in software. Give it another 100 years or so. In the mean time, there are very, very few software engineers out there.
Maybe I have the wrong idea of what Gee-Whiz is. For instance, Clarke has a story about a guy who gets stuck in a bad orbit around the moon that will collide with the surface in a few hours. The story goes into his reaction to things, and his coming to terms with his impending doom, but also features some cool plot turns based on orbital mechanics. I think the latter is kind of a scientific Gee-Whiz.
Slashdot Mirror
In either case, I think they analysis you're asking for is the kind that's easy for a human to do, but harder for an algorithm to do automatically. I don't know what's involved, and it could probably be done, but to prove that these transformations are safe is no mean feat.
Moderators, *please* stop modding up questions that are answered in the article.
Moderators, please stop modding up questions that are answered in the article.
Read the press release. The gas doesn't literally act like a lens; that's just a metaphor for the magnification effect it has.
I guess you've never worked with an SMP?
Nice.
Man, I'm so sad I read Slashdot for months before I bothered to get an ID. By the time I finally got one, they had already moved from punched cards to reel-to-reel magnetic tapes, so I don't have any good stories.
Ha, you 80Kers think you're old-timers.
No, it's torque. Go do a Google search and learn about the experiment.
No matter how big your content-addressable memory is, it is finite, and so there is some input size for which you can't use your contant-time algorithm. When you try to scale your search algorithm to suit these cases, it will become O(log n).
Having said that, my very limited understanding is that this is exactly the case with quantum computing, as you say. Namely, they give some great speedups for problem sizes that fit into a given number of qbits, but they don't give any benefit in asymptotic behaviour.
Yes, you're missing the point. To put it in your terms, these guys are inventing the language that could be used to write your C compiler, or perhaps the code your C compiler generates.
Um, some scientists work on theories. That's what they do. Theories are tools to help scientists find answers.
I think you're wrong. Did you try it? The line starts with "foo(){" which defines a new function called "foo".
Well, I disagree, but gine. You can define "information" and "meaningful" and "patterns" however you want, but your definition has no bearing on data compression, which is what we were discussing.
Yeah, you're right. In hindsight, my post was pretty confusing as it was. :-)
To the curious: to interpret parent's signature, imagine replacing the colon with the word "foo". It's much more legible that way.
No, you still have it wrong. Information is entropy. More information is more entropy. However, imagine the amount of information in a JPEG of your face, compared with a JPEG of bits from /dev/random. The latter will have more information and thus more entropy. That shouldn't give you an inferiority complex. :-)
You remind me of a quote I think came from Gustav Eiffel. He said (in French) that anyone can make a bridge that stands; but it takes an engineer to make a bridge that just barely stands.
Question: would you take legal liability for damage caused by faults in software you write?
Likewise, just because you have an engineering degree doesn't automatically make you an engineer. I sympathize, because I have an MASc in Computer Engineering, but I am not a PEng, so I am very much undecided as to whether I deserve the title.
Real engineers work in a manner and with accountability that the software field lacks. Engineering means having reduced a craft to practice. That has not yet happened in software. Give it another 100 years or so. In the mean time, there are very, very few software engineers out there.
Exactly. I've said the same thing at least once.
After lots of heated debate, let the debates begin.
It was just a little poetic license on the part of the submitter.
Maybe I have the wrong idea of what Gee-Whiz is. For instance, Clarke has a story about a guy who gets stuck in a bad orbit around the moon that will collide with the surface in a few hours. The story goes into his reaction to things, and his coming to terms with his impending doom, but also features some cool plot turns based on orbital mechanics. I think the latter is kind of a scientific Gee-Whiz.