Actually, the post title and summary are fairly accurate, given that Popular Mechanics itself refers to the article as "Science & Tech Guide to the Election" and "Election 2008 Science and Technology Policies".
This sort of article passes as interesting? It's a second-rate hash of the philosophy of science that doesn't even "faith in science" trap that everyone falls into.
It's unfortunate that so many people (like the readers of Slashdot, even) are so misinformed about what scientific laws are and what they mean.
Yeah, you're right re: the speed of Grover's algorithm.
Clearly in the above I was using the definition of powerful that makes sense, not the other typical definition ("capable of solving a larger body of problems").
That's correct, you still need other gates to perform interesting functions. Yet the Hadamard is still the most essential and fundamental quantum operator, and one that has no classical equivalent.
Quantum computing is very different. The details are of course very different (such as the operators, and the need for bit-level error checking), but more important to the software developer, the algorithms are fundamentally different. You approach a problem with an entirely foreign set of tools. With quantum computers, it's not a matter of the quantum computer just being "better" -- it has access to a way of doing things that is more powerful (in the mathematical sense) than classical computing.
I don't remember offhand the set of problems that are trivialized by quantum computing, but the difficulty of many problems changes drastically. For example, you can find an element in an unsorted array using a quantum computer in constant time.
Of course, he didn't say they plugged them into a wide-open connection. Perhaps people inferred that from the statement that they don't use firewalls, but that's their mistake.
The issue is not your tired, bandwagon "America is unimportant" attitude, but your choice of words. If part of the world wants to separate its networks from the Internet, that's fine. (Realistically, no network in their right mind would do so, particularly since our laws would apply to that network as much as they do now.) However, as essentially all of the major Internet backbone structure is within the U.S., of the two separated networks, ours would be the one properly labelled "the Internet".
Of course, since 7/9 of the Tier 1 networks are American companies (one is in Bermuda and the other is an American company wholly owned by a Japanese company), I'm not sure how well that would really work out.
Yes. Do you have the mathematical background or experience with secure systems, six months from your BS, to make these claims?
I didn't say cryptography was magic dust, nor that you could have an electronic voting system without having the internal workings of the system fully public. But to hand-wave at problems with the "anything could happen" defense like you do is amateurish science and engineering at best.
Besides, my main issue is trying to suggest silly things like, apparently, the absence of write-once, tamper-proof protocols and storage.
"With a flash memory card, who knows? A person can't say *anything* about what's stored on it without putting it in a reader, and any reader device can trivially and tracelessly change the data in milliseconds."
First, there are cryptographic means of preventing data from being tampered with tracelessly. Second, there are communication and storage systems (like hardware storage) entirely capable of guaranteed write-once transactions. Just because a problem seems difficult and complicated to you doesn't justify trying to claim that there's no solution.
Unfortunately, the efficiency of heating and cooling is seriously limited by thermodynamics. Cars have been developed that are a ton more efficient, but only some people buy them. Many others buy cars that are less efficient than the ones they had before.
The best thing you can do for heating/cooling is to have a well-insulated house with good air movement and to take advantage of passive heating/cooling.
Actually, in most restaurants in my area, talking on a cellphone will summon someone who asks you politely to take it outside. After that, much less politely. Most lines (banks, take-out food, DMV) do not permit cell phone use and they'll ask you to leave if you ignore that request.
It seems like you're expecting a university science program to train people in a particular, specific skill you want.
A compiler can certainly be written with no knowledge of how an executable header works. Clearly the direct output won't be machine code. More importantly, the specifics of how you need to format some bytes in order to create executable machine code with your compiler is not generic, interesting, or difficult. There's no reason to teach it in a university program.
Knowing assembler seems more reminiscent of a tech school. As a branch of mathematics, a CS academic isn't likely or expected to know all that much about assembler or PE/ELF headers. What you're looking for is a programmer or computer engineer, which they often teach people how to be alongside a proper CS curriculum.
What has changed is that the body of knowledge for a proper understanding of CS and the interest in what a CS program should focus on has shifted. While you may have learned electrical engineering and assembler out of necessity or interest in the 60s, these are no longer CS topics.
Conversely, in my experience, the level to which students are already taught when they reach undergraduate or graduate school has been increasing. Undergrad science and engineering students regularly skip what used to be well over a year's worth of work from having already learned it in high school and enter grad school with substantial graduate-level knowledge.
That's why I neglected to mention them.:-) Soft links are almost a form of label emulation. Hard links are powerful, but just expose that the underlying system is designed as a form of structured labeling meant to look and behave like a folder system.
Copying works fine, but it isn't the same. You're just duplicating an item, instead of having one item with two labels. You can *emulate* labels with folders, yes. But the things you can do with labeling is a superset of the things you can do with folders.
You certainly could make the same complaint about your PC's filesystem, or about how you organize your digital music files. In both cases, newer or in-development systems prefer the label system to the folder system because of its flexibility. In all physical analogies, you're constrained by the fact that no object can be in two places at once, but sorting is often most useful when done physically instead of maintaining lists of where to find items. The file-cabinet design is what originally directed the folder system in the first place, and it's now that people are trying to shed its restrictions.
If searching for items in a label system is more difficult than with a folder system, either you're not labeling properly, or the label system's UI is poorly-designed.
How you assign labels isn't a property of a labelling system, it's a matter of how it's exposed in the UI and can be changed at a whim. Any UI metaphor you use for putting items in folders can be extended to applying labels to items, if you replace the folder with a label. (You'd say, then, that instead of "move item from current location to folder", drag-and-drop would be "apply target label to item and remove the label that is currently exposing this item".)
As far as scaling goes, if you're talking "scaling" in terms of UI presentation, see above. Anything that can be done with folders can be done in the same way with labels. If you're talking scaling in implementation, you're wrong. Putting 10,000 items in one storage location and filtering them all is inefficient if you implement it poorly. But I could follow the folders-are-just-labels thinking and, for each label, maintain a list of items that label has been applied to. Now I have precisely the efficiency folders had.
Well, you're correct. Any sufficiently well-made label system (not too tough, and Gmail seems to have done it) can do everything a folder system can do. The converse is not true -- there's no really elegant way of being able to place an item in multiple folders simultaneously without redefining the folder as a label.
Slashdot Mirror
Actually, the post title and summary are fairly accurate, given that Popular Mechanics itself refers to the article as "Science & Tech Guide to the Election" and "Election 2008 Science and Technology Policies".
This sort of article passes as interesting? It's a second-rate hash of the philosophy of science that doesn't even "faith in science" trap that everyone falls into.
It's unfortunate that so many people (like the readers of Slashdot, even) are so misinformed about what scientific laws are and what they mean.
Not enough 9's imo.
Yeah, you're right re: the speed of Grover's algorithm.
Clearly in the above I was using the definition of powerful that makes sense, not the other typical definition ("capable of solving a larger body of problems").
That's correct, you still need other gates to perform interesting functions. Yet the Hadamard is still the most essential and fundamental quantum operator, and one that has no classical equivalent.
Neither the superposed q-bit nor the Hadamard operator that is the fundamental quantum operator have classical computational equivalents.
Quantum computing is very different. The details are of course very different (such as the operators, and the need for bit-level error checking), but more important to the software developer, the algorithms are fundamentally different. You approach a problem with an entirely foreign set of tools. With quantum computers, it's not a matter of the quantum computer just being "better" -- it has access to a way of doing things that is more powerful (in the mathematical sense) than classical computing.
I don't remember offhand the set of problems that are trivialized by quantum computing, but the difficulty of many problems changes drastically. For example, you can find an element in an unsorted array using a quantum computer in constant time.
Of course, he didn't say they plugged them into a wide-open connection. Perhaps people inferred that from the statement that they don't use firewalls, but that's their mistake.
Wasn't that Ohio?
You are familiar with basic government structure in the US, right?
The issue is not your tired, bandwagon "America is unimportant" attitude, but your choice of words. If part of the world wants to separate its networks from the Internet, that's fine. (Realistically, no network in their right mind would do so, particularly since our laws would apply to that network as much as they do now.) However, as essentially all of the major Internet backbone structure is within the U.S., of the two separated networks, ours would be the one properly labelled "the Internet".
Of course, since 7/9 of the Tier 1 networks are American companies (one is in Bermuda and the other is an American company wholly owned by a Japanese company), I'm not sure how well that would really work out.
Nope, it's definitely "scrith": http://en.wikipedia.org/wiki/Scrith#Scrith
1970 comes before 1994. Imitators need not apply.
Yes. Do you have the mathematical background or experience with secure systems, six months from your BS, to make these claims?
I didn't say cryptography was magic dust, nor that you could have an electronic voting system without having the internal workings of the system fully public. But to hand-wave at problems with the "anything could happen" defense like you do is amateurish science and engineering at best.
Besides, my main issue is trying to suggest silly things like, apparently, the absence of write-once, tamper-proof protocols and storage.
"With a flash memory card, who knows? A person can't say *anything* about what's stored on it without putting it in a reader, and any reader device can trivially and tracelessly change the data in milliseconds."
First, there are cryptographic means of preventing data from being tampered with tracelessly. Second, there are communication and storage systems (like hardware storage) entirely capable of guaranteed write-once transactions. Just because a problem seems difficult and complicated to you doesn't justify trying to claim that there's no solution.
It seems you're not familiar with the percentages involved here.
The approving-by-line-item analogy is dangerous. How would funding for Iraq be if you could choose not to pay for it?
By comparison, the enormously small amount spent on NASA parties would be irrelevant to the average taxpayer.
If you think that kind of strategy would be easy, you haven't worked with engineers and scientists enough.
Unfortunately, the efficiency of heating and cooling is seriously limited by thermodynamics. Cars have been developed that are a ton more efficient, but only some people buy them. Many others buy cars that are less efficient than the ones they had before.
The best thing you can do for heating/cooling is to have a well-insulated house with good air movement and to take advantage of passive heating/cooling.
Actually, in most restaurants in my area, talking on a cellphone will summon someone who asks you politely to take it outside. After that, much less politely. Most lines (banks, take-out food, DMV) do not permit cell phone use and they'll ask you to leave if you ignore that request.
It seems like you're expecting a university science program to train people in a particular, specific skill you want.
A compiler can certainly be written with no knowledge of how an executable header works. Clearly the direct output won't be machine code. More importantly, the specifics of how you need to format some bytes in order to create executable machine code with your compiler is not generic, interesting, or difficult. There's no reason to teach it in a university program.
Knowing assembler seems more reminiscent of a tech school. As a branch of mathematics, a CS academic isn't likely or expected to know all that much about assembler or PE/ELF headers. What you're looking for is a programmer or computer engineer, which they often teach people how to be alongside a proper CS curriculum.
What has changed is that the body of knowledge for a proper understanding of CS and the interest in what a CS program should focus on has shifted. While you may have learned electrical engineering and assembler out of necessity or interest in the 60s, these are no longer CS topics.
Conversely, in my experience, the level to which students are already taught when they reach undergraduate or graduate school has been increasing. Undergrad science and engineering students regularly skip what used to be well over a year's worth of work from having already learned it in high school and enter grad school with substantial graduate-level knowledge.
That's why I neglected to mention them. :-) Soft links are almost a form of label emulation. Hard links are powerful, but just expose that the underlying system is designed as a form of structured labeling meant to look and behave like a folder system.
Copying works fine, but it isn't the same. You're just duplicating an item, instead of having one item with two labels. You can *emulate* labels with folders, yes. But the things you can do with labeling is a superset of the things you can do with folders.
You certainly could make the same complaint about your PC's filesystem, or about how you organize your digital music files. In both cases, newer or in-development systems prefer the label system to the folder system because of its flexibility. In all physical analogies, you're constrained by the fact that no object can be in two places at once, but sorting is often most useful when done physically instead of maintaining lists of where to find items. The file-cabinet design is what originally directed the folder system in the first place, and it's now that people are trying to shed its restrictions.
If searching for items in a label system is more difficult than with a folder system, either you're not labeling properly, or the label system's UI is poorly-designed.
How you assign labels isn't a property of a labelling system, it's a matter of how it's exposed in the UI and can be changed at a whim. Any UI metaphor you use for putting items in folders can be extended to applying labels to items, if you replace the folder with a label. (You'd say, then, that instead of "move item from current location to folder", drag-and-drop would be "apply target label to item and remove the label that is currently exposing this item".)
As far as scaling goes, if you're talking "scaling" in terms of UI presentation, see above. Anything that can be done with folders can be done in the same way with labels. If you're talking scaling in implementation, you're wrong. Putting 10,000 items in one storage location and filtering them all is inefficient if you implement it poorly. But I could follow the folders-are-just-labels thinking and, for each label, maintain a list of items that label has been applied to. Now I have precisely the efficiency folders had.
Well, you're correct. Any sufficiently well-made label system (not too tough, and Gmail seems to have done it) can do everything a folder system can do. The converse is not true -- there's no really elegant way of being able to place an item in multiple folders simultaneously without redefining the folder as a label.