Based on past behavior, the US gov't probably does the same, but doesn't tell anybody. They'll use some creative interpreting of a vague word in the law to justify it if somebody spills the beans, like they've done in the past with phone meta-data.
The only real difference is that the US has to be more clandestine about it among their population to avoid raising suspicion. The Chinese government WANTS their citizens to be suspicious and paranoid, it's how they keep them "in line".
How do you know? Most likely they are generic IT staff who set up a generic email server. Classified info was NOT supposed to be on that office server, so typically you wouldn't prepare them for that contingency.
It's possible they did prepare, but my experience with bureaucracies says, "probably not" because any problems could be blamed on the violator of policy rather than the IT manager. That's how managers in bureaucracies think: "who gets blame if X happens?" The IT manager will only prepare if they themselves get the blame. It's speculation either way, I'm just stating what my office politics observation experience tells me.
Again, a Turing Machine is a virtual machine. It does NOT escape dependence on the compiler and chip/brain/hardware design/implementation levels. Did you forget that point, or did you not understand it?
Adding yet another layer doesn't remove the dependence on the other layers. Something MUST translate Turing notation into machine-level instructions/commands, and some hardware MUST run it. (The equivalent of a compiler could be built into the hardware, I should point out.)
As far as titles, I'm talking about "computer science" as a degree or general field of study. Neither "biology" nor "physics" have the "science" suffix and therefore don't offer comparable examples. If anything they underscore my point by NOT having the "science" suffix. They are basically shortcuts for the study of biology and the study of the physical.
I don't understand your biology analogy. The topic of the course is indeed "biology": it's spot-on. If I go by your analogy, then the "computer science" degree should really be "computology". I'm okay with that. Sold!
But, it's a shortcut for "computer studies", which I don't think will fly for reasons I already stated.
Are you saying any imagined hardware would fit the mentioned speed profile, or is merely "testable"? Everything made is testable; that's not an key fact here. Nor does it mean most "computer science" students will be doing that testing.
It seems to me a ground-based system located on or near Hawaii would be cheaper and possibly more effective because it's easier to test & maintain, and can use volume to compensate for accuracy difficulties.
"known...hardware", yes. It's no more of a science than boat design. You don't call boat designers/engineers "boat scientists" even if there may be a tiny bit of "cutting edge" research involved.
Now if there were a "boat scientist" studying cutting edge theories on boat building materials, automated visual wave inspection and prediction to pre-counter waves, etc., then I'd feel comfortable calling that person a "boat scientist" or "naval scientist" or whatever. But that's only if the majority of his/her work is such cutting edge.
The vast majority of "computer science" students, probably around 99%, are NOT going to be doing something comparable to that.
And Big-O notation is a rule of thumb, not an inherent "law" of the universe, because it makes assumptions about compilers and chips that are not proven to be required.
As a thought experiment, most agree the human mind is a "computer". A human can study an algorithm and find shortcuts or even rewrites to speed it up (on known hardware). As long as the results are correct (or expected) to the user on the other side of the black box, they won't know or care if a human compiler/computer was involved in compiling and/or running the program.
Similarly, but more simply, existing compilers can recognize that the results of say an inner loop are not even used by the program under certain conditions, and skip running the inner loop. Thus, what superficially looks like a "double looping" application doesn't have to be executed as one on the chip.
she was storing classified information on a system not allowed to store or carry classified information. that alone is years in prison.
So if somebody sent YOU classified info HERE AND NOW to your existing email account, you automatically are guilty and go to prison for years? That's what you are implying, as I interpret your claim.
Also note that office server she "should have" been using was not designed for classified materials either. If somebody sent her classified stuff that ended up on that office server, it would be no more or less of a risk than to her home server. It was not a "special" server.
she blatantly ignored the law and should face the punishment.
Your Blatant-A-Meter appears to need recalibration, in light of high readings when applied to the flimsy evidence given.
We can clarify that "step" means a computer instruction.
That depends on the design of the chip. And the physical implementations of a Turing Machine can vary widely.
What's a "step" in the human mind? It's a computer.
because those are the only things that exist in a way that we can study.
Your proof would have to cover "all possible" compilers/chips/computing-hardware to be thorough or relevant here. I already agree with the sort claim for existing/known/typical/familiar technology. But, that's not the issue of contention.
You implied one algorithm is "inherently" faster than another. It's not (or at least not proven), because it also depends on the compiler and hardware, and we DON'T know all possible ways those 2 things can be built, and therefore you cannot make any general statement about their speed.
I didn't say parallelism is the only possible optimization; I mostly gave it an example to illustrate a point. And million-core chips may be something possible or viable in the future. Perhaps they'll find out that many simple CPU's perform certain problems faster than a few complex ones.
Are you saying for all possible chip designs and all possible compiler designs, bubble sort will ALWAYS be slower than merge sort (for non-trivial sort sets of relatively random keys)?
If the compiler doesn't care or ignores that "fact", then that's just an abstract classification. I'm asking for proof of speed, not proof of classification.
As mentioned elsewhere, a sophisticated optimizing compiler may find shortcuts. Even existing compilers can parallelize loops written a certain way. I'm not saying these will help bubble-short in particular, but saying what looks like a sequential loop in the high-level app code does NOT have to be executed as a sequential loop under the hood as long as the results are equivalent. We have existing specimens of this for certain kinds of algorithms/loops. Thus, as a general concept, it's not theoretical.
We can model things based on a simple/typical/common compiler, but that's making assumptions about the compiler.
A thorough proof would need to consider "all possible compilers", which is a very tall order.
Slashdot Mirror
Why would we want to do that? It's not related to this debate that I see.
Notation by itself has no "speed".
Based on past behavior, the US gov't probably does the same, but doesn't tell anybody. They'll use some creative interpreting of a vague word in the law to justify it if somebody spills the beans, like they've done in the past with phone meta-data.
The only real difference is that the US has to be more clandestine about it among their population to avoid raising suspicion. The Chinese government WANTS their citizens to be suspicious and paranoid, it's how they keep them "in line".
How do you know? Most likely they are generic IT staff who set up a generic email server. Classified info was NOT supposed to be on that office server, so typically you wouldn't prepare them for that contingency.
It's possible they did prepare, but my experience with bureaucracies says, "probably not" because any problems could be blamed on the violator of policy rather than the IT manager. That's how managers in bureaucracies think: "who gets blame if X happens?" The IT manager will only prepare if they themselves get the blame. It's speculation either way, I'm just stating what my office politics observation experience tells me.
Again, a Turing Machine is a virtual machine. It does NOT escape dependence on the compiler and chip/brain/hardware design/implementation levels. Did you forget that point, or did you not understand it?
Adding yet another layer doesn't remove the dependence on the other layers. Something MUST translate Turing notation into machine-level instructions/commands, and some hardware MUST run it. (The equivalent of a compiler could be built into the hardware, I should point out.)
As far as titles, I'm talking about "computer science" as a degree or general field of study. Neither "biology" nor "physics" have the "science" suffix and therefore don't offer comparable examples. If anything they underscore my point by NOT having the "science" suffix. They are basically shortcuts for the study of biology and the study of the physical.
You keep claiming that without solid evidence.
I don't understand your biology analogy. The topic of the course is indeed "biology": it's spot-on. If I go by your analogy, then the "computer science" degree should really be "computology". I'm okay with that. Sold!
But, it's a shortcut for "computer studies", which I don't think will fly for reasons I already stated.
I see you are a devout follower. Don't forget to pay your teething.
"Let's see what happens if I send in this odd Trump critter..."
And you will likewise be fired for a "transcendent" reason.
It's been obfuscated for security. He/she doesn't want humans mucking with it.
Are you saying any imagined hardware would fit the mentioned speed profile, or is merely "testable"? Everything made is testable; that's not an key fact here. Nor does it mean most "computer science" students will be doing that testing.
It seems to me a ground-based system located on or near Hawaii would be cheaper and possibly more effective because it's easier to test & maintain, and can use volume to compensate for accuracy difficulties.
The downside is unemployed undertakers.
Rick Perry is thinking, "Oh crap, that's yet another agency I have to remember to kill at debates."
"known...hardware", yes. It's no more of a science than boat design. You don't call boat designers/engineers "boat scientists" even if there may be a tiny bit of "cutting edge" research involved.
Now if there were a "boat scientist" studying cutting edge theories on boat building materials, automated visual wave inspection and prediction to pre-counter waves, etc., then I'd feel comfortable calling that person a "boat scientist" or "naval scientist" or whatever. But that's only if the majority of his/her work is such cutting edge.
The vast majority of "computer science" students, probably around 99%, are NOT going to be doing something comparable to that.
And Big-O notation is a rule of thumb, not an inherent "law" of the universe, because it makes assumptions about compilers and chips that are not proven to be required.
As a thought experiment, most agree the human mind is a "computer". A human can study an algorithm and find shortcuts or even rewrites to speed it up (on known hardware). As long as the results are correct (or expected) to the user on the other side of the black box, they won't know or care if a human compiler/computer was involved in compiling and/or running the program.
Similarly, but more simply, existing compilers can recognize that the results of say an inner loop are not even used by the program under certain conditions, and skip running the inner loop. Thus, what superficially looks like a "double looping" application doesn't have to be executed as one on the chip.
So if somebody sent YOU classified info HERE AND NOW to your existing email account, you automatically are guilty and go to prison for years? That's what you are implying, as I interpret your claim.
Also note that office server she "should have" been using was not designed for classified materials either. If somebody sent her classified stuff that ended up on that office server, it would be no more or less of a risk than to her home server. It was not a "special" server.
Your Blatant-A-Meter appears to need recalibration, in light of high readings when applied to the flimsy evidence given.
That depends on the design of the chip. And the physical implementations of a Turing Machine can vary widely.
What's a "step" in the human mind? It's a computer.
Your proof would have to cover "all possible" compilers/chips/computing-hardware to be thorough or relevant here. I already agree with the sort claim for existing/known/typical/familiar technology. But, that's not the issue of contention.
You implied one algorithm is "inherently" faster than another. It's not (or at least not proven), because it also depends on the compiler and hardware, and we DON'T know all possible ways those 2 things can be built, and therefore you cannot make any general statement about their speed.
Overall, they have less homicides and gun-related homicides than we do.
"Step" is not a clear word, and why rule out slime computers or things that borrow ideas from them?
Addendum: That's arguably closer to how the human brain works.
You might indeed someday, but Kimmy is just bluster most the time.
I didn't say parallelism is the only possible optimization; I mostly gave it an example to illustrate a point. And million-core chips may be something possible or viable in the future. Perhaps they'll find out that many simple CPU's perform certain problems faster than a few complex ones.
Are you saying for all possible chip designs and all possible compiler designs, bubble sort will ALWAYS be slower than merge sort (for non-trivial sort sets of relatively random keys)?
If so, can you provide the proof?
If the compiler doesn't care or ignores that "fact", then that's just an abstract classification. I'm asking for proof of speed, not proof of classification.
Are you saying corporations are NOT people? Blasphemy!
- Mitt
Jail em all and let God sort it out
As mentioned elsewhere, a sophisticated optimizing compiler may find shortcuts. Even existing compilers can parallelize loops written a certain way. I'm not saying these will help bubble-short in particular, but saying what looks like a sequential loop in the high-level app code does NOT have to be executed as a sequential loop under the hood as long as the results are equivalent. We have existing specimens of this for certain kinds of algorithms/loops. Thus, as a general concept, it's not theoretical.
We can model things based on a simple/typical/common compiler, but that's making assumptions about the compiler.
A thorough proof would need to consider "all possible compilers", which is a very tall order.