The distinction I made is valid: the encoders and decoders you use are free only by virtue of the fact that you downloaded them for free. (Trane Francks)
Exactly. It's free for ME (Reality Master 101)
MP3 is not a free standard, regardless of the existence of free downloads. Trane Francks even went so far as to quote the license itself. The "tons of free, licensed encoders" you mention do not exist. They're all unlicensed, unless you've found one whose author is fronting the $5/unit+$15k/yearly royalty to Thomson. Non-commercial use is irrelevant for encoders (and now they're claiming that for decoders, too).
I will happily continue to use LAME, but I'm under no delusions about its legality, and neither are its developers.
This isn't really so off-topic, given Olin's nonexistent tuition.
The Ivies and other such private colleges give aid to well over half of their students. (It's actually much more than half but I don't have the figures with me.) Consider, a typical upper-middle-income family makes $50k per parent. That's $100k per year, $75k after taxes. It's unreasonable to expect $35k of that to go for tuition. All things being equal, this family might receive around $10k of aid or more, not including loans.
Put bluntly, an income of $100k is smack in the middle of the distribution curve at most well-endowed schools.
Yet, I've talked with dozens of families with incomes in that range and the overwhelming majority of them say, "Oh, we make $100k per year, we know that means we're ineligible for aid." And so they don't even bother to apply. It's a deadly myth that's doomed legions of promising students to choose a college based on cost.
Incidentally, the Ivies' elite status is almost entirely a result of their selectivity, and has little to do with their cost or their overrated quality (yes, I went to one).
...for $10K-$30K/year they should be allowed to stockpile as much knowledge as possible in their chosen fields.
This idea is sound, but you're four years off. It is college, in fact, where a wide exposure is most important. Graduate school is where you can stockpile the knowledge you want. The people who do best in graduate school are not the ones who go to a brand-name school with a high-powered department. It's the ones who get four years of growth and maturity as an undergraduate. This is not mushy, wishful thinking; it's been studied extensively.
...but most people know what they want to focus on in college
A decade after graduation, 9 out of 10 people are employed in a field unrelated to their undergraduate major. This statistic is a little lower for engineers -- IMO this is exactly because engineers often lack the background (and, I daresay, cognitive breadth) to successfully change fields. In contrast, people with broad exposure to the liberal arts tend to be able to bootstrap into any number of areas.
Certainly universities should offer choice...
The original poster made the point that, when offered a choice, many engineering students will not make that choice. This is really not okay. It's not always the humanities that are optional... I've looked through several dozen undergraduate CS curricula and I was shocked to see how many of them do not require any theory. I've had friends and students go to these schools and I have to beg them to take the theory classes -- usually to no avail. Such schools, including famous and reputable ones, are content to graduate those aforementioned book-smart number-crunchers with no grasp of fundamentals.
Well, it is no great leap to state that "fundamentals" includes a reasonable understanding of things outside of the chosen major. When chatting with my fellow engineers, I'm constantly correcting them in basic matters of physics, biology, and higher mathematics, to say nothing of politics, sociology, psychology, art, history, and philosophy -- and they are often loath to believe me. They're fantastic engineers, but they are not only ignorant of fairly important things, they often cannot even think critically outside of their field.
The discipline of engineering is an immensely powerful cognitive tool, but it is not a universal one. Engineers are uniquely guilty of the "if you only have a hammer" effect (everything starts to look like a nail). That's what college is for.
Again, I am an engineer. I am not condemning the profession, or my colleagues, but an educational system that cheats its students by allowing them to graduate with nothing more than four years' worth of job skills.
From a technical standpoint, you always have much more footage than will be used in the final cut. This is called the "shooting ratio" and is usually 6:1 or 8:1 or more. 3:1 is considered very low, while some very unusual movies (like Apocalypse Now) ended up being more like 100:1.
From an aesthetic standpoint, it's impossible to tell beforehand how all the parts of the movie will come together. There has to be a lot of leeway for postproduction to make adjustments. Walter Murch tells of a scene that he decided at length to omit. Coppola, the director, agreed with the cut but mentioned (with some regret) that the deleted scene was one of the reasons that he made the movie. Coppola didn't think anything more of it, but Murch took this lesson to heart: scenes may serve purposes other than simply "being in the film." In this case, the scene served as context, backstory, and inspiration to the director, and as such it probably influenced every other shot that was filmed.
As the other poster pointed out very succinctly, a corporation has a seperate existence from the people (or "PEOPLE," as you put it) that comprise it. Under U.S. law, this entity has a number of rights that are based on the rights granted to a human citizen.
None of this impinges on the free speech of its employees. I have no idea where you got that one.
When a corporation "speaks," as through an advertisement or press release, it is expressing the opinion of the corporate entity. This may have nothing to do with the opinions of the people in the corporation. Nonsense, you say? Consider that the "official corporate line" is often a compromise consensus among its senior executives, and may not reflect the complete opinion of any of them (much less the low-level staff who actually write, deliver, or speak this opinion).
Or it may be the opinion of the board of directors, imposed by fiat upon the management (which is required to carry out those decisions). The directors are often not employees of the corporation, but rather shareholders.
Speaking of shareholders, if the constituents ("PEOPLE") of a corporation begin to act against the corporation's interest, the shareholders may remove those constituents. This is often used to oust incompetent or corrupt management, which is good. But it's a double-edged sword. A company that prioritizes ethics over profits is just as likely to find itself the subject of a shareholder lawsuit.
Which brings us all the way back to the beginning of the thread: over the last 150 years, corporations have steadily accrued these human-like rights, but under no circumstances does a corporation behave in a human-like manner.
And what human-like rights does a corporation have?
You must be kidding. Check the law, for starters.
Among other things, a corporation has the right to free speech (of all things!). They typically exercise this "right" to slander people out of office if they can't buy them outright. At the same time, you don't get to do that to them.
A single iteration of the Miller-Rabin test does not take constant time. The time taken involves a number of multiplications mod p (mostly repeated squaring). Both the number of math operations and the difficulty of each one increase with p.
It is therefore polynomial in the size of p, and not constant.
Oops! Yes, this is completely true. Anything that involves multiplications is obviously not going to be constant in the size of the input.
I should offer my head on a platter to Prof. Rabin! I took his class on randomized algorithms...
(If I wanted to be petulant, I could blame the Motwani & Raghavan book for modeling multiplies as unit cost, but that ain't gonna fool anyone. =)
This result, if true, is very interesting from a theory standpoint.
As far as practice, it's fairly irrelevant. Probabilistic primality testing can be done in constant time with bounded error.
The Miller-Rabin test will tell you if a number is prime with at most 1/4 probability of error. That sounds ridiculous, but the catch is that you can iterate it using a random parameter. Do the test twice and your probability drops to 1/16. Do it fifteen times and your chances of being wrong are about one billionth.
If you're truly paranoid, do it 50 times. That'll bring the error rate of the algorithm magnitudes below the error rate of your hardware.
More to the point, you want the frequency-response curve of the speaker to be flat as possible throughout the range of human hearing. That means the dropoff at the end of the curve has to be outside the range of human hearing.
WaiWai is pretty much a tabloid. It doesn't seem to be quite as prone to pure fabrication as the usual American staples, but it's close enough.
--- Dum de dum.
Re:Get your head from out of your arse
on
The Almighty Buck
·
· Score: 1
Hmm, getting offtopic, but I guess an AC post modded up from 0 deserves some response...
They should be applying to Harvard. [...] You get an honors degree if you have a B-minus average. 91% honors graduation rate.
First, let's see you get in. Then, let's see you pick a major where grade inflation applies. Given that you're reading Slashdot, you'd probably choose something techy... and discover, unpleasantly, that the engineering, math, and science departments aren't too shy about giving out the right grades.
I graduated with honors, by the skin of my teeth.
That includes failing a math class that was a few years too advanced for me -- none of this "gentleman's C" stuff there.
Now, what I said earlier about the Ivy League being overrated is true, for undergrads. Harvard is a research university, research happens in graduate schools, and the towering intellectuals produced by the U.S. come out of the graduate system.
But the graduate system is powered by the relatively democratized undergrad system. The ones who do best are not those who get into a fancy undergrad school, but those who spend four undergraduate years honing their mental skills (not to mention work ethic, etc.), and thereby go on to do excellent work in graduate school. Overall, this is much more effective than the paradigm popular outside the U.S., which is to filter students as early as possible -- at 18 or even 14 years of age, before their brain is even finished growing.
By extension, colleges that promote this progression produce better results later on. There is an utterly irrational tendency to judge a school by who they admit rather than who they produce. The late Stephen Jay Gould graduated from a little-known college called Antioch. It's a tiny, tiny school in the middle of Ohio somewhere, that admits mostly B students with average SAT scores around 1000. But if you look at the students it graduates, there are few schools in the world that can match it. There are plenty of such schools around if you look for them, and I for one, do not wonder why this country is so rich.
---
Dum de dum.
Re:Get your head from out of your arse
on
The Almighty Buck
·
· Score: 2
I don't know about Australia, but as a college counselor I've had plenty of students relate the same story: their relatives from other countries rarely recognize any name but Harvard. Literally: "What the hell is {Princeton,MIT,Stanford}? Why aren't you applying to Harvard?"
...her universities are the queens of learning...
This is actually correct in a way that the author of the article probably didn't realize. The key is the word university, which is an institution that awards graduate degrees, as opposed to a college, which grants undergraduate degrees. The U.S. university system -- the graduate system -- is second to none, period. Oxford can certainly compete on equal footing with the best of these universities, but how many Oxfords are there?
In contrast, the U.S. undergraduate experience is only a bit better than mediocre (although there are plenty of excellent exceptions, none of which belong to the overrated Ivy League). The quantity of those attending college greatly compensates for any lack of quality, and is probably crucial to continued abundance -- it seems to me that providing a great education to only a few is less effective than providing a good education to many. But I'd love to see hard numbers if anyone has any...
Seems like you could eliminate a good portion of the MMORPG cheaters by imposing real-world penalties. Like, say, $75 per offense after three warnings. This would be particularly effective against twelve-year-old scrubs who need to learn that anonymity isn't a blank check... and whose parents are probably not thrilled with the $15/month fee in the first place.
Of course, there are obvious obstacles, else we'd have seen this done already... I suppose even a single false positive is unacceptable (for public relations if nothing else). But you don't need to nail every cheater; you don't even have to come close. Stick to verifiable, airtight cases -- by keeping logs, for example, to complement the human GMs used today -- and then make big, flaming examples of them.
This wouldn't replace technological solutions. Ideally, it would bring the amount of cheating down to a level where anti-cheats could be more targeted and perhaps therefore more effective.
I wonder if this might be inviting lawsuits... but considering the Evil that's already present in the typical EULA, I wouldn't expect any problems. IANAL.
I've been involved with college counseling for about seven years now, and the mantra I hear most often is, "Get good grades, get into a good college, get a good job." None of these things hold much water, except maybe the first one, sometimes.
You still need to major in something that will get you a job, but that's a side-effect: in 1985, a survey of Yale graduates from the 1960's showed that about 75% of them were working in jobs that did not exist when they graduated.
The real point of college should be to improve your abilities and talents. And that includes your ability to think, to learn, and to adapt, not to mention secondary skills like discipline, communication, etc. This doesn't always happen, but that's the goal, at least. I had a fabulous education in CS that landed a great job right out of school, but I've always considered that to be the lesser half of my college experience when I look at my life overall. Many who've been through college will say that this is just mushy claptrap. I say, they're the ones who missed out (possibly through no fault of their own).
However, I do hold that liberal arts schools (i.e. essentially anything that is not a "tech" school like RPI or MIT) are generally much better at this, because they deliberately expose you to subjects that, over the centuries, have proven useful in training the mind. Even for the technically-inclined, there is little or no disadvantage to attending a non-tech school at the undergraduate level unless the the department you're interested in is particularly tiny, or has bad lab facilities.
And, as others have mentioned, don't underestimate the friends, mentors, and contacts that are part of the package. The dean of freshmen at Harvard told me once that more than half the value of any college education comes from the people, not the classes -- which I agree with, but it was interesting to see that coming from inside the administration.
Okay, I'll stop ranting now or I'll be here for hours...
Think of mp3 decoding, divx playing, 3d games, etc...
No, SSE or 3DNow is a much better choice than -ffastmath for that kind of thing. I'm pretty sure Winamp supports both instruction sets, and I'd be very surprised if the compute-heavy DivX didn't employ them, too.
You might give up precision when using SIMD (on x86 anyway) but at least you can control that. Using -ffastmath is much more of a toss-up. Case in point, LAME has been observed to produce seriously different results when compiled with -ffastmath.
I wonder what exactly it was that prevented Alpha-based machines from taking a position like Sun or IBM
It's the usual answer: DEC never knew how to market them. And Compaq? As far as I'm concerned, they never even tried.
I joined DEC right out of college. Exactly one week later, it became Compaq. None of the employees knew anything about it until it happened. But Compaq had damage control prepped and ready to go: the line they fed us was that corporate purchasers usually invite the top three companies to bid on a contract, and since DEC was fourth (after Sun, IBM, and HP, apparently), it was disproportionately locked out of the game. But whenever DEC managed to get invited, they would usually win. So the "strategy" for the new behemoth was pretty much that they expected to get invited everywhere and win lots of contracts. Almost without trying.
Well, we all know how Alphaserver sales just took off after that, don't we?
Interestingly, they had well-known DEC execs deliver these fabulously optimistic forecasts... execs that promptly departed before the integration even began. (Not that anything resembling integration actually happened anyway.)
I guess Alpha was pitched as more of a number crunching box.
Not really. Alpha did have that reputation, for obvious reasons, and it had a stable market in the technical computing field (CERN and LLNL come to mind). But that's a fairly small niche, not enough to sustain the business. The wider market penetration just never happened. When I left DEC in, I dunno, 1999 or something, I couldn't tell what marketing was doing at all. It was listless, confused, and worse than directionless.
And even that came to an end, didn't it? I didn't notice at the time, but in retrospect I don't recall seeing any kind of public support of Alpha after 2000 or so.
And so it faded away. My blood, sweat, and tears are in Digital Unix, but I began and concluded my mourning months ago, when Compaq murdered Alpha and handed its head to Intel on a platter. This merger is a postlude, nothing more...
---
I like canned peaches.
Re:A message from the RadLight Admin
on
Spyware Fights Back
·
· Score: 2, Funny
One of these things is not like the others,
one of these things just doesn't belong.
Can you tell which thing is not like the others,
By the time I finish my song?
If I understand you correctly, I think you're saying that making real images look fake is no defense, if there's any corroboration that actual abuse occurred -- even if the corroboration would not have held up by itself.
Hmm, interesting. Point taken. That certainly applies to the second case I quoted, where they actually did catch the guy abducting a kid.
I'm skeptical that it would always be enough to make the case. In this country, at least, the defense only needs to establish a "reasonable doubt" that the physical evidence is related to the images. But IANAL.
In any case, this would likely have done little for the Wonderland Club case, where most of the members were consumers rather than abusers. It's still illegal to own real child porn, and people who purchase it would be unprosecutable.
Preemptive strike against replies: this is orthogonal to the claim that virtual child porn reduces the market for real child porn (with which I will neither agree nor disagree until I see evidence). The situation we're talking about is when a person is found with real pictures that look fake, that person would win in court.
Congress justified the wider ban on grounds that while no real children were harmed in creating the material, real children could be harmed by feeding the prurient appetites of pedophiles or child molesters.
This is sensationalism. The ruling alluded to this, but as I see it, the fundamental intention of the law was to allow enforcement of existing laws.
No, I'm not talking about letting police go on a rampage against anything that might possibly be mistaken for child porn. The law is obviously much, much too broad. But the intent is justifiable: to ensure that the proliferation of fake child porn does not erode our current capability to eliminate real child porn.
Typical of Slashdot, the knee-jerk mantra is "if it's virtual, it's harmless." That's a nice sentiment but it's also a naive analysis coming from a population that prides itself on lateral thinking. Consider, perhaps, that you can now take real child porn and run it through Photoshop to make it indistinguishable from a fake. This would be a powerful defense for people interested in raping children and getting away with it.
Let us inject some concreteness into the discussion. Suppose that the Wonderland Club might have gone free with nothing more than a few strokes of the smudge tool. These guys would have had a more difficult job making their torture videos look fake, but technology would've permitted it in a few years. No, I don't like the idea of restricting anything virtual, but propose a solution then, if you can think of one.
"The youngest victim was about three months and other victims aged up to about 18 years."
You're confusing preemption with mutual exclusion. Being preemptible only means you can be suspended. It does not mean the task that someone gets to break your locks.
If a low-priority task locks a resource, it can still be preempted by a high-priority task... but if the high-priority task also wants that resource, it's going to have to get in line just like everyone else. This is also what leads to the possibility of priority inversion.
Nothing changes if you substitute "kernel task" for "task" in the preceding paragraph.
Corruption of data in memory can be repaired with a reboot while corruption of data on disk cannot.
This is true, but there's no guarantee that corrupted data in memory won't get written to disk, or otherwise affect the disk. This is actually less likely in Linux than in other OSes, since the Linux kernel isn't pageable (but we may see pageable page tables in the future, at least in Rik's VM). NT and Mach-derivatives, being microkernels, are mostly pageable, and bad swap would be a lot riskier there... particularly if it hits filesystem structures or buffers.
Kernels aside, any application that writes to disk may be writing out bad data.
I also consider the risk of a program executing a valid but incorrect operation to be vanishingly small.
This is also true. Corruption is very likely to cause an immediate, simple failure. I wouldn't be concerned about executing an incorrect operation as much as executing the right operation on the wrong data. Most code doesn't perform exhaustive parameter checking at every step. The best you usually find is that stuff gets verified at the external API, while internal APIs do less -- or none. Most software is much worse than that, even widely-deployed code that's considered fairly stable, like C libraries. See the Ballista project. Corruption in RAM (swap) can cause data that was correct at the external API to be wrong later on, and nothing will catch it because it's assumed to have been checked.
This, of course, doesn't even apply to opaque data that the application isn't supposed to examine in the first place. It's unlikely that cp or gzip would have its buffers swapped out and corrupted, but I'd just as soon not take that chance (small as it might be). Corruption may not bite you until long after it's happened.
You can't lose valuable data if you didn't have any in the first place.
I agree, but the principle is invariant. Given a pile of worthless data, what is the worst thing you can do to it? Corruption, I'd submit. Defacing an AOL CD is much more fun than just throwing them away intact.
No, read the post again. In fact, read the part you quoted again.
This isn't about just crashing, this is about corruption. Corruption is the single worst thing that can happen to a machine. Worse than throwing it out the window? Yes, because that will only cause it to fail. Failure is much better than corruption.
Corruption usually results in failure, but not always. This is when the computer starts giving wrong answers and you don't know it.
Let use your example. Without the mod, buggy Netscape crashes anyway, and you don't purchase 150 shares of Enron on eTrade. But that's okay because you know it crashed, so you go and bring up the page again.
On the other hand, corrupted Netscape doesn't crash, and lets you place the order. And then, you gasp in horror when it says you've purchased 47,229 shares of Enron.
Slashdot Mirror
MP3 is not a free standard, regardless of the existence of free downloads. Trane Francks even went so far as to quote the license itself. The "tons of free, licensed encoders" you mention do not exist. They're all unlicensed, unless you've found one whose author is fronting the $5/unit+$15k/yearly royalty to Thomson. Non-commercial use is irrelevant for encoders (and now they're claiming that for decoders, too).
I will happily continue to use LAME, but I'm under no delusions about its legality, and neither are its developers.
--
Dum de dum.
It's free for ME, which is the free I care about.
Sigh. I question whether I should even bother...
Stealing a car does not make cars free.
Copying Photoshop does not make Photoshop free software.
Downloading an unlicensed encoder does not make MP3 a free and open standard.
--
Dum de dum.
This isn't really so off-topic, given Olin's nonexistent tuition.
The Ivies and other such private colleges give aid to well over half of their students. (It's actually much more than half but I don't have the figures with me.) Consider, a typical upper-middle-income family makes $50k per parent. That's $100k per year, $75k after taxes. It's unreasonable to expect $35k of that to go for tuition. All things being equal, this family might receive around $10k of aid or more, not including loans.
Put bluntly, an income of $100k is smack in the middle of the distribution curve at most well-endowed schools.
Yet, I've talked with dozens of families with incomes in that range and the overwhelming majority of them say, "Oh, we make $100k per year, we know that means we're ineligible for aid." And so they don't even bother to apply. It's a deadly myth that's doomed legions of promising students to choose a college based on cost.
Incidentally, the Ivies' elite status is almost entirely a result of their selectivity, and has little to do with their cost or their overrated quality (yes, I went to one).
---
Dum de dum.
...for $10K-$30K/year they should be allowed to stockpile as much knowledge as possible in their chosen fields.
...but most people know what they want to focus on in college
This idea is sound, but you're four years off. It is college, in fact, where a wide exposure is most important. Graduate school is where you can stockpile the knowledge you want. The people who do best in graduate school are not the ones who go to a brand-name school with a high-powered department. It's the ones who get four years of growth and maturity as an undergraduate. This is not mushy, wishful thinking; it's been studied extensively.
A decade after graduation, 9 out of 10 people are employed in a field unrelated to their undergraduate major. This statistic is a little lower for engineers -- IMO this is exactly because engineers often lack the background (and, I daresay, cognitive breadth) to successfully change fields. In contrast, people with broad exposure to the liberal arts tend to be able to bootstrap into any number of areas.
Certainly universities should offer choice...
The original poster made the point that, when offered a choice, many engineering students will not make that choice. This is really not okay. It's not always the humanities that are optional... I've looked through several dozen undergraduate CS curricula and I was shocked to see how many of them do not require any theory. I've had friends and students go to these schools and I have to beg them to take the theory classes -- usually to no avail. Such schools, including famous and reputable ones, are content to graduate those aforementioned book-smart number-crunchers with no grasp of fundamentals.
Well, it is no great leap to state that "fundamentals" includes a reasonable understanding of things outside of the chosen major. When chatting with my fellow engineers, I'm constantly correcting them in basic matters of physics, biology, and higher mathematics, to say nothing of politics, sociology, psychology, art, history, and philosophy -- and they are often loath to believe me. They're fantastic engineers, but they are not only ignorant of fairly important things, they often cannot even think critically outside of their field.
The discipline of engineering is an immensely powerful cognitive tool, but it is not a universal one. Engineers are uniquely guilty of the "if you only have a hammer" effect (everything starts to look like a nail). That's what college is for.
Again, I am an engineer. I am not condemning the profession, or my colleagues, but an educational system that cheats its students by allowing them to graduate with nothing more than four years' worth of job skills.
---
Dum de dum.
From a technical standpoint, you always have much more footage than will be used in the final cut. This is called the "shooting ratio" and is usually 6:1 or 8:1 or more. 3:1 is considered very low, while some very unusual movies (like Apocalypse Now) ended up being more like 100:1.
From an aesthetic standpoint, it's impossible to tell beforehand how all the parts of the movie will come together. There has to be a lot of leeway for postproduction to make adjustments. Walter Murch tells of a scene that he decided at length to omit. Coppola, the director, agreed with the cut but mentioned (with some regret) that the deleted scene was one of the reasons that he made the movie. Coppola didn't think anything more of it, but Murch took this lesson to heart: scenes may serve purposes other than simply "being in the film." In this case, the scene served as context, backstory, and inspiration to the director, and as such it probably influenced every other shot that was filmed.
---
Dum de dum.
You don't work for a corporation, do you?
As the other poster pointed out very succinctly, a corporation has a seperate existence from the people (or "PEOPLE," as you put it) that comprise it. Under U.S. law, this entity has a number of rights that are based on the rights granted to a human citizen.
None of this impinges on the free speech of its employees. I have no idea where you got that one.
When a corporation "speaks," as through an advertisement or press release, it is expressing the opinion of the corporate entity. This may have nothing to do with the opinions of the people in the corporation. Nonsense, you say? Consider that the "official corporate line" is often a compromise consensus among its senior executives, and may not reflect the complete opinion of any of them (much less the low-level staff who actually write, deliver, or speak this opinion).
Or it may be the opinion of the board of directors, imposed by fiat upon the management (which is required to carry out those decisions). The directors are often not employees of the corporation, but rather shareholders.
Speaking of shareholders, if the constituents ("PEOPLE") of a corporation begin to act against the corporation's interest, the shareholders may remove those constituents. This is often used to oust incompetent or corrupt management, which is good. But it's a double-edged sword. A company that prioritizes ethics over profits is just as likely to find itself the subject of a shareholder lawsuit.
Which brings us all the way back to the beginning of the thread: over the last 150 years, corporations have steadily accrued these human-like rights, but under no circumstances does a corporation behave in a human-like manner.
---
Dum de dum.
And what human-like rights does a corporation have?
You must be kidding. Check the law, for starters.
Among other things, a corporation has the right to free speech (of all things!). They typically exercise this "right" to slander people out of office if they can't buy them outright. At the same time, you don't get to do that to them.
---
Dum de dum.
Just looking at the number is already linear time!
Almost. The length of p is log(p). (But yeah, that's more than enough to make Miller-Rabin polynomial, not constant.)
---
Dum de dum.
A single iteration of the Miller-Rabin test does not take constant time. The time taken involves a number of multiplications mod p (mostly repeated squaring). Both the number of math operations and the difficulty of each one increase with p.
It is therefore polynomial in the size of p, and not constant.
Oops! Yes, this is completely true. Anything that involves multiplications is obviously not going to be constant in the size of the input.
I should offer my head on a platter to Prof. Rabin! I took his class on randomized algorithms...
(If I wanted to be petulant, I could blame the Motwani & Raghavan book for modeling multiplies as unit cost, but that ain't gonna fool anyone. =)
---
Dum de dum.
This result, if true, is very interesting from a theory standpoint.
As far as practice, it's fairly irrelevant. Probabilistic primality testing can be done in constant time with bounded error.
The Miller-Rabin test will tell you if a number is prime with at most 1/4 probability of error. That sounds ridiculous, but the catch is that you can iterate it using a random parameter. Do the test twice and your probability drops to 1/16. Do it fifteen times and your chances of being wrong are about one billionth.
If you're truly paranoid, do it 50 times. That'll bring the error rate of the algorithm magnitudes below the error rate of your hardware.
---
Dum de dum.
More to the point, you want the frequency-response curve of the speaker to be flat as possible throughout the range of human hearing. That means the dropoff at the end of the curve has to be outside the range of human hearing.
--
Dum de dum.
WaiWai is pretty much a tabloid. It doesn't seem to be quite as prone to pure fabrication as the usual American staples, but it's close enough.
---
Dum de dum.
Hmm, getting offtopic, but I guess an AC post modded up from 0 deserves some response...
They should be applying to Harvard. [...] You get an honors degree if you have a B-minus average. 91% honors graduation rate.
First, let's see you get in. Then, let's see you pick a major where grade inflation applies. Given that you're reading Slashdot, you'd probably choose something techy... and discover, unpleasantly, that the engineering, math, and science departments aren't too shy about giving out the right grades.
I graduated with honors, by the skin of my teeth. That includes failing a math class that was a few years too advanced for me -- none of this "gentleman's C" stuff there.
Now, what I said earlier about the Ivy League being overrated is true, for undergrads. Harvard is a research university, research happens in graduate schools, and the towering intellectuals produced by the U.S. come out of the graduate system.
But the graduate system is powered by the relatively democratized undergrad system. The ones who do best are not those who get into a fancy undergrad school, but those who spend four undergraduate years honing their mental skills (not to mention work ethic, etc.), and thereby go on to do excellent work in graduate school. Overall, this is much more effective than the paradigm popular outside the U.S., which is to filter students as early as possible -- at 18 or even 14 years of age, before their brain is even finished growing.
By extension, colleges that promote this progression produce better results later on. There is an utterly irrational tendency to judge a school by who they admit rather than who they produce. The late Stephen Jay Gould graduated from a little-known college called Antioch. It's a tiny, tiny school in the middle of Ohio somewhere, that admits mostly B students with average SAT scores around 1000. But if you look at the students it graduates, there are few schools in the world that can match it. There are plenty of such schools around if you look for them, and I for one, do not wonder why this country is so rich.
---
Dum de dum.
I don't know about Australia, but as a college counselor I've had plenty of students relate the same story: their relatives from other countries rarely recognize any name but Harvard. Literally: "What the hell is {Princeton,MIT,Stanford}? Why aren't you applying to Harvard?"
...her universities are the queens of learning...
This is actually correct in a way that the author of the article probably didn't realize. The key is the word university, which is an institution that awards graduate degrees, as opposed to a college, which grants undergraduate degrees. The U.S. university system -- the graduate system -- is second to none, period. Oxford can certainly compete on equal footing with the best of these universities, but how many Oxfords are there?
In contrast, the U.S. undergraduate experience is only a bit better than mediocre (although there are plenty of excellent exceptions, none of which belong to the overrated Ivy League). The quantity of those attending college greatly compensates for any lack of quality, and is probably crucial to continued abundance -- it seems to me that providing a great education to only a few is less effective than providing a good education to many. But I'd love to see hard numbers if anyone has any...
---
Dum de dum.
Seems like you could eliminate a good portion of the MMORPG cheaters by imposing real-world penalties. Like, say, $75 per offense after three warnings. This would be particularly effective against twelve-year-old scrubs who need to learn that anonymity isn't a blank check... and whose parents are probably not thrilled with the $15/month fee in the first place.
Of course, there are obvious obstacles, else we'd have seen this done already... I suppose even a single false positive is unacceptable (for public relations if nothing else). But you don't need to nail every cheater; you don't even have to come close. Stick to verifiable, airtight cases -- by keeping logs, for example, to complement the human GMs used today -- and then make big, flaming examples of them.
This wouldn't replace technological solutions. Ideally, it would bring the amount of cheating down to a level where anti-cheats could be more targeted and perhaps therefore more effective.
I wonder if this might be inviting lawsuits... but considering the Evil that's already present in the typical EULA, I wouldn't expect any problems. IANAL.
---
Dum de dum.
I've been involved with college counseling for about seven years now, and the mantra I hear most often is, "Get good grades, get into a good college, get a good job." None of these things hold much water, except maybe the first one, sometimes.
You still need to major in something that will get you a job, but that's a side-effect: in 1985, a survey of Yale graduates from the 1960's showed that about 75% of them were working in jobs that did not exist when they graduated.
The real point of college should be to improve your abilities and talents. And that includes your ability to think, to learn, and to adapt, not to mention secondary skills like discipline, communication, etc. This doesn't always happen, but that's the goal, at least. I had a fabulous education in CS that landed a great job right out of school, but I've always considered that to be the lesser half of my college experience when I look at my life overall. Many who've been through college will say that this is just mushy claptrap. I say, they're the ones who missed out (possibly through no fault of their own).
However, I do hold that liberal arts schools (i.e. essentially anything that is not a "tech" school like RPI or MIT) are generally much better at this, because they deliberately expose you to subjects that, over the centuries, have proven useful in training the mind. Even for the technically-inclined, there is little or no disadvantage to attending a non-tech school at the undergraduate level unless the the department you're interested in is particularly tiny, or has bad lab facilities.
And, as others have mentioned, don't underestimate the friends, mentors, and contacts that are part of the package. The dean of freshmen at Harvard told me once that more than half the value of any college education comes from the people, not the classes -- which I agree with, but it was interesting to see that coming from inside the administration.
Okay, I'll stop ranting now or I'll be here for hours...
---
Dum de dum.
You might give up precision when using SIMD (on x86 anyway) but at least you can control that. Using -ffastmath is much more of a toss-up. Case in point, LAME has been observed to produce seriously different results when compiled with -ffastmath.
---
Dum de dum.
I joined DEC right out of college. Exactly one week later, it became Compaq. None of the employees knew anything about it until it happened. But Compaq had damage control prepped and ready to go: the line they fed us was that corporate purchasers usually invite the top three companies to bid on a contract, and since DEC was fourth (after Sun, IBM, and HP, apparently), it was disproportionately locked out of the game. But whenever DEC managed to get invited, they would usually win. So the "strategy" for the new behemoth was pretty much that they expected to get invited everywhere and win lots of contracts. Almost without trying.
Well, we all know how Alphaserver sales just took off after that, don't we?
Interestingly, they had well-known DEC execs deliver these fabulously optimistic forecasts... execs that promptly departed before the integration even began. (Not that anything resembling integration actually happened anyway.)
Not really. Alpha did have that reputation, for obvious reasons, and it had a stable market in the technical computing field (CERN and LLNL come to mind). But that's a fairly small niche, not enough to sustain the business. The wider market penetration just never happened. When I left DEC in, I dunno, 1999 or something, I couldn't tell what marketing was doing at all. It was listless, confused, and worse than directionless.
And even that came to an end, didn't it? I didn't notice at the time, but in retrospect I don't recall seeing any kind of public support of Alpha after 2000 or so.
And so it faded away. My blood, sweat, and tears are in Digital Unix, but I began and concluded my mourning months ago, when Compaq murdered Alpha and handed its head to Intel on a platter. This merger is a postlude, nothing more...
---
I like canned peaches.
one of these things just doesn't belong.
Can you tell which thing is not like the others,
By the time I finish my song?
---
Dum de dum.
If I understand you correctly, I think you're saying that making real images look fake is no defense, if there's any corroboration that actual abuse occurred -- even if the corroboration would not have held up by itself.
Hmm, interesting. Point taken. That certainly applies to the second case I quoted, where they actually did catch the guy abducting a kid.
I'm skeptical that it would always be enough to make the case. In this country, at least, the defense only needs to establish a "reasonable doubt" that the physical evidence is related to the images. But IANAL.
In any case, this would likely have done little for the Wonderland Club case, where most of the members were consumers rather than abusers. It's still illegal to own real child porn, and people who purchase it would be unprosecutable.
Preemptive strike against replies: this is orthogonal to the claim that virtual child porn reduces the market for real child porn (with which I will neither agree nor disagree until I see evidence). The situation we're talking about is when a person is found with real pictures that look fake, that person would win in court.
--
Dum de dum.
Congress justified the wider ban on grounds that while no real children were harmed in creating the material, real children could be harmed by feeding the prurient appetites of pedophiles or child molesters.
This is sensationalism. The ruling alluded to this, but as I see it, the fundamental intention of the law was to allow enforcement of existing laws.
No, I'm not talking about letting police go on a rampage against anything that might possibly be mistaken for child porn. The law is obviously much, much too broad. But the intent is justifiable: to ensure that the proliferation of fake child porn does not erode our current capability to eliminate real child porn.
Typical of Slashdot, the knee-jerk mantra is "if it's virtual, it's harmless." That's a nice sentiment but it's also a naive analysis coming from a population that prides itself on lateral thinking. Consider, perhaps, that you can now take real child porn and run it through Photoshop to make it indistinguishable from a fake. This would be a powerful defense for people interested in raping children and getting away with it.
Let us inject some concreteness into the discussion. Suppose that the Wonderland Club might have gone free with nothing more than a few strokes of the smudge tool. These guys would have had a more difficult job making their torture videos look fake, but technology would've permitted it in a few years. No, I don't like the idea of restricting anything virtual, but propose a solution then, if you can think of one.
"The youngest victim was about three months and other victims aged up to about 18 years."
---
Dum de dum.
You're confusing preemption with mutual exclusion. Being preemptible only means you can be suspended. It does not mean the task that someone gets to break your locks.
If a low-priority task locks a resource, it can still be preempted by a high-priority task... but if the high-priority task also wants that resource, it's going to have to get in line just like everyone else. This is also what leads to the possibility of priority inversion.
Nothing changes if you substitute "kernel task" for "task" in the preceding paragraph.
--
I like canned peaches.
Corruption of data in memory can be repaired with a reboot while corruption of data on disk cannot.
This is true, but there's no guarantee that corrupted data in memory won't get written to disk, or otherwise affect the disk. This is actually less likely in Linux than in other OSes, since the Linux kernel isn't pageable (but we may see pageable page tables in the future, at least in Rik's VM). NT and Mach-derivatives, being microkernels, are mostly pageable, and bad swap would be a lot riskier there... particularly if it hits filesystem structures or buffers.
Kernels aside, any application that writes to disk may be writing out bad data.
I also consider the risk of a program executing a valid but incorrect operation to be vanishingly small.
This is also true. Corruption is very likely to cause an immediate, simple failure. I wouldn't be concerned about executing an incorrect operation as much as executing the right operation on the wrong data. Most code doesn't perform exhaustive parameter checking at every step. The best you usually find is that stuff gets verified at the external API, while internal APIs do less -- or none. Most software is much worse than that, even widely-deployed code that's considered fairly stable, like C libraries. See the Ballista project. Corruption in RAM (swap) can cause data that was correct at the external API to be wrong later on, and nothing will catch it because it's assumed to have been checked.
This, of course, doesn't even apply to opaque data that the application isn't supposed to examine in the first place. It's unlikely that cp or gzip would have its buffers swapped out and corrupted, but I'd just as soon not take that chance (small as it might be). Corruption may not bite you until long after it's happened.
--
I like canned peaches.
You can't lose valuable data if you didn't have any in the first place.
I agree, but the principle is invariant. Given a pile of worthless data, what is the worst thing you can do to it? Corruption, I'd submit. Defacing an AOL CD is much more fun than just throwing them away intact.
No, read the post again. In fact, read the part you quoted again.
This isn't about just crashing, this is about corruption. Corruption is the single worst thing that can happen to a machine. Worse than throwing it out the window? Yes, because that will only cause it to fail. Failure is much better than corruption.
Corruption usually results in failure, but not always. This is when the computer starts giving wrong answers and you don't know it.
Let use your example. Without the mod, buggy Netscape crashes anyway, and you don't purchase 150 shares of Enron on eTrade. But that's okay because you know it crashed, so you go and bring up the page again.
On the other hand, corrupted Netscape doesn't crash, and lets you place the order. And then, you gasp in horror when it says you've purchased 47,229 shares of Enron.
(Yeah, yeah, I know it's been delisted. =)
--
I like canned peaches.