I firmly believe transport costs of pure platinum from Mars would be high enough to make extraction from sea water look dirt cheap. Recycling is another thing that will not let the prices go that high. Extraction of gold from used electronics will be cheaper than importing the stuff from Mars.
Okay, but do you have any particular reason to believe this, or is it just a tenet of your faith? If you consider that fuel can be made relatively cheaply from local ingredients (just react some H_2 with the atmosphere, really) and that transport time isn't important for cargo, it might not be too expensive at all. Strap a booster onto your block-o-platinum and loft into Martian orbit (low gravity, so lots easier than for Earth). Fire up an ion/magsail/Vasimir/whatever engine and two years later you're aerobreaking into Earth orbit.
By far the largest cost to mining on Mars is going to be transporting and supporting the human miners -- which, sadly, makes robots a promising alternative. It'll be interesting to see which gets there first, robots sophisticated enough for autonomous mining operations, or launch costs low enough to realistically support human extraterrestrial colonization.
First, magnetostatic attraction drops off as distance cubed.
Second, you can actually do a lot better than that, because planes conduct electricity. Thus, with appropriate dynamic fields, you can set up circulating currents in the skin or frame, which you can then couple to with other fields. We're talking big antenna arrays here, but it might just be possible.
Third, any such attempt is likely to melt the plane via inductive heating before you actually "tractor beam" it to the ground.
Finally, for the energy you're expending, a laser or rail/coil-gun setup is vastly more likely to work. And cooler, too.
it seems to me that knowing how something behave has not the same meaning than knowing how it works, the latter implying, for me, a notion of why it behave this way. Correct me if I'm wrong. For myself, I surely understand how those forces behave, but not why.
You are entirely correct. Unfortunately, this is almost always the case in the world of science, and it's a fact that many people forget. For the most part, we don't know why anything works the way it does.
Of course, that philosophy will get you in trouble with my friends in the string theory department, most of whom take the view that when you get to a deep enough level of understanding, the workings of the universe are entirely defined by the underlying mathematical structures. Their argument is that, at some level, there isn't anything left to describe, and physics reduces to a bunch of staggeringly nonlinear mathematical relationships (mostly dealing with the topology of the spacetime metric at string-theory scales).
Of course, I'm entirely unqualified to speak further on this subject, as I mostly do astrophysics myself. But hey, when you have friends who will randomly drop by to announce that they no longer believe in spacetime, and can plausibly back it up, you learn stuff.:-)
Essentially this collection is the basics of how to build a computer system after the apocalypse.
Okay, I can't resist --
In that case, said post-apocalyptic peoples probably want to know how to build a computer, too. So throw in a good book on semiconductor design. What? They don't have fab plants? Okay, toss in some materials science and electrical engineering texts, and some general engineering texts if they want to know how to build the plant itself. They'll love you forever if you leave the parts to assemble a simple mill machine in a back room.
Of course, if they're going to figure their way through semiconductors (or hell, even tube-based circuitry) they'll want a good E&M text, and probably something on quantum mechanics. Chances are, they won't have the math for this, so throw in math texts (calculus derived from number theory, basic real analysis, and abstract algebra seem like a good bare minimum set).
'Course, maybe they can't even read those! (They were *actually* bombed into the stone age? Humans are extinct, and this library has been found by the descendents of the hyper-intelligent white lab mice?) So throw in a printed copy of some of those SETI transmissions, and you're all set. Oh, don't forget some good science fiction, so they'll know why they should go to the trouble of spending three generations chewing through this library.:-)
From another perspective, suppose you write a program which uses the Microsoft Foundation Classes (MFC). Are you prohibited from releasing that software under the GPL because MFC is not GPL'd?
Strictly speaking, the GPL cannot prevent you from distributing anything that you have written yourself, since you are responsible as the copyright holder for enforcing the license. However, you would need to add a special exception to the license giving others permission to link it against the MFC, or else nobody else would be allowed to distribute it.
Take a look at Galeon's license for an example of this. The Gecko library is under MPL, so they have to give special permission to link to it.
For a company that was just declared a monopoly, MS isn't trying very hard to clean up their reputation. Do you think this'll come up in court?
This is an important point -- MS has been behaving recently like they really don't believe that there's going to be another trial, whereas that is exactly what the Appeals court has ordered: a new evidentiary hearing to determine remedies. The ruling explicitly stated that in a fast moving industry like this one, new evidence should be considered to see how circumstances have changed, to construct a remedy that actually solves the problem.
On the positive side for them is the new set of rules for what OEMs can change on systems. (Which isn't that much.) On the negative side, we have this, the problems with Kodak (can't find a link right now), their fight with AOL, and of course their much-discussed attempt to monopolize web services through.NET. Regarding this last one in particular, a new round of discovery on Microsoft internal documents could be particularly damning, since so far most of what we have to go on is rumors and leaks.
Keep an eye on the antitrust case; MS may have seriously mis-played their hand.
NARR: On May 15th, l911, the Supreme Court of the United States declared that Standard Oil was a monopoly in restraint of trade and should be dissolved. Rockefeller heard of the decision while golfing at Kykuit with a priest from the local Catholic church, Father J.P. Lennon.
CHERNOW: And Rockefeller reacted with amazing aplomb. He turned to the Catholic priest and said, "Father Lennon, have you some money?" And the priest was very startled by the question and said, "No." And then he said, "Why?" And Rockefeller replied, "Buy Standard Oil."
NARR: As Rockefeller foresaw, the individual Standard Oil companies were worth more than the single corporation. In the next few years, their shares doubled and tripled in value. By the time the rain of cash was over, Rockefeller had the greatest personal fortune in history -- nearly two percent of the American economy.
The default clauses for software differ slightly from state to state, but are mostly pretty reasonable. You of course have the usual copyright permissions, which basically say you can't redistribute without permission, but otherwise don't much limit your use or disposal of the software. Until a few years back, there was a legal school of thought that you needed explicit permission from the copyright holder to run it, but Congress has since specified that this is not in fact true; in the absense of other agreements, you have the right to run any software you legally possess.
This, incidentally, is one reason why UCITA is so pernicious: UCITA changes many of these defaults to fairly complex and usually undesirable terms which would practically force anyone who wants to distribute software to hire a lawyer and make sure a binding license agreement is attached (and the GPL wouldn't count, because it is a licence on redistribution, which you don't have to agree to before downloading and using software).
VFS layer cleanups, USB enhancements, oh my! And the march to rock-solid stability continues.
Not that I've had much trouble with the earlier 2.4 kernels on the whole, but I wouldn't run my server farm on them. Soon, though, from the look of things.
The folks and FermiLab and CERN regularly have hardware filter 7 Terabytes per second down to "reasonable" data levels. It's mostly done in hardware. I would think it would be fairly easy to filter out half of the traffic (MP3 files, etc), and use a similar fiber to transport all or a filtered portion of the data streams back to friendly territory.
yes, but have you ever seen the amount of equipment needed to do this? At FermiLab, the first stage of data processing is done in the detector circuitry, and occupies a good chunk of the detector's volume (a three story high by 50 meter long piece of equipment, I should mention). Then, an entire floor of a good sized building is filled with racks of mostly custom-built circuitry processes the output of the first stage filters for interesting events.
It's even worse at CERN. They're currently putting up a new building that will be entirely filled with computing hardware to manage the data produced by the experiments when LHC comes online.
Anyway, sure, it's possible to filter that sort of data stream. But could you do it on the seabed? No. I'm not even convinced the NSA could afford many such installations. The price tag for the current incarnation of CDF (one of the primary detectors at FermiLab): around $700 million. And that's using cheap grad student labor to build a good chunk of it.
The strong and weak force have a relatively (no pun intended) limited range. The strong force has a range of only about 1x10^-15 meters.
This is a common misconception. You are thinking of the characteristic length scale for strong interactions, which is inversely related to the characteristic energy scale of strong interactions. However, the gluon is massless, and thus the strong force has infinite range.
The really cool (or aggravating, depending on your field of interest) thing about the strong force is that its strength actually increases according to some ungodly exponant with increasing range (the potential is somewhat complex, so I'd be lying if I quoted an actual number). Thus, the only allowed states for objects with "color charge" turn out to be "color neutral" combinations.
One consequence of this is that you can't ever make a free quark, because the strong potential just keeps going to infinity. And if you try, you eventually get above the quark pair-creation threshold, so whatever combination of quarks needed to return everything to color neutrality just falls out of the quantum field.
No, you're conflating two different things. The SSC (Superconducting SuperCollider) was to be built in Texas; that project was halted partway through construction because the congress-critters got budget happy. The LEP (Large Electron-Positron collider) is CERN's current machine. It was recently shut down to free up the tunnel for an even larger machine called LHC (Large Hadron Collider).
Name --- Energy(TeV)
LEP ---.2 or so
FermiLab --- 1 - 2
LHC --- 7
SSC --- 40 +
As you can see, SSC was quite a loss to the physics community. Still a lot of people grouchy about that.
Poor example, I'm afraid. The "crack baby" scare is little more than a myth. A number of studies, most prominently one from last year out of the U. of Toronto, have found that "crack babies" are, in terms of intellect, statistically indistinguishable from other individuals in the same population. I have also read studies indicating that, while they start out distractable etc., they seem to grow out of it just fine when they hit adolescence.
Sharing movies is illegal. If someone shares a movie, they aren't going to buy it. That's money lost for the makers. What's wrong with crime prevention.
You don't get front-page articles on slashdot about people putting locks in shops to stop people to stealing the merchandise, so why should it be any different when try and stop other kinds of crime that costs money?
On the first point, it's not even clear that sharing files on Gnutella or other services is illegal. Some places (Canada comes to mind) it is definitely legal; see other posts citing media taxes which are theoretically used to compensate artists for the copying. In the United States, we have the Audio Home Recording ACt, which makes it legal to copy music and share it. The catch in this case is that the law has been interpretted to apply only to devices legally recognized as "audio home recording devices" which means mix tapes are protected, but MP3s aren't.
As for the second point, as has been pointed out before, copying data is fundamentally different from stealing merchandise, in that nobody is deprived of anything (and don't tell me about lost profits -- contrary to popular belief, corporations do NOT have a "right to profit"). Nor is this a case of stores protecting their property, either -- the MPAA is harassing consumers for their post-sale activities, which may or may not be in line with the doctrine of first sale.
Anyway, even if the MPAA is legally within its rights, and the Gnutella users are legally in the wrong, it isn't clear that this situation will continue. Some interpretations of the Constitution (not the Supreme's opinion, but that could change) argue that copyright law as it currently stands violates the 1st Amendment and perhaps the Copyright Clause itself, and there ARE those in Congress who would like to fix it.
It's actually better than that. Turns out, the NSA had been researching linear cryptanalysis sometime prior to 1977. So, when their designers get together with IBM's, they hand over some very particular S-boxes for the algorithm. Until the work of Shamir et al in the 90s, nobody knew where they'd pulled them from. It was one of those "We can't tell you where we got them, or why you should use them -- just trust us" deals.
So around 1995, linear cryptanalysis is discovered in the non-classified world, and applied to DES. To everyone's great surprise, when you factor in the storage requirements, linear cryptanalyzing DES is almost exactly as hard as brute forcing it, because of the particular structure of the S-boxes.
So sure, the NSA almost certainly wouldn't *tell* us if they knew how to break RSA, but their mandate *is* the security of the US. As such, they have an interest in getting security that works into the hands of Americans (business in particular, you might notice from the press release -- same reasoning behind designing DES so strong).
Interestingly, some of the more intelligent animal species exhibit the same behavior. Chimpanzees in particular, having been taught sign language, were observed to look behind themselves when "speaking" about the future.
These black holes may have formed from large clumps of matter relatively soon after the big bang, or from very early, super-large, short-lived stars. They could then have grown by absorbing matter from their vicinities in the relatively dense universe of the time.
It seems unlikely that black holes make up much of the dark matter in the universe, although they could be a part. We know by looking at the way galaxies rotate and move, that most of the dark matter is in galaxies, but reasonably evenly distributed through them rather than collected at the center. If this was small black holes we'd notice their Hawking radiation. If it was bigger ones, we'd notice their gravity in various ways.
No! You were doing great up to the comment about the Hawking radiation. As several people have pointed out, Hawking radiation for stellar black holes is in the neighborhood of a nanokelvin above absolute zero -- colder than the microwave background.
You actually can have a galactic dark halo composed of black holes, but lensing studies rule out most of the possible mass ranges. Earth-mass primordial black holes are still a candidate, but since they only form by fairly exotic processes a couple of seconds after the big bang, nobody takes them seriously.
The bit about the super-massive primordial stars is very possibly right on. An upcoming issue of Astrophysical Journal Letters is going to have an article on Pop. III massive black holes, which formed from >100-solar mass stars that formed at the center of primordial density fluctuations. This is very cool -- not only do the resultant black holes do a nice job of nucleating galaxies, but they shine brighly in the early universe before aggregating through mergers into a central supermassive black hole. Neat, yes?
Damn straight. Hawking radiation is thermal, and for a stellar black hole it is in the neighborhood of a nanokelvin above absolute zero -- so the thing is STILL colder than the microwave background.
Lensing studies have actually detected a few quiescent stellar black holes wandering free in our galaxy, relatively close by at that. Except for the lensing, we detect absolutely nothing from where the black holes should be.
In fact, the Hawking lifespan of a stellar black hole is about 10^60 years.
There might be primordial black holes evaporating now; they started out as sub-earth mass black holes. Nobody's seen this happen yet; see my earlier post.
Multiple Big Bangs are allowed -- see Big Crunch. The trouble is, no information is preserved across the singularity, so you can't have structure -- not even singularities like black holes -- come out on the other side.
Now, in principle you might be able to have a big bang spontaneously occur in our universe, creating a subuniverse -- see Multiverse. Michael Turner has been a big proponent of this lately. However, if this occurred it would either destroy our universe by inflating a huge bubble in it (so points in spacetime that were previously close together would now be separated by billions of parsecs) or you wouldn't see it happen because the new subuniverse would be hidden behind an event horizon (masquerading as an ordinary black hole) and thereby NOT distorting our metric to hell.
Either way, you don't get an "overwriting" of our universe with a new one. Unless you have a false vacuum decay or something, which is an entirely different issue, and probably a load of crap.
The trouble with implementing schemes like this is that the cryptographic protocol is almost always the strongest link in the chain. Much easier to attack other parts of the cryptosystem, such as key management, user interface, etc.
Examples: the article mentions that the decryption keys are stored in the "receiver" app. Okay -- disassemble the app, take key, you're in. Another way to do this would be to develop a patch for the binary that no-ops the syscalls that disable printing and the like, but leaves the crypto in place.
Zero-knowledge protocols are cool stuff, but they share a key weakness with many other crypto algorithms: they are deceptively strong. Thus, it is very easy to lose sight of the fact that the application is pretty much guaranteed to be the point of attack, and should be designed accordingly.
As Schneier has pointed out before, this type of scheme probably cannot be implemented -- PCs are intrinsically untrusted platforms. The application cannot be certain that it isn't running in an emulator. If it tries to "protect" the data by wiping the files, it cannot know whether the user will try again, learning from past mistakes, by restoring everything from a disk image. Is there any general way to distinguish between Word and a pipe to a file programmed to present Word's API? No.
So you're accusing Microsoft of misappropriating GPL'd code.
This is actually fairly unlikely. From everything I've heard, MS actually maintains a pretty strict policy of keeping their developers very far away from GPLed code. I've even heard that sourceforge and some other sites are blocked at their firewalls, for just this reason. If nothing else, "accidentally" incorporating GPLed code into their software would seem to open them to all kinds of nasty shareholder lawsuits for negligence.
The problem is that there is still one person who knows about the keyphrase or certificate...you. A better encryption system would make use of one-time keyphrases or certs that not even you know of.
The problem with this is that you can't have a secure, knowledge-free cryptosystem. The short of it is that in order to securely communicate, the system has to authenticate the receiving end. This can only happen if the sender and receiver share some secret; otherwise, there is no way to prevent man-in-the-middle attacks, or even outright impersonation.
This means you cannot have a system where not even the sender knows the cert/passphrase/whatever -- because then there is fundamentally no way to distinguish between the supposed sender or receiver, and an imposter.
You know, I hear this claim pretty routinely, but I know that the best publically known factoring algorithms are superpolynomial at best. Do you know of any evidence to support the assertion that the NSA can quickly factor integers?
From time to time I have heard about cases where law enforcement tapped some poor sap's lines, and gave the encrypted stuff they found to the NSA, who forwarded back the plaintext after a couple of years. I've never heard whether these were "strong" algorithms being crypanalyzed or exploits on other parts of the cryptosystem.
If you "the NSA can factor quickly" folks have any evidence, please let the rest of the world know! After all, it's not too late to abandon RSA. Factoring may, but doesn't necessarily, solve discrete logs, so DSA/ElGamal encryption might still be safe. If not, elliptic curves or some such might work.
Slashdot Mirror
RedHat.
Okay, but do you have any particular reason to believe this, or is it just a tenet of your faith? If you consider that fuel can be made relatively cheaply from local ingredients (just react some H_2 with the atmosphere, really) and that transport time isn't important for cargo, it might not be too expensive at all. Strap a booster onto your block-o-platinum and loft into Martian orbit (low gravity, so lots easier than for Earth). Fire up an ion/magsail/Vasimir/whatever engine and two years later you're aerobreaking into Earth orbit.
By far the largest cost to mining on Mars is going to be transporting and supporting the human miners -- which, sadly, makes robots a promising alternative. It'll be interesting to see which gets there first, robots sophisticated enough for autonomous mining operations, or launch costs low enough to realistically support human extraterrestrial colonization.
First, magnetostatic attraction drops off as distance cubed.
Second, you can actually do a lot better than that, because planes conduct electricity. Thus, with appropriate dynamic fields, you can set up circulating currents in the skin or frame, which you can then couple to with other fields. We're talking big antenna arrays here, but it might just be possible.
Third, any such attempt is likely to melt the plane via inductive heating before you actually "tractor beam" it to the ground.
Finally, for the energy you're expending, a laser or rail/coil-gun setup is vastly more likely to work. And cooler, too.
Of course, that philosophy will get you in trouble with my friends in the string theory department, most of whom take the view that when you get to a deep enough level of understanding, the workings of the universe are entirely defined by the underlying mathematical structures. Their argument is that, at some level, there isn't anything left to describe, and physics reduces to a bunch of staggeringly nonlinear mathematical relationships (mostly dealing with the topology of the spacetime metric at string-theory scales).
Of course, I'm entirely unqualified to speak further on this subject, as I mostly do astrophysics myself. But hey, when you have friends who will randomly drop by to announce that they no longer believe in spacetime, and can plausibly back it up, you learn stuff. :-)
Okay, I can't resist --
In that case, said post-apocalyptic peoples probably want to know how to build a computer, too. So throw in a good book on semiconductor design. What? They don't have fab plants? Okay, toss in some materials science and electrical engineering texts, and some general engineering texts if they want to know how to build the plant itself. They'll love you forever if you leave the parts to assemble a simple mill machine in a back room.
Of course, if they're going to figure their way through semiconductors (or hell, even tube-based circuitry) they'll want a good E&M text, and probably something on quantum mechanics. Chances are, they won't have the math for this, so throw in math texts (calculus derived from number theory, basic real analysis, and abstract algebra seem like a good bare minimum set).
'Course, maybe they can't even read those! (They were *actually* bombed into the stone age? Humans are extinct, and this library has been found by the descendents of the hyper-intelligent white lab mice?) So throw in a printed copy of some of those SETI transmissions, and you're all set. Oh, don't forget some good science fiction, so they'll know why they should go to the trouble of spending three generations chewing through this library. :-)
Strictly speaking, the GPL cannot prevent you from distributing anything that you have written yourself, since you are responsible as the copyright holder for enforcing the license. However, you would need to add a special exception to the license giving others permission to link it against the MFC, or else nobody else would be allowed to distribute it.
Take a look at Galeon's license for an example of this. The Gecko library is under MPL, so they have to give special permission to link to it.
Explanation from the FSF's GPL FAQ.
This is an important point -- MS has been behaving recently like they really don't believe that there's going to be another trial, whereas that is exactly what the Appeals court has ordered: a new evidentiary hearing to determine remedies. The ruling explicitly stated that in a fast moving industry like this one, new evidence should be considered to see how circumstances have changed, to construct a remedy that actually solves the problem.
On the positive side for them is the new set of rules for what OEMs can change on systems. (Which isn't that much.) On the negative side, we have this, the problems with Kodak (can't find a link right now), their fight with AOL, and of course their much-discussed attempt to monopolize web services through .NET. Regarding this last one in particular, a new round of discovery on Microsoft internal documents could be particularly damning, since so far most of what we have to go on is rumors and leaks.
Keep an eye on the antitrust case; MS may have seriously mis-played their hand.
From American Experience: The Rockefellers (PBS):
The default clauses for software differ slightly from state to state, but are mostly pretty reasonable. You of course have the usual copyright permissions, which basically say you can't redistribute without permission, but otherwise don't much limit your use or disposal of the software. Until a few years back, there was a legal school of thought that you needed explicit permission from the copyright holder to run it, but Congress has since specified that this is not in fact true; in the absense of other agreements, you have the right to run any software you legally possess.
This, incidentally, is one reason why UCITA is so pernicious: UCITA changes many of these defaults to fairly complex and usually undesirable terms which would practically force anyone who wants to distribute software to hire a lawyer and make sure a binding license agreement is attached (and the GPL wouldn't count, because it is a licence on redistribution, which you don't have to agree to before downloading and using software).
VFS layer cleanups, USB enhancements, oh my! And the march to rock-solid stability continues.
Not that I've had much trouble with the earlier 2.4 kernels on the whole, but I wouldn't run my server farm on them. Soon, though, from the look of things.
Keep up the great work.
yes, but have you ever seen the amount of equipment needed to do this? At FermiLab, the first stage of data processing is done in the detector circuitry, and occupies a good chunk of the detector's volume (a three story high by 50 meter long piece of equipment, I should mention). Then, an entire floor of a good sized building is filled with racks of mostly custom-built circuitry processes the output of the first stage filters for interesting events.
It's even worse at CERN. They're currently putting up a new building that will be entirely filled with computing hardware to manage the data produced by the experiments when LHC comes online.
Anyway, sure, it's possible to filter that sort of data stream. But could you do it on the seabed? No. I'm not even convinced the NSA could afford many such installations. The price tag for the current incarnation of CDF (one of the primary detectors at FermiLab): around $700 million. And that's using cheap grad student labor to build a good chunk of it.
This is a common misconception. You are thinking of the characteristic length scale for strong interactions, which is inversely related to the characteristic energy scale of strong interactions. However, the gluon is massless, and thus the strong force has infinite range.
The really cool (or aggravating, depending on your field of interest) thing about the strong force is that its strength actually increases according to some ungodly exponant with increasing range (the potential is somewhat complex, so I'd be lying if I quoted an actual number). Thus, the only allowed states for objects with "color charge" turn out to be "color neutral" combinations.
One consequence of this is that you can't ever make a free quark, because the strong potential just keeps going to infinity. And if you try, you eventually get above the quark pair-creation threshold, so whatever combination of quarks needed to return everything to color neutrality just falls out of the quantum field.
No, you're conflating two different things. The SSC (Superconducting SuperCollider) was to be built in Texas; that project was halted partway through construction because the congress-critters got budget happy. The LEP (Large Electron-Positron collider) is CERN's current machine. It was recently shut down to free up the tunnel for an even larger machine called LHC (Large Hadron Collider).
Name --- Energy(TeV)
LEP --- .2 or so
FermiLab --- 1 - 2
LHC --- 7
SSC --- 40 +
As you can see, SSC was quite a loss to the physics community. Still a lot of people grouchy about that.
Poor example, I'm afraid. The "crack baby" scare is little more than a myth. A number of studies, most prominently one from last year out of the U. of Toronto, have found that "crack babies" are, in terms of intellect, statistically indistinguishable from other individuals in the same population. I have also read studies indicating that, while they start out distractable etc., they seem to grow out of it just fine when they hit adolescence.
The best set of references on this that I have offhand is the bibliography off of this set of slides. Some browsers don't like it, though.
On the first point, it's not even clear that sharing files on Gnutella or other services is illegal. Some places (Canada comes to mind) it is definitely legal; see other posts citing media taxes which are theoretically used to compensate artists for the copying. In the United States, we have the Audio Home Recording ACt, which makes it legal to copy music and share it. The catch in this case is that the law has been interpretted to apply only to devices legally recognized as "audio home recording devices" which means mix tapes are protected, but MP3s aren't.
As for the second point, as has been pointed out before, copying data is fundamentally different from stealing merchandise, in that nobody is deprived of anything (and don't tell me about lost profits -- contrary to popular belief, corporations do NOT have a "right to profit"). Nor is this a case of stores protecting their property, either -- the MPAA is harassing consumers for their post-sale activities, which may or may not be in line with the doctrine of first sale.
Anyway, even if the MPAA is legally within its rights, and the Gnutella users are legally in the wrong, it isn't clear that this situation will continue. Some interpretations of the Constitution (not the Supreme's opinion, but that could change) argue that copyright law as it currently stands violates the 1st Amendment and perhaps the Copyright Clause itself, and there ARE those in Congress who would like to fix it.
You don't consider linear cryptanalysis a threat?
It's actually better than that. Turns out, the NSA had been researching linear cryptanalysis sometime prior to 1977. So, when their designers get together with IBM's, they hand over some very particular S-boxes for the algorithm. Until the work of Shamir et al in the 90s, nobody knew where they'd pulled them from. It was one of those "We can't tell you where we got them, or why you should use them -- just trust us" deals.
So around 1995, linear cryptanalysis is discovered in the non-classified world, and applied to DES. To everyone's great surprise, when you factor in the storage requirements, linear cryptanalyzing DES is almost exactly as hard as brute forcing it, because of the particular structure of the S-boxes.
So sure, the NSA almost certainly wouldn't *tell* us if they knew how to break RSA, but their mandate *is* the security of the US. As such, they have an interest in getting security that works into the hands of Americans (business in particular, you might notice from the press release -- same reasoning behind designing DES so strong).
Interestingly, some of the more intelligent animal species exhibit the same behavior. Chimpanzees in particular, having been taught sign language, were observed to look behind themselves when "speaking" about the future.
No! You were doing great up to the comment about the Hawking radiation. As several people have pointed out, Hawking radiation for stellar black holes is in the neighborhood of a nanokelvin above absolute zero -- colder than the microwave background.
You actually can have a galactic dark halo composed of black holes, but lensing studies rule out most of the possible mass ranges. Earth-mass primordial black holes are still a candidate, but since they only form by fairly exotic processes a couple of seconds after the big bang, nobody takes them seriously.
The bit about the super-massive primordial stars is very possibly right on. An upcoming issue of Astrophysical Journal Letters is going to have an article on Pop. III massive black holes, which formed from >100-solar mass stars that formed at the center of primordial density fluctuations. This is very cool -- not only do the resultant black holes do a nice job of nucleating galaxies, but they shine brighly in the early universe before aggregating through mergers into a central supermassive black hole. Neat, yes?
Damn straight. Hawking radiation is thermal, and for a stellar black hole it is in the neighborhood of a nanokelvin above absolute zero -- so the thing is STILL colder than the microwave background.
Lensing studies have actually detected a few quiescent stellar black holes wandering free in our galaxy, relatively close by at that. Except for the lensing, we detect absolutely nothing from where the black holes should be.
In fact, the Hawking lifespan of a stellar black hole is about 10^60 years.
There might be primordial black holes evaporating now; they started out as sub-earth mass black holes. Nobody's seen this happen yet; see my earlier post.
Multiple Big Bangs are allowed -- see Big Crunch. The trouble is, no information is preserved across the singularity, so you can't have structure -- not even singularities like black holes -- come out on the other side.
Now, in principle you might be able to have a big bang spontaneously occur in our universe, creating a subuniverse -- see Multiverse. Michael Turner has been a big proponent of this lately. However, if this occurred it would either destroy our universe by inflating a huge bubble in it (so points in spacetime that were previously close together would now be separated by billions of parsecs) or you wouldn't see it happen because the new subuniverse would be hidden behind an event horizon (masquerading as an ordinary black hole) and thereby NOT distorting our metric to hell.
Either way, you don't get an "overwriting" of our universe with a new one. Unless you have a false vacuum decay or something, which is an entirely different issue, and probably a load of crap.
The trouble with implementing schemes like this is that the cryptographic protocol is almost always the strongest link in the chain. Much easier to attack other parts of the cryptosystem, such as key management, user interface, etc.
Examples: the article mentions that the decryption keys are stored in the "receiver" app. Okay -- disassemble the app, take key, you're in. Another way to do this would be to develop a patch for the binary that no-ops the syscalls that disable printing and the like, but leaves the crypto in place.
Zero-knowledge protocols are cool stuff, but they share a key weakness with many other crypto algorithms: they are deceptively strong. Thus, it is very easy to lose sight of the fact that the application is pretty much guaranteed to be the point of attack, and should be designed accordingly.
As Schneier has pointed out before, this type of scheme probably cannot be implemented -- PCs are intrinsically untrusted platforms. The application cannot be certain that it isn't running in an emulator. If it tries to "protect" the data by wiping the files, it cannot know whether the user will try again, learning from past mistakes, by restoring everything from a disk image. Is there any general way to distinguish between Word and a pipe to a file programmed to present Word's API? No.
This is actually fairly unlikely. From everything I've heard, MS actually maintains a pretty strict policy of keeping their developers very far away from GPLed code. I've even heard that sourceforge and some other sites are blocked at their firewalls, for just this reason. If nothing else, "accidentally" incorporating GPLed code into their software would seem to open them to all kinds of nasty shareholder lawsuits for negligence.
The problem with this is that you can't have a secure, knowledge-free cryptosystem. The short of it is that in order to securely communicate, the system has to authenticate the receiving end. This can only happen if the sender and receiver share some secret; otherwise, there is no way to prevent man-in-the-middle attacks, or even outright impersonation.
This means you cannot have a system where not even the sender knows the cert/passphrase/whatever -- because then there is fundamentally no way to distinguish between the supposed sender or receiver, and an imposter.
You know, I hear this claim pretty routinely, but I know that the best publically known factoring algorithms are superpolynomial at best. Do you know of any evidence to support the assertion that the NSA can quickly factor integers?
From time to time I have heard about cases where law enforcement tapped some poor sap's lines, and gave the encrypted stuff they found to the NSA, who forwarded back the plaintext after a couple of years. I've never heard whether these were "strong" algorithms being crypanalyzed or exploits on other parts of the cryptosystem.
If you "the NSA can factor quickly" folks have any evidence, please let the rest of the world know! After all, it's not too late to abandon RSA. Factoring may, but doesn't necessarily, solve discrete logs, so DSA/ElGamal encryption might still be safe. If not, elliptic curves or some such might work.