Redundant comment: TWM _was_ ugly as hell by
contemporary standards.
On the other hand, if I remember correctly,
I switched from whatever non-X windowing system Sun was using (can't remember the name) directly to X with TWM sometime in the late 80's. I hated Sun's system and, comparatively, TWM and X seemed wonderful, elegant, versatile and aesthetically
inspired.
FVWM was another huge leap forward. Oddly
enough, although I like KDE and Gnome and currently use KDE, they
haven't made as big a difference as TWM and FVWM
have to me. I can't even imagine anything as
horrible as UWM sounds; so maybe even the old
Sun's had a comparatively nice interface.
For those who want to knock Unix for horrible
interfaces and design-by-committee, remember that
most non-Unix types were using DOS at this time and later they were using pre-95 versions of Windows. DOS was (and is) a shockingly bad and difficult to use command line interface despite its simplicity and
early versions of Windows IMO claim the title for worst graphical interface of all time. VMS generally meant a command line over a text terminal.
Macs may have been ahead of most other
commercial offerings, but
Unix not being at least reasonably competitive on the user interface
front is a relatively recent thing (from 1995 on); and of course, just a few short years later, without much help from vendors, we have
competitive interfaces again. Unix hasn't done as badly in this area as many seem to think. Unix has either been ahead of MS (mid 80's to 95)
or quickly catching up.
Of course this is intended to address only the windowing system. Unix people still like apps that send most Windows people running and crying to mommy.
Indeed, although there was a lot of smoke
and mirrors about whether the P/E ratio was
really relevant in the "new economy". Hence
arguments that stock price vs. revenue _growth_ should be the proper measure rather than price vs. earnings. It sounds good on the surface,
but it still doesn't give much of a clue if the
business model is viable. IMO anyone who didn't expect the bubble to burst about a year before
it did needs to get a better source of news---one without glossy pages, banner adds or extensive coverage of movie stars. I recommend the Economist.
To be fair to tech companies, many companies have been
taking on debt, consumer confidence ain't great, consumers have been taking on massive
debt, there was a media frenzy about the "R word", etc. Not a pretty picture. Let's just say "mistakes were made".
Re:The decision is obvious, different buyers targe
on
Is It OK To Sucks?
·
· Score: 1
Correct, Nelson Mandela never said anything
of the sort. Mbeki, however, backed off quickly
enough that it was hard for me to figure out
exactly what he did say. One moment the media
were claiming he'd read on the web that HIV doesn't cause AIDS
and was basing South African health
care policy on that, and the next he was saying that HIV does cause AIDS, Africa needs anti-virals, and all he meant to imply was that there are social
factors in its spread that are uniquely African
that need to be dealt with, quite apart from the
virus. I couldn't tell if it was the media out
of control or one of the fastest spin control
jobs I'd ever seen.
FWIW, if it's the latter, I'm disappointed.
Mbeki always sounded smart to me, if not as charismatic or surrounded by an aura of
sainthood as Mandela.
Re:The decision is obvious, different buyers targe
on
Is It OK To Sucks?
·
· Score: 1
I've made plenty of my own. IMO Guinness is
about on the same level as Murphy's. Yeungling,
while a nice brewery with an interesting history
(the oldest US brewery to survive
prohibition if I remember correctly), was really
doing a better job targeting the cheaper end
of the decent beer market. I'll buy a Yeungling
and enjoy it but not at anywhere near the same price as Guinness or Murphys.
OTOH maybe I've missed
some new development from living in Australia the last few
years (I deimmigrated last week). Does Yeungling make a proper stout now
or are you talking about the porter? If they
do make something they call a stout, I'll apologize and retract this paragraph until I
get my hands on it.
Sierra Nevada is superb if you like a hoppier
stout, but it ain't the same type of beast as
Guinness. Brooklyn Black Chocolate Stout is
amazing too. Rogue makes nice stouts. Anchor
porter might well be considered a stout depending
on your standards---and it's great. The list
is endless, but there aren't many American examples that are much like the Irish variety (approx. 1040OG, solid hop bitterness, but not much in the way of finishing hops, nitrogen serve, and
a heavy does of black rather than chocolate malts). There are vaguely similar alternatives but the closest and most readily available ones are all Irish. (Murphy's and Beamish).
To steer myself back to a state of relevant
grace, I think doing harm to the company is a red herring. Companies seem to be claiming that
every use of a trademark as part of a domain name is a commercial use
that falls under trademark law. Call me Pangloss "Pig" Bodine, but, while it might be legal now, trademark protection has traditionally been
rather narrowly construed. I think
that in the best of all possible worlds it should only apply to commercial
use of a name. Hell, in the distant past, the
fact that this guy isn't selling stout would
have been enough to nullify a claim of trademark
violation.
Does anyone think if my real name is Steve
Guinness I shouldn't be able to start up The
Guinness Window company without having Bill
Gates and the shade
of Arthur Guinness on my ass? If some guy falls
out his 2nd story window, shattering his leg and rupturing
his spleen, due to my shoddy refusal
to build a window that doesn't cave in
whenever someone leans on the window frame, shouldn't he be able
to grab a GuinnessSucks domain name and give me
hell? If we let every corporation grab such
exclusive rights to their name, then I'd have
to change my name to Steve Defenestration and
start selling Defenestration brand windows. (Wait---Bill G. probably has already
protectively trademarked defenestration as
a slur on Windows.) Scratch that: how about
XTRAPRerjij123 Brand Dual Home/Outdoor Photon
Portals? Would you buy my product once you
figured out what it was?
Sorry for the rant. Reductio ad absurdum
is dangerous stuff, but tempting.
You seem to have missed the word bandlimited. Audio signals are always filtered in advance to ensure that they are bandlimited and the response of the human ear is itself bandlimited, so this isn't a big assumption. You are talking about sampling a signal with exactly one frequency component at exactly its Nyquist frequency. To recover a signal exactly you need to sample at faster than the Nyquist frequency. Samples that are only marginally faster than you propose will suffice to exactly recover the signal you are describing. A way to do this is: make an analog square wave from the digital samples and then apply an analog filter to remove harmonics and (at the same time) eliminate multiplication of the frequency components by sin(x)/x. If the signal was bandlimited and you design your filter well enough (admittedly a non-trivial design problem), this gives you the signal exactly. If you don't know what the function sin(x)/x has to do with it, you need to go back and study sampling theory before spouting off.
What you call "mathematical justification" not only explains your example, but shows that you are talking about sampling at a threshold for which for all higher sampling frequencies you can recover the bandlimited signal exactly. You've pinpointed the fastest sampling for which reconstruction can still fail (and, in fact, sampling at exactly the Nyquist frequency wouldn't have failed here if you had selected a signal with a continuous frequency distribution instead of a single frequency---this isn't the simplest case, but, in fact, it is an extreme case). Use the frequency domain. You can't get much intuition about this without looking at Fourier transforms.
I was wondering the same damn thing. Slashdot moderation is pretty useful, but on technical areas that are a little outside the mainstream here it occasionally falls flat on its face in a really spectacular way.
This isn't just sour grapes: My post wasn't all that great in terms of clarity (I shouldn't post at 2am after a night out drinking), but others made the same points better than I did and didn't get moderated up. They were both lucid and right while the highly rated post they were responding to was just plain wrong.
Sorry for the rant... For the most part I like the moderation here, but every now and then it seems more important to post early and be able to pretend you know what you are talking about than to be able to get the facts right.
Re:Mr. Chiariglione and Encryption/Watermarking
on
SDMI *NOT* Cracked!?
·
· Score: 1
The question, of course, is how big those two "ifs" actually are.
My guess is that they are pretty big "ifs". Assuming that the technology isn't patented (this is a big "if" too), someone is certain to fill a market niche by providing a player that doesn't give a rat's ass about watermarks. OTOH, I said this in another thread and someone pointed out that even on PCs, this protection is starting to be implemented in hardware: someone is making a soundcard that scans for watermarks. If you manage to make a copy and run it through an open source, watermark-ignoring player, it will still fail when pumped through your sound card. They can certainly make it inconvenient to make copies for reasons of fair use.
BTW, does anyone know if the SDMI technology, in whatever form it finally manifests itself around our throats, will be patented?
Re:Mr. Chiariglione and Encryption/Watermarking
on
SDMI *NOT* Cracked!?
·
· Score: 3
I don't know if you are on crack (you don't sound like you are, but who can tell on the internet?), but cracking an encryption scheme is a much better defined problem than removing a watermark. You have a reasonable standard to decide if you have succeeded. A watermark, on the other hand, introduces some distortion in the music and removing it is going to result in more distortion. Removing it is always possible in the trivial sense that you can write a program to take a music file and output all zeros; it will be a "distortion" of the original music with no watermark. The point of this observation: there is some ambiguity about how good the cracked music has to sound before you call it a legitimate crack. No matter what hackers come up with, these guys are going to swear with their hands on their wallets that cracked music has shitty sound quality and that their watermark doesn't harm the sound quality at all.
The guy in the link mentioned this fact. I'm betting it's something they'll hide behind, even if all their schemes are cracked.
This is either misleading or wrong depending on how generous my interpretation is. Here is the best I can do at explaining the sampling theorem.
The Nyquist sampling theorem: a signal that is bandlimited to a maximum frequency F is EXACTLY determined by samples taken at a frequency 2*F
In vague terms that might not be all that clear (I wish I could draw a picture!), the reason for this is that if you look at the the frequency components of a sampled signal they are the same as for the original signal, but replicated periodically along the frequency axis. The faster the sampling, the wider the spacing. If they are sampled fast enough then because they are all bandlimited these replications don't intersect and you can recover a pure replication of the original signal by just filtering a portion of the spectrum that includes the copy you want and excludes the others. If they intersect you have what is called aliasing, where each copy of the signal corrupts the others and none is retrievable. The Nyquist frequency is the frequency at which aliasing sets in; anything faster than 2*F and your samples will suffice to recover the bandlimited signal; anything slower than that and you are screwed.
It's actually a bit more complicated than that, but not in the way you suggest. In practice, the process for reconstruction of an analog signal from a digital signal (i.e. DAC) involves applying a zero-order-hold (i.e. producing an analog signal by holding the value of the previous digital signal until the next comes along). This gives a square-shaped signal that has the same frequency components as the original signal, and still replicated over the entire frequency spectrum, but multiplied by the sinc function sin(x)/x. You get the original signal by low-pass filtering to pick up the frequency-replicated version of the signal you want while compensating for the distortion caused by multiplication by sin(x)/x. If you can design sufficiently good low-pass filters with an x/sin(x) shape, you will still get the original signal exactly, assuming only that it was bandlimited to a maximum frequency F.
Further, there are ways to deal with both aliasing and the restriction that you have a good low-pass filter at the end. For the first, you just low-pass filter the original analog signal before sampling. This ensures that there is no aliasing and that you, at least in theory, have enough information to reconstruct all frequency components less than F exactly. For the second, you use oversampling purely in the digital domain. The idea for oversampling is that if the signal really was limited to a maximum frequency of F and your sampling rate is 2*F, the signal is completely determined by its samples and you can accurately interpolate what the samples would have been if you had sampled at 4*F, 8*F, 16*F. This process is exact and (as I mentioned before about the effect of a higher sampling rate on the spacing of replications of the spectrum) it creates a significant gap between the replicated versions of the signal. This gap means that the cut-off on your low-pass filter doesn't have to be exact---it can gradually (rather than sharply) drop off in the regions in which the oversampled signal has no frequency components. Note that oversampling is done through digital interpolation of a signal sampled at a lower rate and has nothing to do with the sampling rate on a CD (or whatever medium). In fact the chief purpose of oversampling is not to improve sound quality, but to allow people to use crappier analog filters for the DAC process.
It's late here in Australia. That's long-winded and probably not clear to anyone who doesn't know it, but the gist of it is what I put in bold above: 2*F is sufficient to recover the signal exactly if you want to put enough care into designing the hardware. Contrary to what you say, Nyquists theorem states that for a band-limited signal, the shape of the wave form CAN be accurate if you put enough care into designing the hardware.
FWIW, none of this is obvious. My jaw dropped the first time I saw an explanation of the sampling theorem. It's really an amazing, counter-intuitive idea.
I may be wrong here and I'm sure I'm dating myself back into the 6502 and Z-80 era, but wasn't BallBlaster a Lucas game? Simple game: a ball, a court, and two "pods" that could push the ball around and shoot it between two goal posts to try to score. It had a split screen (one for each player) and a rudimentary first-person 3-d view of the court that was pretty impressive for the technology of the time. And if you had enough people to compete against, it was addictive as hell. I still fondly remember matches in my dorm room.
At the time, I was studying electrical engineering and trying to write games on the side. My most notable (Heh!) accomplishment was "BrickShitter", a version of Space War superimposed on the cellular automata program Life: It was hand coded in 6502 assembly for an old Atari 8-bit; the object was the same as space war, but you could "shit bricks" (i.e. lay down blocks on the screen) which would form cells and evolve according to the rules of "Life". The idea was to create obstacles that could protect you from your opponents shots and, to make things more complicated, would vary with time. I debugged it and got it working, but the playability was shocking. After all that work for an ill-conceived idea, I have a lot of respect for people who can think up playable games. I was just trying to learn how to write something relatively big in 6502 assembly and didn't spend a lot of time thinking the concept over.
\begin{geezer}Ah! It takes me back. I still think the old games were more fun, even if my attempts sucked. Anyone else remember typing in the code for games from computer magazines? Those Data statements in Basic giving Hex code for a game coded in Assembly were hell.\end{geezer}
Well, I mostly agree. I'll try to balance both sides, as I see them. IMO the most substantial concern about Redhat dominating the market is a question of when a company strong-arms the standards to suit their own purposes: we don't want Netscape playing continual catch-up to a standard set by MS (FWIW, I hate to use Crashscape and, while I use Linux for all my serious work, I'm reading this in IE; having your favorite browser run on your least favorite OS really sucks). IMO diversity is a way out of the fundamental problem of companies wanting to capture their market by not supporting open standards: if no distribution is dominant, they all have to work together to respect common standards or have their user base so fragmented that no one will want to use Linux at all. Redhat's only crime is being succesful when free software types are (with reason) suspicious of success; grabbing too large a piece of the pie always has the potential for undermining standards.
That being said, I really admire the way Redhat has respected the idea of an open source distribution and the way they have contributed to open source/linux related projects. As much as I trust any company to not behave like bastards, I trust them. My history with Linux has been Slackware in 1995 with a 1.2.13 kernel, Redhat 3.x shortly thereafter and, over the last year (I went through a few years without a PC, only using Sparc workstations), Mandrake 6.5. I only switched to Mandrake after trying Redhat 6.0 and finding Gnome 1.0 to be dubiously stable and the KDE setup in Redhat to be weak. I don't stay on the bleeding edge, so it might be a while, but I'll probably try Redhat again next time I upgrade because I respect what they are trying to contribute (even Gnome, although I haven't tried it since 1.0). No one is perfect, but Bob Young is right; comparing Redhat to Microsoft is ridiculous---they're just unusually succesful good guys who occasionally make disputable calls on when to release things. I'm less certain of the viability of open source business models, but if Bob Young wants to justify Redhat being a nice company by arguing that he would be cutting off his nose to spite his face by giving up on open source, I'm not going argue with him.
The idea assumes that they can keep some control of most players/recorders and make sure they respect the watermarks. Every piece of music is supposed to have two watermarks, one fragile and one indestructible. The indestructible one is supposed to survive compression into other formats and indicates that the piece of music is copyrighted. The fragile one disappears under compression and indicates that the music is a copy and is in the wrong format or has been temporarily transmitted in the wrong format. If only one watermark is there, the music shouldn't play on any player made by any company playing along with the SDMI. Thus it doesn't protect against exact copies (if you have a recorder that will make exact copies of anything watermarked...), but it is intended to protect against putting the music into other formats for transmission, archival, or use of another player on another platform. No watermarks at all means that you can play and copy, but the copy will have two watermarks put on. (Thus you can't make a copy of a copy that will play on commercial players, even if it's a recording from your own garage band.)
This is summarized from the Oct. 7-13 edition of the Economist; I didn't really understand the point of watermarking before reading that either.
Personally, in my ignorance, I never thought this was a problem for hackers wanting to play music on their PCs. I don't know for sure about commercial players on stereo systems, but there's no way that all players for PCs will respect the watermark. And I'd be surprised if some enterprising company didn't end up making a player/recorder for your stereo that ignored the watermarks as well. (Do all DVD players respect the region codes? There's at least a comfortable niche market for products that don't respect these security standards.)
Yeah. The guy who, earlier in this discussion, referred to number 5 as almost a one-time pad (i.e. provably secure encryption) had it right. At least some of us know algorithms for factoring, even if we don't have the computer power to do it and possibly don't know the best algorithms; trying to find the right text to decode #5 must have been hellish for everyone who tried, whether they had a network of computers or an Atari 400 (my first computer at 1.79 MHz...). If, as would be reasonable for anyone trying to REALLY encrypt something, Singh had selected a random sequence of letters instead, the code would have been unbreakable.
I found the trouble people had with #5 illuminating: public key cryptography isn't everything. If you can distribute your key secretly, or not distribute it at all, symmetric cryptography can be pretty powerful. I guess that's part of why public key methods are chiefly used to distribute keys for DES and the like. (The other reason being computational complexity---would anyone use RSA to encrypt a reasonably long message?)
FWIW I've been really impressed with Singh as a science writer who tries to get it right. There aren't that many of them (and one of my hobbies is poking holes in popular accounts of science and mathematics.)
I'm not an expert here, but aren't all NP-complete problems reducible to a satisfiability problem? I'm less sure with this, but I thought that the satisfiability problem was also, in fact, reducible to any one of these problems. This would be sufficient to show that all NP-complete problems are isomorphic under a polynomial time reduction. And I've heard people say that a solution to any one NP-complete problem yields a solution to all the others. That would also seem to suggest isomorphism.
Re:This is an incorrect definition of NP
on
Does P = NP?
·
· Score: 2
Call me stupid, but can someone explain how a Turing machine can possibly be non-deterministic?
The nondeterministic Turing machine is a separate creature and, to the best of my knowledge, I don't know that Turing ever thought about it. (Though I wouldn't swear that he didn't---he was an amazing man). It's an abstract generalization of a Turing machine. The guy who said that you should imagine a computer in which you can use an arbitrary number of fork()'s which produce completely parallel processes, any one of which might solve the problem, had the right idea. Perhaps even more roughly, think of a parallel computer that always has as many nodes as you need to create any new process you might want to run on a new node without conflicting with other nodes in any way.
It's a rather abstract model that isn't equivalent to anything we have built. I suppose if we ever build good enough parallel computers, it might be taken as a model for massively parallel computation---but I've never seen anyone pushing this line. The number of nodes would need to grow exponentially with the problem size and this degree of parallelism for problems we might be interested in is probably fairy dust unless DNA or quantum based computers work out. (And no one has proven that either of these amount to general non-determinstic Turing machines anyway, even if they can be made to work at all). So for the moment it's an abstract model of computation that lets us consider whether we can reduce a large number of exponential time problems that are polynomial time on the abstract model to polynomial time on a real computer. In other words, don't hold your breath waiting for a non-deterministic P9 from Intel.
Warning: I'm not an expert on this field, but I am a mathematician and I did study this stuff in a basic way long, long ago. Apologies if I've mangled some facts.
This is related to the current nobel prize for chemistry. From watching an interview with the recipient on the News Hour with Jim Lehrer, apparently whoever won the prize (the name escapes me...) did so for work on conducting polymers. Apparently they dope a polymer with something (he mentioned iodine, though I don't know if that is what is actually used in any real devices) and it automagically conducts. He mentioned plastic batteries and plastic light emitting diodes as two of the main applications.
It'd be nice to see a major leap forward in battery technology. I hope this pans out.
So, is this a non-story, or are TW _really_ up to no good? Because if this is true, it smacks of strongarming ISPs into becoming part of the AOL/TW collective.
It defies belief. If they were really doing this, they'd have the FTC and the DOJ on them like stink on shit. It sounds like a blatant abuse of monopoly power to me; surely they can't be that stupid?
Those elegant, sturdy, indestructable IBM keyboards that you could spill coffee on and they would still work
They're still there. I'm typing on one now. Sure it's missing the Escape key cap and a few function keys, but all you have to do is go to a junkyard and select one out of the scrap heap for $1.50; they clean up well and they'll last forever. It's one stop shopping: I also got an alternator for my car while I was there. I just wish they had put the control key in the right place. (For youngsters and PC people, that's to the left of the "A", where "Caps Lock" is now).
Just out of curiosity, there seems to be an assumption here that inbreeding results in the propagation of bad recessive traits. I made the assumption in an earlier post. But the Hardy-Weinberg law states that the relative proportion of dominant and recessive genes remains constant if breeding is random. Obviously breeding isn't random (Slashdot is 90%+ male so I'll ask the question: Swimsuit model vs. Janet Reno... You choose...), but what I've read about sexual selection suggests that it would favor beneficial genes. So can a species be coddled for a while in an artificial environment, letting individuals choose to mate with the members of the population with good genes(or even letting humans make the call), to eliminate bad recessive genes? How do they make the call as to what constitutes a viable breeding population?
Food for thought. Anyway, it's food for thought for me; I don't know the answer.
That's an excellent point and I was going to make it before I saw that you had already done so. A genetically viable breeding population requires more individuals than most people might think. Adam and Eve or Noah's Ark aren't viable options. Two is not enough and several hundred are a bit dicey.
I also wanted to point out that this doesn't solve the problem of habitat decline, which is one of the biggest threats facing endangered species in the first place. If you don't have a plan for a sustainable environment in which the species can
live, that doesn't cut it either. A male and a female in a zoo aren't of much use except for providing the opportunity of stunning photography of an extinct species.
D'oh! Thanks. I know about protective oxide layers with steel and I should have been able to figure that out. If I were a moderator, I'd try to bump you up to 5 for concisely answering a basic question about something relevant to the discussion. But instead I'll just sit here feeling stupid...
Disclaimer for the marginal relevance of this post: I don't know whether or not Malaysian arcades are suitable for children or whether they should be banned. Different culture, different values; if they think they need to ban arcades because of gambling or the fact that there are unsavoury characters, I won't give them a hard time.
But the prime minister of Malaysia, Mahathir, is a nasty piece of work and the current system there is at best quasi-democratic, without any of the political freedoms most of us are used to. (Just read up on the politically inspired and politically manipulated sodomy trial of his former lieutentant Anwar Ibrahim.) Mahathir also has a habit of ranting about international conspiracies of first world governments and currency speculators to keep Asia under the heel of the west. I'm not that impressed with currency speculators either, but Mahathir is a shrill, rabid, xenophobic nationalist who holds on to power by any means necessary and blames others for his own economic mistakes.
The irony I see in the Malaysian govt. accusing kids of stealing that inspired my subject header: in 1997 I was a post-doc at Stanford. Mahathir was visiting Stanford at the time and, in particular, was visiting the Bill Gates building (I know, this is Slashdot---I spat on the sidewalk every single time I entered). He sent some of his security people out to find some reading material. They came into my office and asked around. Wanting to express good will to other nations and promote peace on earth, I offered my New York Times, which I was going to read on my lunch break, on the condition that it be returned to me when Mahathir was finished. I never saw it again and, to this day, I don't know what happened in the world on January xx 1997. I know Mahathir didn't know the conditions under which I lent him the paper and I should blame the security guys, but I don't like him and as far as I'm concerned Mahathir pinched my paper and doesn't have any moral right to crack down on any sort of crime. I'd much rather have a gay friend (i.e. Anwar) than a friend who pinches stuff (i.e. Mahathir).
Sorry for the rant. I don't like his politics and he has personally annoyed me. I was looking for an excuse to go off. If it had been anyone other than a prime minister I would have forgotten about it long ago.
I don't disagree in general, but it must rust under some circumstances or given enough time. My chemistry is pretty rusty (Hahahaha!) and I'm not sure I ever understood anything about metals but I'm a religous reader of the Economist and the whole point of the article was that you get titanium dioxide and the new electrolytic process for getting pure titanium is more efficient than the old chemical reduction method. So under what circumstances does titanium rust (i.e. titanium dioxide) form in the first place?
I realize this is a tangent, but I'm honestly curious to hear an understandable reconciliation of the anti-rusting properties of titanium with the prevalence of titanium dioxide. Maybe I'm just trying to make up for not paying enough attention to chemistry courses when I was an undergraduate...
The problem is that without common carrier protection for ISPs you could sue other people's ISPs for delivering content. If you send an offensive E-mail or set up an offensive web site and your ISP could be held liable and sued by whomever you sent the E-mail to or by the hot-tempered and protective parent whose kid visited your porn/bomb making site. That's what people are worried about.
I guess there is an up-side here in avoiding some regulation, but opening up the possibility that ISPs could be sued for the actions of their users strikes me as a bigger down-side.
IANAL, but it seems to me that we are already paying for it everytime we buy a recording device or medium. Check out the US code Title 17 Sec 1001 and following at US Copyright Code. Manufacturers and importers seem to be paying into a royalty fund for every recording device or medium they provide. Presumably this is to cover the cost to artists of the noncommercial copying that, to my untrained eye, appears to be explicitly permitted in Sec. 1008.
Given that copyrights are constitutionally limited rights, noncommercial copying is permitted in a passage in the US code and there is a royalty fund for compensating artists, words like "steal" seem unfair and inflammatory. I realize you might be a troll, but it would be more constructive to debate whether the current provisions of the law are adequate to compensate artists for noncommercial copying.
Note that none of my comments should be taken as being in defence of for-profit distribution of copyrighted works. This is illegal and IMO should be illegal.
Slashdot Mirror
Redundant comment: TWM _was_ ugly as hell by contemporary standards. On the other hand, if I remember correctly, I switched from whatever non-X windowing system Sun was using (can't remember the name) directly to X with TWM sometime in the late 80's. I hated Sun's system and, comparatively, TWM and X seemed wonderful, elegant, versatile and aesthetically inspired. FVWM was another huge leap forward. Oddly enough, although I like KDE and Gnome and currently use KDE, they haven't made as big a difference as TWM and FVWM have to me. I can't even imagine anything as horrible as UWM sounds; so maybe even the old Sun's had a comparatively nice interface.
For those who want to knock Unix for horrible interfaces and design-by-committee, remember that most non-Unix types were using DOS at this time and later they were using pre-95 versions of Windows. DOS was (and is) a shockingly bad and difficult to use command line interface despite its simplicity and early versions of Windows IMO claim the title for worst graphical interface of all time. VMS generally meant a command line over a text terminal.
Macs may have been ahead of most other commercial offerings, but Unix not being at least reasonably competitive on the user interface front is a relatively recent thing (from 1995 on); and of course, just a few short years later, without much help from vendors, we have competitive interfaces again. Unix hasn't done as badly in this area as many seem to think. Unix has either been ahead of MS (mid 80's to 95) or quickly catching up.
Of course this is intended to address only the windowing system. Unix people still like apps that send most Windows people running and crying to mommy.
Indeed, although there was a lot of smoke and mirrors about whether the P/E ratio was really relevant in the "new economy". Hence arguments that stock price vs. revenue _growth_ should be the proper measure rather than price vs. earnings. It sounds good on the surface, but it still doesn't give much of a clue if the business model is viable. IMO anyone who didn't expect the bubble to burst about a year before it did needs to get a better source of news---one without glossy pages, banner adds or extensive coverage of movie stars. I recommend the Economist.
To be fair to tech companies, many companies have been taking on debt, consumer confidence ain't great, consumers have been taking on massive debt, there was a media frenzy about the "R word", etc. Not a pretty picture. Let's just say "mistakes were made".
Correct, Nelson Mandela never said anything of the sort. Mbeki, however, backed off quickly enough that it was hard for me to figure out exactly what he did say. One moment the media were claiming he'd read on the web that HIV doesn't cause AIDS and was basing South African health care policy on that, and the next he was saying that HIV does cause AIDS, Africa needs anti-virals, and all he meant to imply was that there are social factors in its spread that are uniquely African that need to be dealt with, quite apart from the virus. I couldn't tell if it was the media out of control or one of the fastest spin control jobs I'd ever seen.
FWIW, if it's the latter, I'm disappointed. Mbeki always sounded smart to me, if not as charismatic or surrounded by an aura of sainthood as Mandela.
I've made plenty of my own. IMO Guinness is about on the same level as Murphy's. Yeungling, while a nice brewery with an interesting history (the oldest US brewery to survive prohibition if I remember correctly), was really doing a better job targeting the cheaper end of the decent beer market. I'll buy a Yeungling and enjoy it but not at anywhere near the same price as Guinness or Murphys. OTOH maybe I've missed some new development from living in Australia the last few years (I deimmigrated last week). Does Yeungling make a proper stout now or are you talking about the porter? If they do make something they call a stout, I'll apologize and retract this paragraph until I get my hands on it.
Sierra Nevada is superb if you like a hoppier stout, but it ain't the same type of beast as Guinness. Brooklyn Black Chocolate Stout is amazing too. Rogue makes nice stouts. Anchor porter might well be considered a stout depending on your standards---and it's great. The list is endless, but there aren't many American examples that are much like the Irish variety (approx. 1040OG, solid hop bitterness, but not much in the way of finishing hops, nitrogen serve, and a heavy does of black rather than chocolate malts). There are vaguely similar alternatives but the closest and most readily available ones are all Irish. (Murphy's and Beamish).
To steer myself back to a state of relevant grace, I think doing harm to the company is a red herring. Companies seem to be claiming that every use of a trademark as part of a domain name is a commercial use that falls under trademark law. Call me Pangloss "Pig" Bodine, but, while it might be legal now, trademark protection has traditionally been rather narrowly construed. I think that in the best of all possible worlds it should only apply to commercial use of a name. Hell, in the distant past, the fact that this guy isn't selling stout would have been enough to nullify a claim of trademark violation.
Does anyone think if my real name is Steve Guinness I shouldn't be able to start up The Guinness Window company without having Bill Gates and the shade of Arthur Guinness on my ass? If some guy falls out his 2nd story window, shattering his leg and rupturing his spleen, due to my shoddy refusal to build a window that doesn't cave in whenever someone leans on the window frame, shouldn't he be able to grab a GuinnessSucks domain name and give me hell? If we let every corporation grab such exclusive rights to their name, then I'd have to change my name to Steve Defenestration and start selling Defenestration brand windows. (Wait---Bill G. probably has already protectively trademarked defenestration as a slur on Windows.) Scratch that: how about XTRAPRerjij123 Brand Dual Home/Outdoor Photon Portals? Would you buy my product once you figured out what it was?
Sorry for the rant. Reductio ad absurdum is dangerous stuff, but tempting.
You seem to have missed the word bandlimited. Audio signals are always filtered in advance to ensure that they are bandlimited and the response of the human ear is itself bandlimited, so this isn't a big assumption. You are talking about sampling a signal with exactly one frequency component at exactly its Nyquist frequency. To recover a signal exactly you need to sample at faster than the Nyquist frequency. Samples that are only marginally faster than you propose will suffice to exactly recover the signal you are describing. A way to do this is: make an analog square wave from the digital samples and then apply an analog filter to remove harmonics and (at the same time) eliminate multiplication of the frequency components by sin(x)/x. If the signal was bandlimited and you design your filter well enough (admittedly a non-trivial design problem), this gives you the signal exactly. If you don't know what the function sin(x)/x has to do with it, you need to go back and study sampling theory before spouting off.
What you call "mathematical justification" not only explains your example, but shows that you are talking about sampling at a threshold for which for all higher sampling frequencies you can recover the bandlimited signal exactly. You've pinpointed the fastest sampling for which reconstruction can still fail (and, in fact, sampling at exactly the Nyquist frequency wouldn't have failed here if you had selected a signal with a continuous frequency distribution instead of a single frequency---this isn't the simplest case, but, in fact, it is an extreme case). Use the frequency domain. You can't get much intuition about this without looking at Fourier transforms.
I was wondering the same damn thing. Slashdot moderation is pretty useful, but on technical areas that are a little outside the mainstream here it occasionally falls flat on its face in a really spectacular way. This isn't just sour grapes: My post wasn't all that great in terms of clarity (I shouldn't post at 2am after a night out drinking), but others made the same points better than I did and didn't get moderated up. They were both lucid and right while the highly rated post they were responding to was just plain wrong.
Sorry for the rant... For the most part I like the moderation here, but every now and then it seems more important to post early and be able to pretend you know what you are talking about than to be able to get the facts right.
My guess is that they are pretty big "ifs". Assuming that the technology isn't patented (this is a big "if" too), someone is certain to fill a market niche by providing a player that doesn't give a rat's ass about watermarks. OTOH, I said this in another thread and someone pointed out that even on PCs, this protection is starting to be implemented in hardware: someone is making a soundcard that scans for watermarks. If you manage to make a copy and run it through an open source, watermark-ignoring player, it will still fail when pumped through your sound card. They can certainly make it inconvenient to make copies for reasons of fair use.
BTW, does anyone know if the SDMI technology, in whatever form it finally manifests itself around our throats, will be patented?
I don't know if you are on crack (you don't sound like you are, but who can tell on the internet?), but cracking an encryption scheme is a much better defined problem than removing a watermark. You have a reasonable standard to decide if you have succeeded. A watermark, on the other hand, introduces some distortion in the music and removing it is going to result in more distortion. Removing it is always possible in the trivial sense that you can write a program to take a music file and output all zeros; it will be a "distortion" of the original music with no watermark. The point of this observation: there is some ambiguity about how good the cracked music has to sound before you call it a legitimate crack. No matter what hackers come up with, these guys are going to swear with their hands on their wallets that cracked music has shitty sound quality and that their watermark doesn't harm the sound quality at all.
The guy in the link mentioned this fact. I'm betting it's something they'll hide behind, even if all their schemes are cracked.
This is either misleading or wrong depending on how generous my interpretation is. Here is the best I can do at explaining the sampling theorem.
The Nyquist sampling theorem: a signal that is bandlimited to a maximum frequency F is EXACTLY determined by samples taken at a frequency 2*F
In vague terms that might not be all that clear (I wish I could draw a picture!), the reason for this is that if you look at the the frequency components of a sampled signal they are the same as for the original signal, but replicated periodically along the frequency axis. The faster the sampling, the wider the spacing. If they are sampled fast enough then because they are all bandlimited these replications don't intersect and you can recover a pure replication of the original signal by just filtering a portion of the spectrum that includes the copy you want and excludes the others. If they intersect you have what is called aliasing, where each copy of the signal corrupts the others and none is retrievable. The Nyquist frequency is the frequency at which aliasing sets in; anything faster than 2*F and your samples will suffice to recover the bandlimited signal; anything slower than that and you are screwed.
It's actually a bit more complicated than that, but not in the way you suggest. In practice, the process for reconstruction of an analog signal from a digital signal (i.e. DAC) involves applying a zero-order-hold (i.e. producing an analog signal by holding the value of the previous digital signal until the next comes along). This gives a square-shaped signal that has the same frequency components as the original signal, and still replicated over the entire frequency spectrum, but multiplied by the sinc function sin(x)/x. You get the original signal by low-pass filtering to pick up the frequency-replicated version of the signal you want while compensating for the distortion caused by multiplication by sin(x)/x. If you can design sufficiently good low-pass filters with an x/sin(x) shape, you will still get the original signal exactly, assuming only that it was bandlimited to a maximum frequency F.
Further, there are ways to deal with both aliasing and the restriction that you have a good low-pass filter at the end. For the first, you just low-pass filter the original analog signal before sampling. This ensures that there is no aliasing and that you, at least in theory, have enough information to reconstruct all frequency components less than F exactly. For the second, you use oversampling purely in the digital domain. The idea for oversampling is that if the signal really was limited to a maximum frequency of F and your sampling rate is 2*F, the signal is completely determined by its samples and you can accurately interpolate what the samples would have been if you had sampled at 4*F, 8*F, 16*F. This process is exact and (as I mentioned before about the effect of a higher sampling rate on the spacing of replications of the spectrum) it creates a significant gap between the replicated versions of the signal. This gap means that the cut-off on your low-pass filter doesn't have to be exact---it can gradually (rather than sharply) drop off in the regions in which the oversampled signal has no frequency components. Note that oversampling is done through digital interpolation of a signal sampled at a lower rate and has nothing to do with the sampling rate on a CD (or whatever medium). In fact the chief purpose of oversampling is not to improve sound quality, but to allow people to use crappier analog filters for the DAC process.
It's late here in Australia. That's long-winded and probably not clear to anyone who doesn't know it, but the gist of it is what I put in bold above: 2*F is sufficient to recover the signal exactly if you want to put enough care into designing the hardware. Contrary to what you say, Nyquists theorem states that for a band-limited signal, the shape of the wave form CAN be accurate if you put enough care into designing the hardware.
FWIW, none of this is obvious. My jaw dropped the first time I saw an explanation of the sampling theorem. It's really an amazing, counter-intuitive idea.
I may be wrong here and I'm sure I'm dating myself back into the 6502 and Z-80 era, but wasn't BallBlaster a Lucas game? Simple game: a ball, a court, and two "pods" that could push the ball around and shoot it between two goal posts to try to score. It had a split screen (one for each player) and a rudimentary first-person 3-d view of the court that was pretty impressive for the technology of the time. And if you had enough people to compete against, it was addictive as hell. I still fondly remember matches in my dorm room.
At the time, I was studying electrical engineering and trying to write games on the side. My most notable (Heh!) accomplishment was "BrickShitter", a version of Space War superimposed on the cellular automata program Life: It was hand coded in 6502 assembly for an old Atari 8-bit; the object was the same as space war, but you could "shit bricks" (i.e. lay down blocks on the screen) which would form cells and evolve according to the rules of "Life". The idea was to create obstacles that could protect you from your opponents shots and, to make things more complicated, would vary with time. I debugged it and got it working, but the playability was shocking. After all that work for an ill-conceived idea, I have a lot of respect for people who can think up playable games. I was just trying to learn how to write something relatively big in 6502 assembly and didn't spend a lot of time thinking the concept over.
\begin{geezer}Ah! It takes me back. I still think the old games were more fun, even if my attempts sucked. Anyone else remember typing in the code for games from computer magazines? Those Data statements in Basic giving Hex code for a game coded in Assembly were hell.\end{geezer}
Well, I mostly agree. I'll try to balance both sides, as I see them. IMO the most substantial concern about Redhat dominating the market is a question of when a company strong-arms the standards to suit their own purposes: we don't want Netscape playing continual catch-up to a standard set by MS (FWIW, I hate to use Crashscape and, while I use Linux for all my serious work, I'm reading this in IE; having your favorite browser run on your least favorite OS really sucks). IMO diversity is a way out of the fundamental problem of companies wanting to capture their market by not supporting open standards: if no distribution is dominant, they all have to work together to respect common standards or have their user base so fragmented that no one will want to use Linux at all. Redhat's only crime is being succesful when free software types are (with reason) suspicious of success; grabbing too large a piece of the pie always has the potential for undermining standards.
That being said, I really admire the way Redhat has respected the idea of an open source distribution and the way they have contributed to open source/linux related projects. As much as I trust any company to not behave like bastards, I trust them. My history with Linux has been Slackware in 1995 with a 1.2.13 kernel, Redhat 3.x shortly thereafter and, over the last year (I went through a few years without a PC, only using Sparc workstations), Mandrake 6.5. I only switched to Mandrake after trying Redhat 6.0 and finding Gnome 1.0 to be dubiously stable and the KDE setup in Redhat to be weak. I don't stay on the bleeding edge, so it might be a while, but I'll probably try Redhat again next time I upgrade because I respect what they are trying to contribute (even Gnome, although I haven't tried it since 1.0). No one is perfect, but Bob Young is right; comparing Redhat to Microsoft is ridiculous---they're just unusually succesful good guys who occasionally make disputable calls on when to release things. I'm less certain of the viability of open source business models, but if Bob Young wants to justify Redhat being a nice company by arguing that he would be cutting off his nose to spite his face by giving up on open source, I'm not going argue with him.
The idea assumes that they can keep some control of most players/recorders and make sure they respect the watermarks. Every piece of music is supposed to have two watermarks, one fragile and one indestructible. The indestructible one is supposed to survive compression into other formats and indicates that the piece of music is copyrighted. The fragile one disappears under compression and indicates that the music is a copy and is in the wrong format or has been temporarily transmitted in the wrong format. If only one watermark is there, the music shouldn't play on any player made by any company playing along with the SDMI. Thus it doesn't protect against exact copies (if you have a recorder that will make exact copies of anything watermarked...), but it is intended to protect against putting the music into other formats for transmission, archival, or use of another player on another platform. No watermarks at all means that you can play and copy, but the copy will have two watermarks put on. (Thus you can't make a copy of a copy that will play on commercial players, even if it's a recording from your own garage band.)
This is summarized from the Oct. 7-13 edition of the Economist; I didn't really understand the point of watermarking before reading that either.
Personally, in my ignorance, I never thought this was a problem for hackers wanting to play music on their PCs. I don't know for sure about commercial players on stereo systems, but there's no way that all players for PCs will respect the watermark. And I'd be surprised if some enterprising company didn't end up making a player/recorder for your stereo that ignored the watermarks as well. (Do all DVD players respect the region codes? There's at least a comfortable niche market for products that don't respect these security standards.)
Yeah. The guy who, earlier in this discussion, referred to number 5 as almost a one-time pad (i.e. provably secure encryption) had it right. At least some of us know algorithms for factoring, even if we don't have the computer power to do it and possibly don't know the best algorithms; trying to find the right text to decode #5 must have been hellish for everyone who tried, whether they had a network of computers or an Atari 400 (my first computer at 1.79 MHz...). If, as would be reasonable for anyone trying to REALLY encrypt something, Singh had selected a random sequence of letters instead, the code would have been unbreakable.
I found the trouble people had with #5 illuminating: public key cryptography isn't everything. If you can distribute your key secretly, or not distribute it at all, symmetric cryptography can be pretty powerful. I guess that's part of why public key methods are chiefly used to distribute keys for DES and the like. (The other reason being computational complexity---would anyone use RSA to encrypt a reasonably long message?)
FWIW I've been really impressed with Singh as a science writer who tries to get it right. There aren't that many of them (and one of my hobbies is poking holes in popular accounts of science and mathematics.)
I'm not an expert here, but aren't all NP-complete problems reducible to a satisfiability problem? I'm less sure with this, but I thought that the satisfiability problem was also, in fact, reducible to any one of these problems. This would be sufficient to show that all NP-complete problems are isomorphic under a polynomial time reduction. And I've heard people say that a solution to any one NP-complete problem yields a solution to all the others. That would also seem to suggest isomorphism.
The nondeterministic Turing machine is a separate creature and, to the best of my knowledge, I don't know that Turing ever thought about it. (Though I wouldn't swear that he didn't---he was an amazing man). It's an abstract generalization of a Turing machine. The guy who said that you should imagine a computer in which you can use an arbitrary number of fork()'s which produce completely parallel processes, any one of which might solve the problem, had the right idea. Perhaps even more roughly, think of a parallel computer that always has as many nodes as you need to create any new process you might want to run on a new node without conflicting with other nodes in any way.
It's a rather abstract model that isn't equivalent to anything we have built. I suppose if we ever build good enough parallel computers, it might be taken as a model for massively parallel computation---but I've never seen anyone pushing this line. The number of nodes would need to grow exponentially with the problem size and this degree of parallelism for problems we might be interested in is probably fairy dust unless DNA or quantum based computers work out. (And no one has proven that either of these amount to general non-determinstic Turing machines anyway, even if they can be made to work at all). So for the moment it's an abstract model of computation that lets us consider whether we can reduce a large number of exponential time problems that are polynomial time on the abstract model to polynomial time on a real computer. In other words, don't hold your breath waiting for a non-deterministic P9 from Intel.
Warning: I'm not an expert on this field, but I am a mathematician and I did study this stuff in a basic way long, long ago. Apologies if I've mangled some facts.
This is related to the current nobel prize for chemistry. From watching an interview with the recipient on the News Hour with Jim Lehrer, apparently whoever won the prize (the name escapes me...) did so for work on conducting polymers. Apparently they dope a polymer with something (he mentioned iodine, though I don't know if that is what is actually used in any real devices) and it automagically conducts. He mentioned plastic batteries and plastic light emitting diodes as two of the main applications.
It'd be nice to see a major leap forward in battery technology. I hope this pans out.
It defies belief. If they were really doing this, they'd have the FTC and the DOJ on them like stink on shit. It sounds like a blatant abuse of monopoly power to me; surely they can't be that stupid?
They're still there. I'm typing on one now. Sure it's missing the Escape key cap and a few function keys, but all you have to do is go to a junkyard and select one out of the scrap heap for $1.50; they clean up well and they'll last forever. It's one stop shopping: I also got an alternator for my car while I was there. I just wish they had put the control key in the right place. (For youngsters and PC people, that's to the left of the "A", where "Caps Lock" is now).
Just out of curiosity, there seems to be an assumption here that inbreeding results in the propagation of bad recessive traits. I made the assumption in an earlier post. But the Hardy-Weinberg law states that the relative proportion of dominant and recessive genes remains constant if breeding is random. Obviously breeding isn't random (Slashdot is 90%+ male so I'll ask the question: Swimsuit model vs. Janet Reno... You choose...), but what I've read about sexual selection suggests that it would favor beneficial genes. So can a species be coddled for a while in an artificial environment, letting individuals choose to mate with the members of the population with good genes(or even letting humans make the call), to eliminate bad recessive genes? How do they make the call as to what constitutes a viable breeding population?
Food for thought. Anyway, it's food for thought for me; I don't know the answer.
That's an excellent point and I was going to make it before I saw that you had already done so. A genetically viable breeding population requires more individuals than most people might think. Adam and Eve or Noah's Ark aren't viable options. Two is not enough and several hundred are a bit dicey.
I also wanted to point out that this doesn't solve the problem of habitat decline, which is one of the biggest threats facing endangered species in the first place. If you don't have a plan for a sustainable environment in which the species can live, that doesn't cut it either. A male and a female in a zoo aren't of much use except for providing the opportunity of stunning photography of an extinct species.
D'oh! Thanks. I know about protective oxide layers with steel and I should have been able to figure that out. If I were a moderator, I'd try to bump you up to 5 for concisely answering a basic question about something relevant to the discussion. But instead I'll just sit here feeling stupid...
Disclaimer for the marginal relevance of this post: I don't know whether or not Malaysian arcades are suitable for children or whether they should be banned. Different culture, different values; if they think they need to ban arcades because of gambling or the fact that there are unsavoury characters, I won't give them a hard time.
But the prime minister of Malaysia, Mahathir, is a nasty piece of work and the current system there is at best quasi-democratic, without any of the political freedoms most of us are used to. (Just read up on the politically inspired and politically manipulated sodomy trial of his former lieutentant Anwar Ibrahim.) Mahathir also has a habit of ranting about international conspiracies of first world governments and currency speculators to keep Asia under the heel of the west. I'm not that impressed with currency speculators either, but Mahathir is a shrill, rabid, xenophobic nationalist who holds on to power by any means necessary and blames others for his own economic mistakes.
The irony I see in the Malaysian govt. accusing kids of stealing that inspired my subject header: in 1997 I was a post-doc at Stanford. Mahathir was visiting Stanford at the time and, in particular, was visiting the Bill Gates building (I know, this is Slashdot---I spat on the sidewalk every single time I entered). He sent some of his security people out to find some reading material. They came into my office and asked around. Wanting to express good will to other nations and promote peace on earth, I offered my New York Times, which I was going to read on my lunch break, on the condition that it be returned to me when Mahathir was finished. I never saw it again and, to this day, I don't know what happened in the world on January xx 1997. I know Mahathir didn't know the conditions under which I lent him the paper and I should blame the security guys, but I don't like him and as far as I'm concerned Mahathir pinched my paper and doesn't have any moral right to crack down on any sort of crime. I'd much rather have a gay friend (i.e. Anwar) than a friend who pinches stuff (i.e. Mahathir).
Sorry for the rant. I don't like his politics and he has personally annoyed me. I was looking for an excuse to go off. If it had been anyone other than a prime minister I would have forgotten about it long ago.
I don't disagree in general, but it must rust under some circumstances or given enough time. My chemistry is pretty rusty (Hahahaha!) and I'm not sure I ever understood anything about metals but I'm a religous reader of the Economist and the whole point of the article was that you get titanium dioxide and the new electrolytic process for getting pure titanium is more efficient than the old chemical reduction method. So under what circumstances does titanium rust (i.e. titanium dioxide) form in the first place?
I realize this is a tangent, but I'm honestly curious to hear an understandable reconciliation of the anti-rusting properties of titanium with the prevalence of titanium dioxide. Maybe I'm just trying to make up for not paying enough attention to chemistry courses when I was an undergraduate...
The problem is that without common carrier protection for ISPs you could sue other people's ISPs for delivering content. If you send an offensive E-mail or set up an offensive web site and your ISP could be held liable and sued by whomever you sent the E-mail to or by the hot-tempered and protective parent whose kid visited your porn/bomb making site. That's what people are worried about.
I guess there is an up-side here in avoiding some regulation, but opening up the possibility that ISPs could be sued for the actions of their users strikes me as a bigger down-side.
IANAL, but it seems to me that we are already paying for it everytime we buy a recording device or medium. Check out the US code Title 17 Sec 1001 and following at US Copyright Code. Manufacturers and importers seem to be paying into a royalty fund for every recording device or medium they provide. Presumably this is to cover the cost to artists of the noncommercial copying that, to my untrained eye, appears to be explicitly permitted in Sec. 1008.
Given that copyrights are constitutionally limited rights, noncommercial copying is permitted in a passage in the US code and there is a royalty fund for compensating artists, words like "steal" seem unfair and inflammatory. I realize you might be a troll, but it would be more constructive to debate whether the current provisions of the law are adequate to compensate artists for noncommercial copying.
Note that none of my comments should be taken as being in defence of for-profit distribution of copyrighted works. This is illegal and IMO should be illegal.