If you're sending it into space, weevils and damp really aren't problems you have to deal with, so paper would be much more practical than here on Earth.:)
Re:Updates Anyone?
on
Jaguar is Over
·
· Score: 2, Insightful
Yes, but Apple seems to EOL products in 12-18 months, as compared to 5 years for Microsoft (YES, Microsoft is the benchmark here.)
If 10.1 were still a supported product then people wouldn't complain so much about the forced upgrades. They could still happily run 10.1 and be reasonably confident that their machine was secure and stable. Instead, people must pay $130 to upgrade to Jaguar if they want a secure OS.
My Windows 98 machine at home STILL gets security updates from Microsoft. I would expect 10.1 to still be getting updates until at least 10.3 is released, but it hasn't had one since mid-2002.
I agree. Problem is, they put a touchscreen on the back of the device. This nullifies most of the advantages of the piezoelectric interface.
If the ONLY (primary) interface to the device is piezoelectric, then piezoelectric can make a lot of sense. However, their sample application already included a touch screen, so the piezoelectric component adds cost without adding a durability benefit.
What I didn't get from this picture was one simple thing:
If you can put a touch-sensitive area on the back, why not just have touch-sensitive areas for zoom in and zoom out?
Something like: Tap behind an area to zoom in on it, drag your finger to scroll, and tap a corner to zoom out would be every bit as effective. No, it's not a button "per se" but a touch screen is WAY more sophisticated than a button, and can emulate one quite reasonably. Adding bend detection is just going to increase manufacturing costs for no significant usability benefit.
Now, I can think of a few applications where this would be useful (i.e. consider a color or brightness adjustable glowstick) but only on objects that require only one or two independent input variables, not for something so sophisticated that it's already going to have an integrated touchscreen.
For simply indexing files on a network and not getting paid for it, contributory infringement is pretty clearly not applicable, and vicarious infringement is a little questionable since it's not clear that they had the responsibility to police their list. The students could probably claim the DMCA safe-harbor anyways, so long as the RIAA never contacted them telling them what they had that needed to be removed.
If they were sharing and/or downloading songs themselves, then that clearly IS illegal, and this is probably pretty reasonable for a "show everybody we mean business" lawsuit. It's high enough to mean something, but low enough to keep them from declaring bankruptcy and to keep it from really affecting their futures (assuming they have enough money to finish college.)
Unless (as noted in my response to notasheep, as well as by one of the supreme court justices in the Eldred case) the only reason content producers produce is because they believe that they will be able to extend copyright terms before their copyright expires.
In that case the right thing to do is of course to extend copyright terms and then lock them down. If there is a noticeable drop in content production for an extended period, then you can consider extending the term (since it isn't long enough,) but for new works only.
You made a subtle distinction there that's important. I didn't say the extensions themselves are detrimental, but the whole system where copyrights are extendable, for the correct fee paid to the correct legislators.
Lobbying congress to extend copyright terms costs money, and lots of it. Unfortunately, under the current system lobbying is a profitable investment, but it doesn't create any new content (besides a few poorly thought-out laws.) If terms were set in stone and not open to debate, then some (I dare say most) of the money spent lobbying congress to extend copyrights on existing works would instead be spent on creating new works since the new works would fall under copyright.
Many if not most of the holders of existing works are also content producers. Certainly this applies to the movie and record industries (although somewhat indirectly.) These are also the industries most strongly lobbying for extended copyright terms. Preventing the retroactive extension of copyrights would encourage these industries to constantly produce new content, in order to make up for the loss of profit from works put in the public domain.
What we're doing here is removing a substitute from one market (Disney, as and example, investing in new laws as a substitute for new works) and adding a substitute to another (Wal-mart's version of Snow White after a fixed number of years as a substitute for Disney's now.) Thus, we are encouraging Disney to invest in content production instead of lobbying or reproduction of older content by making that lobbying and older content reproduction less profitable to invest in than new content production. This was the whole point of American copyright, to make production of new works more profitable than other endeavours, not to make reproduction of existing works profitable.
Now, the only good counterargument I can come up with is that if copyrights really still ARE too SHORT, and the only reason content producers are producing works is that they believe they will continue to be able to extend copyright. If you believe this, then it makes sense to continue retroactively extending copyrights up to a point, as new content producers will just stop creating if we stop extending copyrights, as it won't be profitable. However, the correct solution to this problem is to extend copyrights sufficiently now, and then lock things down.
Actually, it's not even the length of copyrights that causes the lack of new content, it's the continual extensions, and it's a system that benefits politicians as much as the content industries.
Look at it like this: You have 2 choices of investments.
Investment 1: For a few million dollars, you can make an investment that has consistently paid off, and which if successful will continue producing significant profit for the next 20 years.
Investment 2: For that same million dollars, you can invest in a work that has proven to be profitable only around 70 percent of the time, and the term of those profits is unclear. There's maybe a 50 percent chance that it will produce profits for more than 2 years, and a 0.5 percent chance that it will be profitable for up to 100 years.
Most organizations would choose investment 1, although the really rich ones would hedge their bets and invest in both.
Well, what are these investments?
1.) Investing in your congressman to encourage them to extend copyrights on the existing works that you already make profits on.
2.) Investing in a new movie.
For the congressman, this is a good investment also. For the cost of a few votes, you get a LOT of campaign funds from the companies, which typically translates into significantly more votes than you lost, and then into getting to keep your job.
If existing copyrights were really for a fixed, non-negotiable time period, ANY fixed time period, the result would be that the money spent on extending copyrights on existing works would instead be spent on creating new works, to replace those that are falling into the public domain.
For any reasonably long copyright term there is no significant economic incentive to investing in extending the term for new copyrights because that money would be better invested in creating new works. With an unreasonably long term it may be true that inappropriate amounts of money is spent maintaining old works, but it is unclear that this money is badly spent, or that this money would not be spent by someone even if the works were in the public domain, since maintaining old works does have value to society as a whole. Certainly some money goes to lawyers to protect additional works under a longer copyright term, but again it isn't clear that this wouldn't be spent with a shorter copyright term, just to protect the works that remain under copyright.
Extendable copyright terms are the real problem preventing the creation of new works under the current copyright system.
[NOTE: The numbers for investment 2 are guesses, but I suspect I'm being generous with some of them. Also, I intentionally didn't mention the DMCA, etc. Those are a whole new type of copyright law, and here I'm talking about traditional copyright.]
The judge makes it very clear in the ruling that he is making no decisions on people who operate an index of copyrighted files. In fact, that seems like a reasonably clear case of contributory infringement. Writing a program that indexes copyrighted files is okay, but the moment you provide that index to others you have to take steps to restrict entries with copyrighted content. At a minimum, you may be able to claim DMCA safe harbor if you eliminate entries after the copyright holder tells you those entries are infringing.
Actually, it's worse than that. The decision centered on two things. First, neither of these companies is currently operating a Kazaa "root supernode," and each of them lacks control over either the underlying protocol(Morpheus/Gnutella) or the software itself (Grokster). Things look pretty bad for Kazaa, since it has control over all of the items listed above, and it presumably operates Kazaa "root supernodes." I suspect that they are liable, at least for vicarious infringement, based on the operation of root supernodes.
The judge declines judgement on people who provide (for instance) lists of gnutella nodes. These people may still be liable for contributory and/or vicarious infringement.
Of course, the users actually doing the copying are direct infringers, so they should not expect protection by the courts.
I agree with you about being able to remove the INVARIANT sections as being fundamental to really having free documentation. I also agree that having the ability to lock down portions of the document from change (not removal) has good uses to prevent people from modifying your political statements or personal views.
However, I don't think it's unreasonable for the DFSG to prevent these personal views from being included in main. By making INVARIANT sections removable, Debian would be able to just remove the INVARIANT sections and leave the rest of the documentation intact. No problems there. The author doesn't have their views misrepresented, and the consistency of Debian's modifiability is maintained.
The only thing that really calls for a non-removable invariant section is attribution of the author. This should be included in a small invariant area in the license, along with reasonable terms for combination of multiple works with attribution clauses. If someone abuses the attribution clause, their work will simply be moved to non-free, just as Debian is discussing doing with current documents containing an invariant section.
Trouble is, even the original author can't remove an invariant section. The moment the original author picks up changes from a second contributor, they are now under license from that second contributor never to remove the invariant sections that they themselves put in.
This is why invariant sections are such a problem... they're PERMANENT.
What's wrong with requiring an always-up internet connection at each poll-site?
You simply have the person make a request to vote in that location, and you check the database. Once you have established the person has not previously voted, you mark them as "voted" and give them a randomly numbered ballot.
If the database connection is lost, you failsafe to voter-id marked ballots, as you say.
There are various ways to make the failsafe ballots less onerous (like maintaining separate databases, as was suggested in another post.) In this case you would need two databases, 1 with a list of voters who cast failsafe ballots, a second, heavily restricted one which links voter names in database 1 with randomly selected voter IDs, and finally the votes themselves, which are marked with the randomly selected voter ID.
There are various other schemes of keeping the link between voter ID and voter as secret as possible, but when it comes down to it, they all involve trusting the government not to use a database or to destroy all copies of a database. If you don't trust the government to be able to handle voter data provided they have reasonable safeguards, what makes you think they won't just subvert the system at some other level? They're the ones running the election, aren't they? How hard is it to void the votes for an entire precinct simply by illegally opening the ballot box, and do so in a county that is sure to support your opponent?
But in DRM, the secret isn't being sent to YOU, it's being sent to YOUR COMPUTER. It's that whole trusted client thing. Software can probably never be a trusted client BY ITSELF (too many opportunities for corrupting the computer it's running on), but software in conjunction with a TCPA-like device gets pretty close.
I agree, you can't keep a secret from yourself (i.e. the music has to be output to you, and the video has to be displayed on your monitor. That secret you can always get, and many consider it the weak point in DRM), but this isn't the secret most DRM schemes seem designed to keep.
I.E. you have three layers on any given piece of data, for instance an E-book.
1.) The text of the book itself. This must be displayed on the user's screen, thus allowing it to be copied by at least some sort of device that the attacker attaches to the monitor... or by simple retyping.
2.) The unencrypted file containing the text of the book itself. This is what DRM is trying to keep from you. So long as you have no direct access to this file, it's much more difficult to copy, requiring either retyping the text or using a device as described above.
3.) The encrypted copy of the unencrypted file. This doesn't need to be protected, and since it can only be opened by trusted clients, it's relatively safe to let go in the wild.
So, DRM can be successful at protecting the unencrypted file, just not at protecting the human-readable data in that file.
Something like OCRing a screenshot lies in the fuzzy area between 1 and 2, but a good trusted client can keep screenshots from the user, so you have to assume that the only interfaces for the attacker to use to generate his file are the interfaces provided to the user (i.e. the keyboard, mouse, speaker output, and display output.)
Now, there's another problem with DRM. If the trusted client also refuses to play untrusted CONTENT, then the data that the attacker generates from the screen reader will be unusable on all trusted clients. Such a client isn't too difficult to imagine, as video game consoles already do this. In a hypothetical future where most people don't have access to a "general-purpose" computer, DRM could work. Sure, you could still make casette tapes of music coming out of your computer speakers and you could still make photocopies of text scraped from your monitor, but that kind of copying isn't what the big content producers are worried about.
I admit that the above scenario doesn't seem likely since at least one computer manufacturer (Apple) has made home content production a major part of their product marketing. Still, to say that all DRM schemes are nonsense is not looking at the concept closely enough.
I would love to use Pine, but both it and MUTT don't seem to handle unread pointers in the same way the other clients (i.e. Mozilla, OE, Kmail) do, so my read/unread pointers on my IMAP server don't get tracked correctly. Maybe it's a setup issue.
I would use them for the bulk of my mail if it weren't for this limitation, but since I like to use the GUI mail clients sometimes, neither of these works.
Actually, conceptually this could be used as a good thing. An "evil bit" that meant "drop this packet" could be used as a safety on network security testing. If I want to simulate an attack on machine A, I tell machine A to ignore the "evil bit" while machine B will continue to honor it. I don't have to worry about machine B doing anything based on my test packets, since it will drop them on the floor based on the "evil bit."
It could also be used to test firewalls for any possible damage done by malicious packets hitting the firewall and being dropped, in this case due to the "evil bit."
From what I could tell from reading those papers, softupdates can't recover if the corruption occurs while writing a sector to the drive. With that kind of error a journal is necessary to be able to recover. The other solution would be to always rewrite the superblock for every update and maintain several copies of the superblock (effectively journalling only the superblock.) This is essentially what TUX2 does.
I could be mistaken, but my understanding is that softupdates may allow you to update an on-disk inode or directory structure that has other things pointing to it, which could corrupt a filesystem if the system crashed during the update.
Let me get this straight. First you say that the average person is too dumb to be trusted with a vote. And then you say that our representatives should be selected at random from that same mass of people who are too stupid to vote in a general election, and given a vote in a legislative assembly? Are you sure about that? You'd be giving huge amounts of power to people who are, by your reasoning, incapable of wielding even the lesser power of a vote diluted by thousands of others. Perhaps you should put that one back to bake for a bit more, 'cause it ain't done yet.
Your political system evidently would be controlled by a small minority -- the "intelligent" ones who can "do the right thing" for all the countless others. You say you don't want to be lorded over by an aristocrat; but what I hear is not that you hate aristocrats . . . merely that you want to be the one doing the lording
I agree that most people are too stupid to vote for political candidates. I'm not even sure that I'm smart enough to judge a person's character well enough to vote for a person. However, I think most people ARE smart enough to vote on specific issues. Voting on specific issues is much easier. That's why selecting people at random from the general populace (NOT some ethereal bourgiouse, which is just plain elitism) would actually work. Sort of like jury duty, but within the legislative branch instead of the judicial. Executive branch will probably still have to be filled through elections or legislative selection, but this is generally the least powerful branch.
There are some problems in the system, but those can be worked out. You need enough random senators so that no single interest is likely to be overrepresented. Terms will have to be long enough so that people don't push through half-baked laws that only mildly represent their interests before they leave office. This can be helped by tying "senators" to specific bills or classes of law instead of time periods, etc. The executive branch will probably still have to be either elected or appointed, but their power is already constrained by the legislative branch, and could be further constrained.
The biggest problem is who oversees the selection of random people. I'm not sure that there is a clear way to do this, as my experience with jury duty shows that current methods don't work, and result in the same person getting selected more often than is reasonable.
All of this is distracting the point that selecting random members of the populace is a reasonable solution to the problem of people being too stupid to vote in elections, as voting on issues is easier than voting for people.
Actually, it's not that complex. IP ids are a strictly one-way phoenomenon. If none of the packets going to the NAT are fragmented (common for most home NATs), all you have to do is set the IPid field to something independent of the contents of the incoming packet (say, a random number generator or even a counter) and you're done. If you do receive fragmented packets, it's a little more complex because you have to have a way to translate from the incoming packets to outgoing ones. The easiest way to do this is:
1.) On bootup choose a large random key. 2.) When you get a packet, do a secure hash on the incoming IP, the IPid field on the incoming packet, and the key, and store the result in the IPid field of the outgoing packet. 3.) Rekey as needed. Choosing a sufficiently large random key should reduce the need to rekey, and, if possible, rekeying should be done during times of network inactivity.
I don't think either Stevens or Breyer would agree with your assessment of their arguments. Both of them (though Stevens moreso than Breyer) consider abiding by the constitution as their primary purpose.
Breyer's argument is the weaker of the two because he gives the most strength to the "promote the progress of science" part of the statute. He argues that even the extension on copyrights for new works fails the constitutional requirements for "limited times" and promotion of the "progress of science." He makes a convincing argument that the courts have an obligation to set limits on what "limited times" means, and then spends a lot of time arguing that life of the author plus 70 years is so long as to fail the "limited times" requirement. He makes a reasonable argument that the courts do have the obligation to set limits (they're the ones who have to strike down a law that extends copyright to life of the author plus 10,000 years,) but fails to fully convince at least me that the current extension is so grossly out of balance that it crosses the line between being merely poorly conceived to being unconstitutional.
To understand Stevens' argument (which is substantially more convincing thand Breyer's) it is necessary to understand the majority opinion. The Ginsburg opinion seems to rely primarily on early patent cases and the existence of previous copyright term extensions by congress to decide that the framers did not intend the reading of the copyright clause that Eldred requests. Their argument primarily rests on three things: the Copyright Act of 1790, which established copyrights in the United States, a number of individual patent extensions passed between 1790 and 1875, and the Copyright Act of 1831, which was the first extension of copyright terms on existing works. From these, as well as continued congressional practice in extending copyrights on existing works, the court concludes that in both the framers' and in the modern legal framework, the CTEA is constitutional.
Stevens argues that the constitution, not early congressional actions must be the basis of our law, and that many of the actions that the majority uses to support retroactive extensions either don't apply or are blatantly unconstitutional. The copyright act of 1790, he argues, does not apply in this case because, while it did give copyright protection to existing works, did so in the context of establishing a national system of copyrights, and the founders were keenly aware of the difference between this establishment of copyrights and the extension of existing ones. Many of the patent extensions used as evidence of the framers intent were blatently unconstitutional extensions of patents on inventions that had already entered the public domain, and so are unconvincing as a basis for modern case law. The copyright extension act of 1831 cannot be used to derive the framers' intent because none of the original delgates were in the 1831 congress. Further, the 1831 act was based on a view of copyright judged unconstitutional in the 1834 case of Wheaton v. Peters. All of this goes to show that the historical precedent for constitutionality of the extension of copyrights is inconclusive at best.
Stevens further points out that protection against ex post facto laws should protect the interests of both the patentee and the public with respect to copyrights. Just as congress should be unable to shorten the term of existing copyrights (thus harming the patentee), they should also be unable to extend the term of existing copyrights (thus harming the public.)
The point of all this is to show that both Stevens and Breyer very much had the constitution in mind when forming their opinions, and they are based on reasonable interpretation of the text. Stevens makes it very clear that he considers congressional practice an inappropriate way of deciding constitutionality and is not uncomfortable with the possibility of this decision putting previous copyright term extensions on similar shaky ground, if they are indeed unconstitutional. Breyer's arguments are less clear on this, and he seems to go out of his way to show how the copyright term extinsions in 1976 could have served a constitutional purpose, while the current statute does not. This interchange makes me wonder whether some of the other justices' opinions were based primarily on a desire not to unravel 170 years worth of copyright term extension acts.
It's somewhere in-between. You get the "anything can happen in any order" that you would get with a modern processor, but cache synchronization bugs can be hidden. This is because the two threads share a common cache, and so there is minimal latency between when one thread does a cache write and when the other thread has that data in its cache, hiding bugs in the sychronization code that prevents using stale cache data.
Most likely you'd catch any bugs that were in a higher-level language just fine, but you might miss bugs at the assembly level. For instance, you'd catch bugs in the code that uses a mutex, but might not catch bugs in the code that implements said mutex.
I bet a LOT more companies would be using Linux/OpenOffice. It's kind of like the issue of oil vs. alternative fuel. We
would all like to be able to use alternatively powered vehicles, but right now Oil is just cheaper b/c there is an existing infrastructure (gas pumps everyhwere). Economically, Microsoft stuff is cheaper than
Linux, because of the advantages gained by the fact that everyone uses it. When that changes, Linux will win. One can already see some effects of Linux's success: Microsoft is moving to more and more open
standards, IBM has embraced Linux, and there are fewer and fewer reasons not to use linux.
Ahhh, but this is another example of monopoly maintenance, and the entire purpose of monopoly maintenance on the Windows product is to prevent other OSes (like GNU/Linux) from being able to become dominant.
If all Microsoft did was stay the leading OS product, adding bugfixes, upgrading the File System, etc., this would be acceptable. I'll admit that there are tendencies towards a natural monopoly in the Operating Systems business. The problem is that Microsoft goes beyond this. They integrate Windows-only variations on existing or developing middleware technologies and sell them at little to no cost in order to preserve their monopoly in the OS market.
I didn't understand exactly how this worked until I read the original verdict in the Microsoft trial. Multiplatform middleware techologies are a threat to Microsoft's Windows monopoly because they allow users to switch to other operating systems and still achieve the same functionality they have in Windows. Thus, Microsoft must replace multiplatform middleware technologies with Windows-only equivalents.
This is where Microsoft's unique position as the monopoly seller of operating systems comes in. By integrating Windows-only middleware technology into the the base Windows product, Microsoft can offer that middleware at a cost below what any competitor can possibly offer (i.e. free, and with no installation costs.) Since Microsoft Windows is the monopoly OS, content producers are virtually guaranteed to produce content for the Windows-only technology, since it will be installed on almost every computer in their potential market. The multiplatform technology thus falls by the wayside, and becomes an expensive afterthought for both consumers and content producers. In this way Microsoft harms the competitive abilities of the middleware companies and more importantly of alternative OSes that run that Middleware.
PDF is one of the remaining big middleware technology threats, and so Microsoft is coming up with a Windows-only version. By integrating it into the OS (I guarantee you, the reader for this format will be a free download and will be included with the next version of Windows) Microsoft will guarantee that content providers will produce content for this format, at the expense of PDF and more importantly at the expense of other OSes that rely on PDF for document exchange.
Are consumers or content producers dumb for choosing to use the free Microsoft product instead of the multiplatform alternative? No. At worst they are a little shortsighted, but even that is questionable. Most consumers can't conceive of using another OS besides Windows. It does most of what they want, and there are no reasonable alternatives that provide the same capabilities at the same complete system cost. GNU/Linux could conceivably get there, but it isn't there yet, and Microsoft can keep it that way by preventing the development of popular multiplatform middleware that would allow consumers to switch (even if they wanted to, which they don't.) The government had the opportunity to try and prevent further monopoly maintenance in this fashion, but it chose not to, and as the parent poster noted, gave Microsoft carte blanche to integrate whatever Windows-only technologies it wants into the OS in order to prevent the development of middleware that would lead to competition in the OS market.
How is all this bad for consumers? I can think of two reasons. First, it allows Microsoft to extract monopoly profits from its OS products, thus costing consumers money. Second, it prevents the innovation that real market competition usually leads to. There are probably others, but both of these are real consumer harms that could be fixed by restoring competition in the OS market.
Slashdot Mirror
If you're sending it into space, weevils and damp really aren't problems you have to deal with, so paper would be much more practical than here on Earth. :)
Yes, but Apple seems to EOL products in 12-18 months, as compared to 5 years for Microsoft (YES, Microsoft is the benchmark here.)
If 10.1 were still a supported product then people wouldn't complain so much about the forced upgrades. They could still happily run 10.1 and be reasonably confident that their machine was secure and stable. Instead, people must pay $130 to upgrade to Jaguar if they want a secure OS.
My Windows 98 machine at home STILL gets security updates from Microsoft. I would expect 10.1 to still be getting updates until at least 10.3 is released, but it hasn't had one since mid-2002.
I agree. Problem is, they put a touchscreen on the back of the device. This nullifies most of the advantages of the piezoelectric interface.
If the ONLY (primary) interface to the device is piezoelectric, then piezoelectric can make a lot of sense. However, their sample application already included a touch screen, so the piezoelectric component adds cost without adding a durability benefit.
What I didn't get from this picture was one simple thing:
If you can put a touch-sensitive area on the back, why not just have touch-sensitive areas for zoom in and zoom out?
Something like: Tap behind an area to zoom in on it, drag your finger to scroll, and tap a corner to zoom out would be every bit as effective. No, it's not a button "per se" but a touch screen is WAY more sophisticated than a button, and can emulate one quite reasonably. Adding bend detection is just going to increase manufacturing costs for no significant usability benefit.
Now, I can think of a few applications where this would be useful (i.e. consider a color or brightness adjustable glowstick) but only on objects that require only one or two independent input variables, not for something so sophisticated that it's already going to have an integrated touchscreen.
No, 64-bit PPC instructions are still 32-bits long. It's just the data accessed by the instructions is 64-bits wide.
What exactly were they guilty of?
For simply indexing files on a network and not getting paid for it, contributory infringement is pretty clearly not applicable, and vicarious infringement is a little questionable since it's not clear that they had the responsibility to police their list. The students could probably claim the DMCA safe-harbor anyways, so long as the RIAA never contacted them telling them what they had that needed to be removed.
If they were sharing and/or downloading songs themselves, then that clearly IS illegal, and this is probably pretty reasonable for a "show everybody we mean business" lawsuit. It's high enough to mean something, but low enough to keep them from declaring bankruptcy and to keep it from really affecting their futures (assuming they have enough money to finish college.)
Unless (as noted in my response to notasheep, as well as by one of the supreme court justices in the Eldred case) the only reason content producers produce is because they believe that they will be able to extend copyright terms before their copyright expires.
In that case the right thing to do is of course to extend copyright terms and then lock them down. If there is a noticeable drop in content production for an extended period, then you can consider extending the term (since it isn't long enough,) but for new works only.
You made a subtle distinction there that's important. I didn't say the extensions themselves are detrimental, but the whole system where copyrights are extendable, for the correct fee paid to the correct legislators.
Lobbying congress to extend copyright terms costs money, and lots of it. Unfortunately, under the current system lobbying is a profitable investment, but it doesn't create any new content (besides a few poorly thought-out laws.) If terms were set in stone and not open to debate, then some (I dare say most) of the money spent lobbying congress to extend copyrights on existing works would instead be spent on creating new works since the new works would fall under copyright.
Many if not most of the holders of existing works are also content producers. Certainly this applies to the movie and record industries (although somewhat indirectly.) These are also the industries most strongly lobbying for extended copyright terms. Preventing the retroactive extension of copyrights would encourage these industries to constantly produce new content, in order to make up for the loss of profit from works put in the public domain.
What we're doing here is removing a substitute from one market (Disney, as and example, investing in new laws as a substitute for new works) and adding a substitute to another (Wal-mart's version of Snow White after a fixed number of years as a substitute for Disney's now.) Thus, we are encouraging Disney to invest in content production instead of lobbying or reproduction of older content by making that lobbying and older content reproduction less profitable to invest in than new content production. This was the whole point of American copyright, to make production of new works more profitable than other endeavours, not to make reproduction of existing works profitable.
Now, the only good counterargument I can come up with is that if copyrights really still ARE too SHORT, and the only reason content producers are producing works is that they believe they will continue to be able to extend copyright. If you believe this, then it makes sense to continue retroactively extending copyrights up to a point, as new content producers will just stop creating if we stop extending copyrights, as it won't be profitable. However, the correct solution to this problem is to extend copyrights sufficiently now, and then lock things down.
Actually, it's not even the length of copyrights that causes the lack of new content, it's the continual extensions, and it's a system that benefits politicians as much as the content industries.
Look at it like this:
You have 2 choices of investments.
Investment 1: For a few million dollars, you can make an investment that has consistently paid off, and which if successful will continue producing significant profit for the next 20 years.
Investment 2: For that same million dollars, you can invest in a work that has proven to be profitable only around 70 percent of the time, and the term of those profits is unclear. There's maybe a 50 percent chance that it will produce profits for more than 2 years, and a 0.5 percent chance that it will be profitable for up to 100 years.
Most organizations would choose investment 1, although the really rich ones would hedge their bets and invest in both.
Well, what are these investments?
1.) Investing in your congressman to encourage them to extend copyrights on the existing works that you already make profits on.
2.) Investing in a new movie.
For the congressman, this is a good investment also. For the cost of a few votes, you get a LOT of campaign funds from the companies, which typically translates into significantly more votes than you lost, and then into getting to keep your job.
If existing copyrights were really for a fixed, non-negotiable time period, ANY fixed time period, the result would be that the money spent on extending copyrights on existing works would instead be spent on creating new works, to replace those that are falling into the public domain.
For any reasonably long copyright term there is no significant economic incentive to investing in extending the term for new copyrights because that money would be better invested in creating new works. With an unreasonably long term it may be true that inappropriate amounts of money is spent maintaining old works, but it is unclear that this money is badly spent, or that this money would not be spent by someone even if the works were in the public domain, since maintaining old works does have value to society as a whole. Certainly some money goes to lawyers to protect additional works under a longer copyright term, but again it isn't clear that this wouldn't be spent with a shorter copyright term, just to protect the works that remain under copyright.
Extendable copyright terms are the real problem preventing the creation of new works under the current copyright system.
[NOTE: The numbers for investment 2 are guesses, but I suspect I'm being generous with some of them. Also, I intentionally didn't mention the DMCA, etc. Those are a whole new type of copyright law, and here I'm talking about traditional copyright.]
Unfortunately, no.
The judge makes it very clear in the ruling that he is making no decisions on people who operate an index of copyrighted files. In fact, that seems like a reasonably clear case of contributory infringement. Writing a program that indexes copyrighted files is okay, but the moment you provide that index to others you have to take steps to restrict entries with copyrighted content. At a minimum, you may be able to claim DMCA safe harbor if you eliminate entries after the copyright holder tells you those entries are infringing.
Actually, it's worse than that. The decision centered on two things. First, neither of these companies is currently operating a Kazaa "root supernode," and each of them lacks control over either the underlying protocol(Morpheus/Gnutella) or the software itself (Grokster). Things look pretty bad for Kazaa, since it has control over all of the items listed above, and it presumably operates Kazaa "root supernodes." I suspect that they are liable, at least for vicarious infringement, based on the operation of root supernodes.
The judge declines judgement on people who provide (for instance) lists of gnutella nodes. These people may still be liable for contributory and/or vicarious infringement.
Of course, the users actually doing the copying are direct infringers, so they should not expect protection by the courts.
I agree with you about being able to remove the INVARIANT sections as being fundamental to really having free documentation. I also agree that having the ability to lock down portions of the document from change (not removal) has good uses to prevent people from modifying your political statements or personal views.
However, I don't think it's unreasonable for the DFSG to prevent these personal views from being included in main. By making INVARIANT sections removable, Debian would be able to just remove the INVARIANT sections and leave the rest of the documentation intact. No problems there. The author doesn't have their views misrepresented, and the consistency of Debian's modifiability is maintained.
The only thing that really calls for a non-removable invariant section is attribution of the author. This should be included in a small invariant area in the license, along with reasonable terms for combination of multiple works with attribution clauses. If someone abuses the attribution clause, their work will simply be moved to non-free, just as Debian is discussing doing with current documents containing an invariant section.
Trouble is, even the original author can't remove an invariant section. The moment the original author picks up changes from a second contributor, they are now under license from that second contributor never to remove the invariant sections that they themselves put in.
This is why invariant sections are such a problem... they're PERMANENT.
What's wrong with requiring an always-up internet connection at each poll-site?
You simply have the person make a request to vote in that location, and you check the database. Once you have established the person has not previously voted, you mark them as "voted" and give them a randomly numbered ballot.
If the database connection is lost, you failsafe to voter-id marked ballots, as you say.
There are various ways to make the failsafe ballots less onerous (like maintaining separate databases, as was suggested in another post.) In this case you would need two databases, 1 with a list of voters who cast failsafe ballots, a second, heavily restricted one which links voter names in database 1 with randomly selected voter IDs, and finally the votes themselves, which are marked with the randomly selected voter ID.
There are various other schemes of keeping the link between voter ID and voter as secret as possible, but when it comes down to it, they all involve trusting the government not to use a database or to destroy all copies of a database. If you don't trust the government to be able to handle voter data provided they have reasonable safeguards, what makes you think they won't just subvert the system at some other level? They're the ones running the election, aren't they? How hard is it to void the votes for an entire precinct simply by illegally opening the ballot box, and do so in a county that is sure to support your opponent?
But in DRM, the secret isn't being sent to YOU, it's being sent to YOUR COMPUTER. It's that whole trusted client thing. Software can probably never be a trusted client BY ITSELF (too many opportunities for corrupting the computer it's running on), but software in conjunction with a TCPA-like device gets pretty close.
I agree, you can't keep a secret from yourself (i.e. the music has to be output to you, and the video has to be displayed on your monitor. That secret you can always get, and many consider it the weak point in DRM), but this isn't the secret most DRM schemes seem designed to keep.
I.E. you have three layers on any given piece of data, for instance an E-book.
1.) The text of the book itself.
This must be displayed on the user's screen, thus allowing it to be copied by at least some sort of device that the attacker attaches to the monitor... or by simple retyping.
2.) The unencrypted file containing the text of the book itself.
This is what DRM is trying to keep from you. So long as you have no direct access to this file, it's much more difficult to copy, requiring either retyping the text or using a device as described above.
3.) The encrypted copy of the unencrypted file. This doesn't need to be protected, and since it can only be opened by trusted clients, it's relatively safe to let go in the wild.
So, DRM can be successful at protecting the unencrypted file, just not at protecting the human-readable data in that file.
Something like OCRing a screenshot lies in the fuzzy area between 1 and 2, but a good trusted client can keep screenshots from the user, so you have to assume that the only interfaces for the attacker to use to generate his file are the interfaces provided to the user (i.e. the keyboard, mouse, speaker output, and display output.)
Now, there's another problem with DRM. If the trusted client also refuses to play untrusted CONTENT, then the data that the attacker generates from the screen reader will be unusable on all trusted clients. Such a client isn't too difficult to imagine, as video game consoles already do this. In a hypothetical future where most people don't have access to a "general-purpose" computer, DRM could work. Sure, you could still make casette tapes of music coming out of your computer speakers and you could still make photocopies of text scraped from your monitor, but that kind of copying isn't what the big content producers are worried about.
I admit that the above scenario doesn't seem likely since at least one computer manufacturer (Apple) has made home content production a major part of their product marketing. Still, to say that all DRM schemes are nonsense is not looking at the concept closely enough.
I would love to use Pine, but both it and MUTT don't seem to handle unread pointers in the same way the other clients (i.e. Mozilla, OE, Kmail) do, so my read/unread pointers on my IMAP server don't get tracked correctly. Maybe it's a setup issue.
I would use them for the bulk of my mail if it weren't for this limitation, but since I like to use the GUI mail clients sometimes, neither of these works.
Actually, conceptually this could be used as a good thing. An "evil bit" that meant "drop this packet" could be used as a safety on network security testing. If I want to simulate an attack on machine A, I tell machine A to ignore the "evil bit" while machine B will continue to honor it. I don't have to worry about machine B doing anything based on my test packets, since it will drop them on the floor based on the "evil bit."
It could also be used to test firewalls for any possible damage done by malicious packets hitting the firewall and being dropped, in this case due to the "evil bit."
From what I could tell from reading those papers, softupdates can't recover if the corruption occurs while writing a sector to the drive. With that kind of error a journal is necessary to be able to recover. The other solution would be to always rewrite the superblock for every update and maintain several copies of the superblock (effectively journalling only the superblock.) This is essentially what TUX2 does.
I could be mistaken, but my understanding is that softupdates may allow you to update an on-disk inode or directory structure that has other things pointing to it, which could corrupt a filesystem if the system crashed during the update.
What's your understanding of this? Am I mistaken?
There are some problems in the system, but those can be worked out. You need enough random senators so that no single interest is likely to be overrepresented. Terms will have to be long enough so that people don't push through half-baked laws that only mildly represent their interests before they leave office. This can be helped by tying "senators" to specific bills or classes of law instead of time periods, etc. The executive branch will probably still have to be either elected or appointed, but their power is already constrained by the legislative branch, and could be further constrained.
The biggest problem is who oversees the selection of random people. I'm not sure that there is a clear way to do this, as my experience with jury duty shows that current methods don't work, and result in the same person getting selected more often than is reasonable.
All of this is distracting the point that selecting random members of the populace is a reasonable solution to the problem of people being too stupid to vote in elections, as voting on issues is easier than voting for people.
Actually, it's not that complex. IP ids are a strictly one-way phoenomenon. If none of the packets going to the NAT are fragmented (common for most home NATs), all you have to do is set the IPid field to something independent of the contents of the incoming packet (say, a random number generator or even a counter) and you're done. If you do receive fragmented packets, it's a little more complex because you have to have a way to translate from the incoming packets to outgoing ones. The easiest way to do this is:
1.) On bootup choose a large random key.
2.) When you get a packet, do a secure hash on the incoming IP, the IPid field on the incoming packet, and the key, and store the result in the IPid field of the outgoing packet.
3.) Rekey as needed. Choosing a sufficiently large random key should reduce the need to rekey, and, if possible, rekeying should be done during times of network inactivity.
I don't think either Stevens or Breyer would agree with your assessment of their arguments. Both of them (though Stevens moreso than Breyer) consider abiding by the constitution as their primary purpose.
Breyer's argument is the weaker of the two because he gives the most strength to the "promote the progress of science" part of the statute. He argues that even the extension on copyrights for new works fails the constitutional requirements for "limited times" and promotion of the "progress of science." He makes a convincing argument that the courts have an obligation to set limits on what "limited times" means, and then spends a lot of time arguing that life of the author plus 70 years is so long as to fail the "limited times" requirement. He makes a reasonable argument that the courts do have the obligation to set limits (they're the ones who have to strike down a law that extends copyright to life of the author plus 10,000 years,) but fails to fully convince at least me that the current extension is so grossly out of balance that it crosses the line between being merely poorly conceived to being unconstitutional.
To understand Stevens' argument (which is substantially more convincing thand Breyer's) it is necessary to understand the majority opinion. The Ginsburg opinion seems to rely primarily on early patent cases and the existence of previous copyright term extensions by congress to decide that the framers did not intend the reading of the copyright clause that Eldred requests. Their argument primarily rests on three things: the Copyright Act of 1790, which established copyrights in the United States, a number of individual patent extensions passed between 1790 and 1875, and the Copyright Act of 1831, which was the first extension of copyright terms on existing works. From these, as well as continued congressional practice in extending copyrights on existing works, the court concludes that in both the framers' and in the modern legal framework, the CTEA is constitutional.
Stevens argues that the constitution, not early congressional actions must be the basis of our law, and that many of the actions that the majority uses to support retroactive extensions either don't apply or are blatantly unconstitutional. The copyright act of 1790, he argues, does not apply in this case because, while it did give copyright protection to existing works, did so in the context of establishing a national system of copyrights, and the founders were keenly aware of the difference between this establishment of copyrights and the extension of existing ones. Many of the patent extensions used as evidence of the framers intent were blatently unconstitutional extensions of patents on inventions that had already entered the public domain, and so are unconvincing as a basis for modern case law. The copyright extension act of 1831 cannot be used to derive the framers' intent because none of the original delgates were in the 1831 congress. Further, the 1831 act was based on a view of copyright judged unconstitutional in the 1834 case of Wheaton v. Peters. All of this goes to show that the historical precedent for constitutionality of the extension of copyrights is inconclusive at best.
Stevens further points out that protection against ex post facto laws should protect the interests of both the patentee and the public with respect to copyrights. Just as congress should be unable to shorten the term of existing copyrights (thus harming the patentee), they should also be unable to extend the term of existing copyrights (thus harming the public.)
The point of all this is to show that both Stevens and Breyer very much had the constitution in mind when forming their opinions, and they are based on reasonable interpretation of the text. Stevens makes it very clear that he considers congressional practice an inappropriate way of deciding constitutionality and is not uncomfortable with the possibility of this decision putting previous copyright term extensions on similar shaky ground, if they are indeed unconstitutional. Breyer's arguments are less clear on this, and he seems to go out of his way to show how the copyright term extinsions in 1976 could have served a constitutional purpose, while the current statute does not. This interchange makes me wonder whether some of the other justices' opinions were based primarily on a desire not to unravel 170 years worth of copyright term extension acts.
It's somewhere in-between. You get the "anything can happen in any order" that you would get with a modern processor, but cache synchronization bugs can be hidden. This is because the two threads share a common cache, and so there is minimal latency between when one thread does a cache write and when the other thread has that data in its cache, hiding bugs in the sychronization code that prevents using stale cache data.
Most likely you'd catch any bugs that were in a higher-level language just fine, but you might miss bugs at the assembly level. For instance, you'd catch bugs in the code that uses a mutex, but might not catch bugs in the code that implements said mutex.
I agree with you with respect to Purify and the like. This seems to really be where most of the innovation is happening in debugging tools.
They're less general purpose, but they tend to catch things that are substantially more difficult to debug with a normal debugger.
Ahhh, but this is another example of monopoly maintenance, and the entire purpose of monopoly maintenance on the Windows product is to prevent other OSes (like GNU/Linux) from being able to become dominant.
If all Microsoft did was stay the leading OS product, adding bugfixes, upgrading the File System, etc., this would be acceptable. I'll admit that there are tendencies towards a natural monopoly in the Operating Systems business. The problem is that Microsoft goes beyond this. They integrate Windows-only variations on existing or developing middleware technologies and sell them at little to no cost in order to preserve their monopoly in the OS market.
I didn't understand exactly how this worked until I read the original verdict in the Microsoft trial. Multiplatform middleware techologies are a threat to Microsoft's Windows monopoly because they allow users to switch to other operating systems and still achieve the same functionality they have in Windows. Thus, Microsoft must replace multiplatform middleware technologies with Windows-only equivalents.
This is where Microsoft's unique position as the monopoly seller of operating systems comes in. By integrating Windows-only middleware technology into the the base Windows product, Microsoft can offer that middleware at a cost below what any competitor can possibly offer (i.e. free, and with no installation costs.) Since Microsoft Windows is the monopoly OS, content producers are virtually guaranteed to produce content for the Windows-only technology, since it will be installed on almost every computer in their potential market. The multiplatform technology thus falls by the wayside, and becomes an expensive afterthought for both consumers and content producers. In this way Microsoft harms the competitive abilities of the middleware companies and more importantly of alternative OSes that run that Middleware.
PDF is one of the remaining big middleware technology threats, and so Microsoft is coming up with a Windows-only version. By integrating it into the OS (I guarantee you, the reader for this format will be a free download and will be included with the next version of Windows) Microsoft will guarantee that content providers will produce content for this format, at the expense of PDF and more importantly at the expense of other OSes that rely on PDF for document exchange.
Are consumers or content producers dumb for choosing to use the free Microsoft product instead of the multiplatform alternative? No. At worst they are a little shortsighted, but even that is questionable. Most consumers can't conceive of using another OS besides Windows. It does most of what they want, and there are no reasonable alternatives that provide the same capabilities at the same complete system cost. GNU/Linux could conceivably get there, but it isn't there yet, and Microsoft can keep it that way by preventing the development of popular multiplatform middleware that would allow consumers to switch (even if they wanted to, which they don't.) The government had the opportunity to try and prevent further monopoly maintenance in this fashion, but it chose not to, and as the parent poster noted, gave Microsoft carte blanche to integrate whatever Windows-only technologies it wants into the OS in order to prevent the development of middleware that would lead to competition in the OS market.
How is all this bad for consumers? I can think of two reasons. First, it allows Microsoft to extract monopoly profits from its OS products, thus costing consumers money. Second, it prevents the innovation that real market competition usually leads to. There are probably others, but both of these are real consumer harms that could be fixed by restoring competition in the OS market.
Because this is all anyone really needs to know.
That one sentence explains the entire ruling, barring a single technicality.