Yes I know that he means well and believes he is helping the cause but his extreme black and white mentality on the issue of OSS does just the opposite. I respect his contributions to the OSS movement and even agree that it's beneficial for someone to raise these various concerns. However, I think the OSS movement would benefit far more by having more pragmatic and moderate leaders while retaining individuals like Stallman as gadflys.
In particular I take the goal of open source software to be providing as much functionality and power as possible through open source software. That is the point of the GPL vs. liscenses like the BSD is to encourage future contributions of code by corporate and individual code. Sure some individual coders may have emotional reasons for not wanting their effort used by another for private gain but the point of these liscenses for the community is merely to promote future development and contribution by forcing later modifications to be released. In other words if we discovered that liscensing software under a BSD style liscense generated more contributions we should do that.
Thus insisting on idealogy can sometimes be detrimental. Consider insisting on a GPL style liscense when you had a trustworthy promise from IBM to donate many programmers to the project if it was liscensed in the BSD style. Supposing you knew that IBM (and other people) would simply write a proprietary competitor rather than be forced to open source a proprietary technology they wish to offer as an add on to the project. Sure the ideaology can be useful and in many cases a GPL liscense will result in more contributions but it is stupid to ignore the fact that this ideology is directed at a goal, more OSS and thus it is counterproductive to insist on this ideology when it discourages OSS.
Yet this is exactly what Stallman's vociferous black and white view of the world does. By effectively equating any company or product that does not live up to his idea of open source with proprietary software he diminishes any benefit these companies might have to contribute to the open source world. In fact choosing to contribute to OSS in certain ways can actually result in negative publicity since all the sudden Stallman is critisizing you for not going all the way and people start looking at you as an evil corporation.
As a result of this attitude companies like Sun get far less benefit out of open sourceing their software then they might otherwise. If instead of postive PR for releasing the code to open office (a very valuable free software tool I might add) they are instead ripped for not fully open sourceing their java compilers/specs and the benefits to sun of the open source programming community are drasticly reduced by a fork designed to avoid Sun's java they will be much less inclined to donate code in the future.
Sure it is perfectly withing Stallman's rights to fork an OSS project and for this particular project it might even result in more purely open source driven software. However, in the long term this sort of attitude actually harms the OSS movement by sending companies a message that it isn't worthwhile to open source some of their systems or fund the development of an open source framework they can augment with their proprietary technology. It's alot like the problem with enviornmentalists who critique cleaner power plants/companies for not doing enough or for otherwise not living up to the perfect enviornmental ideal while leaving old fashioned dirty coal plants alone because they either see no hope of change or lack of interest (a new cleaner power generation tech will attract interest that one of a 100 coal plants won't). As a result this good intentioned concern to make the planet as clean as possible can actually end up making it dirtier by reducing the incentive for companies to take partial steps in the right direction.
Finally let me remark that in this particular project we are not seeing an attempt by Sun to deviously lock open office into
First of all I agree with the first response as well. I know plenty of people who have spent their entire life on the hamster wheel working and trying to get rich (and mostly succeeding) but it doesn't make them happy and often makes them less happy. Moreover, this isn't just my opinion they actually do scientific studies releating income to reported happiness and for people at least of middle class more money doesn't have any positive correlation with happiness.
Secondly it is really easy when you are old to look back and say it would have been better if you had worked harder. That's because it would be *you* who are getting the benefits but *past you* who would have done the work. After all I find it really easy to wish I had spent the *past* year going to the gym regularly or writing the papers I have to write now. It's really easy to wish you had gotten some unpleasent task out of the way with earlier but that doesn't mean it is necessarily better that you had done it.
But if your goal in life is just accomplishment then working hard and starting up your buisness is a great idea. My point is just that many of us are interested in happiness rather than success.
So I agree with Scheiner REAL ID is an absolutely terrible idea (it combines the worst security aspects of national and state IDs) it really isn't clear to me that a national ID card could not make identification more reliable as well as realizing significant economic savings by standardization.
In particular while I agree that using one ID system introduces a common point of high value failure it also economiclly feasible to invest a great deal more in the ID system. If one ID replaces n IDs you can make the ID cost roughly about the sum of the costs of all those other IDs. If one national ID replaced all our driver's liscensces, passports, credit cards and so forth it could afford more sophisticated safegaurds than any of the former IDs individually.
So while REAL ID seems to introduce the single point of failure without benefitting from economies of scale it seems perfectly possible that at some point in the future the increased safety provided by cryptographic smart card features, biometrics, and other possible features would outweigh the safety disadvantages of one point of attack. Furthermore the amount of resources spent to verify the card holder at issuance or for replacement could be similarily increased.
Furthermore it seems to me that our current system already has the problem of a single point of attack insofar as is relevant to terror. I let my drivers liscensce expire and prove my identity entierly via my passport. I have never had to produce any other ID for airline flights or other government related authentication. I have no doubt I could get a credit card or SSN number with a passport plus some easy to aquire items (bills to your address in that name etc..).
So while one wouldn't want to implement a system like REAL ID, or any system that hasn't already gone through some extensive real world testing. It seems at least possible that the increased resources availible by combining IDs could be used to more than outweigh any disadvantges, especially since the relevant ID systems already suffer from many of the purported disadvantages.
Admitedly I may be somewhat biased against this guy because of his stupid claims about nerds and web service applications. It is obviously outright false that nerds are not popular in secondary school because they have better things to do than spend time trying to be popular. In both my personal and observed experience nerds often try despartly hard to be popular (in fact they often have a problem of trying too hard and too transparently). Moreover, it just isn't true that other groups with large time comitments and interests outside of school are automatically as unpopular (music people etc..). While I am uncertain about the future of web service applications I expect a better analysis than the same crap which has been used to predict network computers for years.
As for this actual point I agree there is a small grain of truth in it. The advent of computer programming and other large profitable fields of intellectual endeavor makes intelligence hugely more valuable for a company. The productivity difference between someone who is just average and someone who is really fucking brilliant can be very large. So it really does make sense for companies to pay new recruiters big bucks just for being smart and train them on the job. Oracle is already doing this recruiting people from caltech with no programming experience for 80k starting salaries. However, people with those sort of smarts are extremely rare and so this trend does not hold out hope for the vast majority of CS students much less undergrads in general. As a TA at Berkeley it is clear that even here most of the CS students are not of the caliber necessery to launch compelling new products.
Moreover, I think the fundamental flaw in this analysis is the authors assumption that people, even young just out of college types, are interested in maximizing their expected revenue. Sure these people would like to earn more but alot of them just don't find it worthwhile to live for years eating Raman and living in a hovel for the promise of later riches. Happiness and utility are sub-linear in wealth and so great riches later in life don't necessarily compensate for prior poverty. In short many people really would prefer to be comfortably well off for most of their life and have the time and resources to start a family or pursue other interests out of college rather than being poor and working 100 hour weeks for several horrible years for later money. Heck I sure as hell wouldn't want to waste my youth as a workaholic just to end up as one of those rich bachelors at 35.
Also while the author touches on risk he radiclly misunderstands most people's attitudes towards risk. Sure studies have shown that there is some percent of society who are naturally thrill seekers and thrive on risk but many of the rest of us find risk itself (and not just the bad consequences if things go bad) unpleasent. Most people don't like risking alot for the hope of a big reward but would prefer a comfortable sure thing.
Big companies with stable employment and paychecks which don't depend entierly on project success exist for a reason. Most people prefer that sort of comfortable stable life. It isn't all about accumulating the biggest bank account but also about knowing you can provide for a family, have free time and safely plan for the future. It isn't just the economic situation that needs to change to make start-ups as prevalent as the author imagines but human nature itself
So this article got me thinking about what it would take to make a program which automatically scans binary software for OS code. I imagine it is possible but it would be an interesting programming problem.
One early thought is that you could scan for matching arithmetic operations. Walk through the assembely and keep a table of register contents/memory contents/constant loads to regenerate algabraic operations. By transforming these operations to some canonical form one could match algabraic operations from the source regardless of compiler optimization or variable renaming.
Of course there are several problems with this approach. First implemented in the obvious fashion it is horribly slow (like N^2M^2 N=binary size M=Source code files). Secondly some programs may do very little explicit algabraic manipulations. Finally common snippets of array bounds logic or pointer arithmetic may trigger false positives.
Well perhaps in some technical sense but in any realistic sense his would work. For one most copyright holdders aren't going to bother to sue you, Secondly I expect courts will be quite loath to award damages in this case. What the copyright holder intended is now being done so no injunctive relief will be given and it is unlikely you could prove damages for their violation of the GPL.
Finally, one could presumably just assert that it was under the GPL all along and you are just now releasing the code.
The real answer is that these situations are legally questionable. I'm not sure what the result of an actual court case about someone who copies an single algorithm (and not overall code organization) from copyrighting code. However I do know that when companies reimplement another companies copyrighted code they make sure to go out of their way not to hire anyone who has looked at the original just to make sure they have no copyright problems. Even if you are ultimately able to win such a copyright suit would still be very complex and expensive.
Thus if you have seen MS source code it is a bad idea to work on a FOSS product as it gives MS legal leverage against any related code you write. Of course with 5 million dollars in lawyers fees you could probably win the case but that wouldn't be much help to most FOSS products.
On the other hand it is quite unlikely that the GNU or OSDF people are going to go after a commercial product which copied an algorithm but not the code itself. Not only do they not have any incentive (they don't benefit by trying to bankcrupt their competitors) but these organizations are generally opposed to the idea that stealing code ideas (i.e. software patents) should be a crime.
First while surely one can't derive the principle I suggested from pure logic this has not bearing on whether or not it is objectively true. One shouldn't confuse epistemology and metaphysics. All moral theories make a claim to objective truth (well except nihlism/pure relativism) the question is just which one is right.
Also since I tend to believe the average happiness is positive killing off everyone would reduce the future average happiness.
Secondly, I did not mean to imply that any societal deciscion about who is human and who is not is equally good. In particular the entire point is to avoid the suffering of grief and forseen doom/danger. Thus excluding any group who is capable of realizing their exclusion and be upset about it from the societal definition of human is therefore wrong. What perhaps would have been a better way to say what I meant is that the minimum group that needs to not be murdered etc.. are those people capable of understanding and fearing their murder but the extension to children and the severely retarded is just societal convention (though of course in some socities one convention may pragmatically be better than another).
I don't know if I would actually kill said depressed person. Since I'm sure I would feel guilty (irrationally or not) it would probably be on whole a bad thing for me. In any case this is irrelevant, most christians would not actually martyr themselves rather than renounce their faith but this doesn't mean this isn't still the standard they strive towards.
Finally in the Chimera case I suspect the very existance of the Chimera would cause a great deal of distress in many people. Killing it early would alleviate much of this distress because they would not see it acting fully human like (b/c it is killed before fully developed) and this would probably reduce the number of people who get emotionally involved in the question. Moreover, I suspect that people like you would feel some distress at this result but I submit this is alot less intense than the distress disabled people would feel upon learning that someone had been killed just because they are retarded.
Regardless of what you think is right clearly many people would not accord a chimera the full emotional regard they do a human. This means the harm from killing a chimera will be much lower.
Ohh yes I agree completely that we may have to face this question eventually. If for instance instead of growing human brain cells in a sheep they were doing so in higher primates I would not be so unconcerned.
So yes the ethical questions relating to chimeras need to be dealt with at some point. I was just saying this isn't one of those cases. While it does suggestively raise the issue I don't think this is much different than putting human genes in bacteria to synthesize antibodies. In fact the resulting brain will differ from a normal human brain to a much greater extent than the monkey brain differs from ours. So unless you plan to stop all monkey research until we figure out whether they are really just harry 'humans' or are animals it doesn't make sense to worry about this particular case. Our moral worries should concentrate around the animals which are the closest to us and despite this sheep case that will still be monkeys.
---
I was previously avoiding bringing up my own ethical views about retarded people and people in general since I have some very uncommon views. In essence I am a complete and total utilitarian. So I really do think all ethical problems are about encouraging happy experiences and discouraging unhappy ones and that cognitive ability is only ethically relevant insofar as it affects the human/animals experiental enjoyment. In particular while I expect the intensity of experiences correlates fairly well with cognitive ability in species (smarter creaters generally have more of the violitional/attentional/emotive features we think are associated with conciouss experience) they are not one and the same thing. For instance medical procedures which completely sever the frontal lobes from the rest of the brain destroy most of an individuals reasoning and planning ability but leave intact identity, violition and reported subjective states (for a variety of reasons I tend to believe the neural correlate of conciousness is in the thalamus specifically the intra-laminar nucleas but this is just an educated guess by a top neurosurgen/scientist I found compelling).
As a result I think the primary reason for moral prohibitions against killing, withholding medical care, or using them in organ experiments in regards to people but not other mammals is because of their societal impact on humans. Thus while both non-human mamals (and people have experiences and suffer pain (though people perhaps more intensly than animals) the reason it is okay to raise animals for food and not people is the suffering this practice would induce in the friends and relatives of the eaten person. Not to mention the suffering which would inflicted by the knowledge a person had they were being raised for food.
In short the reason it isn't okay just to kill unhappy people (killing happy people would reduce average utility) is because such actions would make their relatives sad and cause other people to fear they too might be killed. Similarly the reason we accord very retarded people all the same protections as other humans is not because of some essential essence of humanity they posses. Rather it is because not according them these rights causes many other people to become upset and worry that they or their loved ones might be treated in a similar manner. Since few people are retarded the cost of caring for them properly and granting them rights is small compared to the amount of distress it would cause if others knew they were being treated as animals.
So in a very real sense who should get the moral protections afforded to living humans is determined by what society regards as a living human. If society really and truly did not regard someone as a person until they were 1 year of age and felt no more emotional attachment to newborns than to fetuses I would have no problem with 'abortion' up until 1 year. Of course as a matter of fact people are much more bothered by the idea of a newborn being killed than a fetus not to mention the strong attachment people start to d
First it just isn't the case that there are no ethical questions experimenting in sheep. Our best evidence indicates sheep have experiences and suffer pain similarly to the way we do and thus we have rules about unnecessary cruelty and using anesthisia and the like in labs.
Secondly you are confusing the question I was answering, will these animals be human like in their cognitive capabilities, with the ethical notion of being human. Likely these are related issues but many people attach ethical import to being born from human parents and other factors for a variety of reasons (avoiding slipperly slopes and the irrational emotional attachment we have to those things we recognize as human).
Finally when I say that it is the whole brain development system which makes someone human it clearly does not mean there is a sharp dividing line and if you have the full system you are human but if you fall one iota short you are not human. I mean it is the whole development system that makes us human in the same sense that it is all the water in the atlantic which makes it an ocean. Sure taking a cup of water out of the atlantic would reduce its oceanness but not very much.
So the point is that being human requires a huge number of factors. Just as missing only a few of these does not make someone totally non-human (retarded people) giving a sheep only one of these does not suddenly make it human.
One of the things I really liked about the movie was that it was nice and cheery unlike the last book in the series. The magic of HG2G is in the lighthearted humor and fun style if they try and copy the depressing last book it would ruin the movies even more than it did the books.
I agree this is an interesting question but human intellectual capacity consists of alot more than one linguistic gene. After all people who have had strokes and suffered severe damage to linguistic ability do not revert to being mere apes and still retain many other complex human tool usage and non-verbal reasoning capabilities.
Of course it is quite possible that the change in linguistic capability was an evolutionary spark that encouraged our other brain changes but it isn't just that on its own. If only one gene was able to make this significant a difference in other primates I would be extremely surprised that they had not evolved this mutation on their own. The state with mice is even more extreme as they likely don't even have a large enough body to hold the sort of brain structure humans have.
Still, your point that the line between humans and animals is not so sharp is valid. If this experiment had been about growing human brain cells in some kind of primate I would be concerned since the developmental triggers might be similar enough to produce some kind of hybrid (assuming you somehow just replaced the cells differentiating into the monkey brain with human ones in some sensible fashion). This still might be impossible, I just don't know, but I would not be so confident in such a case.
This wasn't the question. The question was a purely scientific one of whether this would result in sheep with the reasoning and cognitive abilities of a person.
The question of how subjective mental experience arises is an entierly other matter. In fact I would argue that we have nearly as much evidence that sheep have experiences, suffer pain and otherwise have subjective states as we do for other humans. Of course it is logically possible that sheep have no subjective experiences (as we assume is true for things like tables) but it is also logically possible that other people don't either. Since in both cases we see a strong resembalance in both underlying anatomy (sheep neurons are much like human ones) and observe reactions we associate with subjective experience in ourselves (we can fairly easily recognize when a mammal is in pain) I think there is almost as strong a case that sheep have experiences as there is that other people have experiences.
Thus this issue is pretty much irrelevant to the moral question of whether we should put human brain cells in sheep. Given our state of knowledge we should already be treating those sheep as if they had experiences (though being unaware of abstract concepts like death this does not mean we need to treat them like people).
In short *anything* is logically possible to cause conciouss experience but you don't worry that your computer might be suffering when you install the newest kind of software. If we want to make any choices we need to make the pragmatic assumption that morally relevant states can be deduced from objectively availible information otherwise we could never clip our toenails for fear they would feel pain.
Just injecting some human brain cells into a sheep or even transplanting an early human brain from a fetus is unlikely to produce any kind of human sheep. The human brain doesn't just develop from a genetic blueprint but also requires a huge amount of deveelopmental cues and responds to hormones and signaling molecules (like sonic the hedgehog) to develop properly. Not to mention a host of enviornmental stimuli needed to encourage the brain to wire itself correctly.
In short it isn't just human neurons which make us human but the whole brain development system at work in babies. This isn't the sort of thing which could be duplicated in a sheep without extensive genetic modification or hand controlling all the developmental signals. If this is possible at all it is far beyond our current level of technology.
So don't get freaked out yet people. They are just growing human neurons in sheep at the moment there is no chance we will make a person trapped in a sheep body.
Ohh and in case anyone is curious there is a pretty efficent algorithm to calculate the n-th hexadecimal digit of pi without calculating any of the previous ones
Clearly pi is not random in any absolute sense. One can give an algorithm to calculate it. Of course if you don't have the benefits of a hardware random number generator you are stuck choosing some psuedo-random source but the idea that there is some universal hierarchy of psuedo-random algorithms is just downright idiotic.
Whether a particular psuedo-random algorithm is good or not depends quite heavily on what it is being used for. The requirements for a psuedo-random number for a monte-carlo simulation are much different than those needed for cryptography (in one you want an even distribution over some relevant space while in the other it is usually unpredictability that is desired). Moreover, even within one type of application the specific algorithm being used can make a huge difference. For instance if I am trying to use monte-carlo integration on some nice smooth polynomial the standard psuedo-random technique of feeding back the output of a mod operation to itself will work quite well but if you are trying to integrate a function defined by a similar sort of recursion on mods it may perform horribly.
So there simply can't be any answer of what psuedo-random algorithm is better or worse. Of course one might still ask what would be a good choice for the standard psuedo-random algorithm in an OS given most people's software choices. For this application you want some source that won't displat a pattern relevant to normal computation. Thus, since trigonometric and other pi related functions are very commonly used pi is probably a fairly dangerous choice even if it works well on test cases.
In other words any number or algorithm that like pi pops up in many computer applications is going to be a bad choice. Not because you can run some ridiculous test which will rank its 'randomness' but because it is much more likely to be relevant to the program someone is running. In other words pi is a bad pragmatic choice because someone is alot more likely to use monte-carlo techniques to calculate pi or something to do with pi than with some recurrent mod generator (or something more sophisticated but not so common).
Thus from a pragmatic point of view these tests are just dumb. They would be better off running an array of common software and seeing how it performs. From a mathematical point of view there is nothing to be said as all of these sources (except the hardware one) aren't random at all so the only question is pragmatic.
I dunno about the impact. I expect the iraqi insurgents already have a much better idea of how our security operations work than they can gain through this document. After all they have the opportunity to actually observe this operations and how they react to their attacks.
I expect the information was just concealed as a matter of course in case it might contain something of value to the enemy. Still though your overall point is valid. There is no reason the public really needs to know this stuff and it is better safe than sorry with information which might let them kill more soldiers.
Hmm, really? sendmail can't pass a switch to make sure it gets a response from the authoritative server not just a cache?
I wonder then why my MX records seem to work almost immediatly while other changes take much longer to propogate. Can MX records have different TTL than the domain itself? If so maybe that is the explanation.
Well that sounds like a good explanation for why major ISPs might ignore the TTL. That's alot of wasted lookups and server load. For a large ISP we can expect that someone is going to most Akamai cites every minute so that is at least one lookup per minute per Akamai DNS record.
If other people do this as well ISPs might not have much of a choice but to ignore TTL and maybe there isn't any easy way to force the machine to use a minimum TTL but respect TTL above that limit.
My understanding is that many mail delivery programs do their own recursive resolve and often ignore cached entries. At least when I have updated MX records they have seemed to propogate almost instantly. I presumed this was because sendmail and the like insist on reading the MX from servers authoritative for that domain but perhaps I'm mistaken.
Anyone care to clarify how DNS cache works with MX records?
While this is certainly a disturbing development the constitutionality of it is unclear and quite interesting.
On the one hand the first ammendment certainly does not require libraries to provide pornographic magazines or otherwise provide some unbiased representation of viewpoints. In general the first ammendment does not restrict the government from providing some content but not others (except when this infringes on the establishment clause).
However, while library filtering has been deemed constitional (link here) the supreme court has also ruled that libraries must allow adults to bypass the filters. In other words apparently the supreme court has recognized that internet filtering is significantly different than buying library books. The library has legitimate financial constraints in what books it provides but does not in internet filtering.
So the question becomes very unclear in the case of truck stops. Since these are entierly automated they can't very well demand a librarian turn the filtering off. Still, since one does need to be at least 16 to drive and because of the real possibility that by providing enough government internet access filtering could stifle free speech I imagine it would be declared unconstitutional but it is a tough call.
While this is certainly a disturbing development the constitutionality of it is unclear and quite interesting.
On the one hand the first ammendment certainly does not require libraries to provide pornographic magazines or otherwise provide some unbiased representation of viewpoints. In general the first ammendment does not restrict the government from providing some content but not others (except when this infringes on the establishment clause).
So the question becomes very unclear in the case of truck stops. Since these are entierly automated they can't very well demand a librarian turn the filtering off. Still, since one does need to be at least 16 to drive and because of the real possibility that by providing enough government internet access filtering could stifle free speech I imagine it would be declared unconstitutional but it is a tough call.
Sure, this may be the same thing that happened with Samba but linus never made any claims about the legality. Something which may be technically and legally similar is not necessarily just as good a strategic idea.
In the case of microsoft we had a widely deployed piece of software that the open source community needed to interact with for compatibility reasons. Nothing of the kind is true with BitKeeper. In the case of BitKeeper the open source community could have simply built their own incompatible protocol and not have to worry about being shut out of the market by a BitMover monopoly. On the other hand in the case of microsoft the open source community couldn't simply build a better protocol than Samab but really needed to be compatible.
Secondly, while it's possible I very much doubt that the BitKeeper protocol was being reverse engineered from the expensive pro version. Most likely it being reverse enginered from the free versions (or at least comped versions). Unlike microsoft which needs to keep Samba out there in every windows box BitMover was just allowing this free usage as a donation/PR move and could easily revoke it without comprimising their buisness model.
In short by trying to reverse engineer this protocol it seems that Andrew? gave the impression that the 'price' of donating expensive software to open source projects is to have your market advantages reverse engineered and probably implemented in free projects. So while sure he has just as much right to reverse engineer in samba the first instance is an important blow against a monopoly trying to use propietary protocols to unfairly strangle competition. In this case there was no similar monopolistic pressure (there isn't a strong installed base of BK users who we need to be compatible with) and made it look like there was a steep price for trying to help the open source community.
Regardless of what you think of the deciscion to use BK or the need to reverse engineer this project having someone paid by the SAME organization which is the beneficiary of the free software (or at least appears to be in the media) is surely a bad move politically. It certainly would give me pause if I was a manager at a big corporation thinking of donating some helpful development tools to some open source project.
Yes the key word here is 'comprimised set of device keys'
The way this worked in CSS and probably works similarly here is that at the begining to the disk they encrypt a disk key with many different device keys. Then each device decrypts the disk key using their own device key.
However if you work out the math it simply isn't plausible to include a seperate key for every HD DVD player that might ever be sold (imagine 128 bits for an AES key). Instead each manufacturer, or perhaps even DVD player model in this new system, gets one key. They can then 'revoke' these keys by just refusing to encrypt future DVD keys with these device keys but since each DVD player doesn't have its own key they can't disable movies player by player.
On another point I would find it to be really unlikely that any major DVD player would truly get this penalty imposed against it. It would be a huge loss to be the first movie that doesn't work on sony blah players so no movie company is going to be the one who takes that first step.
Instead this is really a measure to deter manufacturers from 'accidently' making their DVD players ignore copy protection or otherwise violate their rules. Thus it is likely to be used when a player first hits the market or not at all.
Really, so if in the 1800's people had managed to find a high density propelant they should have devoted a significant fraction of their resources to building colonies on the moon out of welded metal and rivets?
No, I think the better response would have been to devot that money to improving the underlying technology and they would actually get self sustaining colonies faster. I am arguing that we might actually make the colonies faster by devoting research money to the underlying technology that might let us make real self sustaining colonies rather than wasting money support people on mars when we know we aren't yet at the point of being able to make self-sustaining colonies.
This is kinda like the rendering dilema. Suppose you have a big scene to render should you immediatly buy a bunch of computers and start to run the job? Will that get it done the fastest supposing you only have Y dollars to spend on the job? Not necessarily if the rendering project would take really long on today's computers sometimes it takes less time just to invest that money and then buy new faster computers in a year which can finish the job before the original computers would have been able to do.
I'm arguing that this is the situation now in spaceflight (though to an even greater extreme since the people doing rendering don't usually have the choice of investing in processor research). It can actually get us self sustaining colonies faster by investing the money now in the underlying technologies than by wasting it sending people into space atop very expensive rockets.
Slashdot Mirror
Yes I know that he means well and believes he is helping the cause but his extreme black and white mentality on the issue of OSS does just the opposite. I respect his contributions to the OSS movement and even agree that it's beneficial for someone to raise these various concerns. However, I think the OSS movement would benefit far more by having more pragmatic and moderate leaders while retaining individuals like Stallman as gadflys.
In particular I take the goal of open source software to be providing as much functionality and power as possible through open source software. That is the point of the GPL vs. liscenses like the BSD is to encourage future contributions of code by corporate and individual code. Sure some individual coders may have emotional reasons for not wanting their effort used by another for private gain but the point of these liscenses for the community is merely to promote future development and contribution by forcing later modifications to be released. In other words if we discovered that liscensing software under a BSD style liscense generated more contributions we should do that.
Thus insisting on idealogy can sometimes be detrimental. Consider insisting on a GPL style liscense when you had a trustworthy promise from IBM to donate many programmers to the project if it was liscensed in the BSD style. Supposing you knew that IBM (and other people) would simply write a proprietary competitor rather than be forced to open source a proprietary technology they wish to offer as an add on to the project. Sure the ideaology can be useful and in many cases a GPL liscense will result in more contributions but it is stupid to ignore the fact that this ideology is directed at a goal, more OSS and thus it is counterproductive to insist on this ideology when it discourages OSS.
Yet this is exactly what Stallman's vociferous black and white view of the world does. By effectively equating any company or product that does not live up to his idea of open source with proprietary software he diminishes any benefit these companies might have to contribute to the open source world. In fact choosing to contribute to OSS in certain ways can actually result in negative publicity since all the sudden Stallman is critisizing you for not going all the way and people start looking at you as an evil corporation.
As a result of this attitude companies like Sun get far less benefit out of open sourceing their software then they might otherwise. If instead of postive PR for releasing the code to open office (a very valuable free software tool I might add) they are instead ripped for not fully open sourceing their java compilers/specs and the benefits to sun of the open source programming community are drasticly reduced by a fork designed to avoid Sun's java they will be much less inclined to donate code in the future.
Sure it is perfectly withing Stallman's rights to fork an OSS project and for this particular project it might even result in more purely open source driven software. However, in the long term this sort of attitude actually harms the OSS movement by sending companies a message that it isn't worthwhile to open source some of their systems or fund the development of an open source framework they can augment with their proprietary technology. It's alot like the problem with enviornmentalists who critique cleaner power plants/companies for not doing enough or for otherwise not living up to the perfect enviornmental ideal while leaving old fashioned dirty coal plants alone because they either see no hope of change or lack of interest (a new cleaner power generation tech will attract interest that one of a 100 coal plants won't). As a result this good intentioned concern to make the planet as clean as possible can actually end up making it dirtier by reducing the incentive for companies to take partial steps in the right direction.
Finally let me remark that in this particular project we are not seeing an attempt by Sun to deviously lock open office into
First of all I agree with the first response as well. I know plenty of people who have spent their entire life on the hamster wheel working and trying to get rich (and mostly succeeding) but it doesn't make them happy and often makes them less happy. Moreover, this isn't just my opinion they actually do scientific studies releating income to reported happiness and for people at least of middle class more money doesn't have any positive correlation with happiness.
Secondly it is really easy when you are old to look back and say it would have been better if you had worked harder. That's because it would be *you* who are getting the benefits but *past you* who would have done the work. After all I find it really easy to wish I had spent the *past* year going to the gym regularly or writing the papers I have to write now. It's really easy to wish you had gotten some unpleasent task out of the way with earlier but that doesn't mean it is necessarily better that you had done it.
But if your goal in life is just accomplishment then working hard and starting up your buisness is a great idea. My point is just that many of us are interested in happiness rather than success.
So I agree with Scheiner REAL ID is an absolutely terrible idea (it combines the worst security aspects of national and state IDs) it really isn't clear to me that a national ID card could not make identification more reliable as well as realizing significant economic savings by standardization.
In particular while I agree that using one ID system introduces a common point of high value failure it also economiclly feasible to invest a great deal more in the ID system. If one ID replaces n IDs you can make the ID cost roughly about the sum of the costs of all those other IDs. If one national ID replaced all our driver's liscensces, passports, credit cards and so forth it could afford more sophisticated safegaurds than any of the former IDs individually.
So while REAL ID seems to introduce the single point of failure without benefitting from economies of scale it seems perfectly possible that at some point in the future the increased safety provided by cryptographic smart card features, biometrics, and other possible features would outweigh the safety disadvantages of one point of attack. Furthermore the amount of resources spent to verify the card holder at issuance or for replacement could be similarily increased.
Furthermore it seems to me that our current system already has the problem of a single point of attack insofar as is relevant to terror. I let my drivers liscensce expire and prove my identity entierly via my passport. I have never had to produce any other ID for airline flights or other government related authentication. I have no doubt I could get a credit card or SSN number with a passport plus some easy to aquire items (bills to your address in that name etc..).
So while one wouldn't want to implement a system like REAL ID, or any system that hasn't already gone through some extensive real world testing. It seems at least possible that the increased resources availible by combining IDs could be used to more than outweigh any disadvantges, especially since the relevant ID systems already suffer from many of the purported disadvantages.
Admitedly I may be somewhat biased against this guy because of his stupid claims about nerds and web service applications. It is obviously outright false that nerds are not popular in secondary school because they have better things to do than spend time trying to be popular. In both my personal and observed experience nerds often try despartly hard to be popular (in fact they often have a problem of trying too hard and too transparently). Moreover, it just isn't true that other groups with large time comitments and interests outside of school are automatically as unpopular (music people etc..). While I am uncertain about the future of web service applications I expect a better analysis than the same crap which has been used to predict network computers for years.
As for this actual point I agree there is a small grain of truth in it. The advent of computer programming and other large profitable fields of intellectual endeavor makes intelligence hugely more valuable for a company. The productivity difference between someone who is just average and someone who is really fucking brilliant can be very large. So it really does make sense for companies to pay new recruiters big bucks just for being smart and train them on the job. Oracle is already doing this recruiting people from caltech with no programming experience for 80k starting salaries. However, people with those sort of smarts are extremely rare and so this trend does not hold out hope for the vast majority of CS students much less undergrads in general. As a TA at Berkeley it is clear that even here most of the CS students are not of the caliber necessery to launch compelling new products.
Moreover, I think the fundamental flaw in this analysis is the authors assumption that people, even young just out of college types, are interested in maximizing their expected revenue. Sure these people would like to earn more but alot of them just don't find it worthwhile to live for years eating Raman and living in a hovel for the promise of later riches. Happiness and utility are sub-linear in wealth and so great riches later in life don't necessarily compensate for prior poverty. In short many people really would prefer to be comfortably well off for most of their life and have the time and resources to start a family or pursue other interests out of college rather than being poor and working 100 hour weeks for several horrible years for later money. Heck I sure as hell wouldn't want to waste my youth as a workaholic just to end up as one of those rich bachelors at 35.
Also while the author touches on risk he radiclly misunderstands most people's attitudes towards risk. Sure studies have shown that there is some percent of society who are naturally thrill seekers and thrive on risk but many of the rest of us find risk itself (and not just the bad consequences if things go bad) unpleasent. Most people don't like risking alot for the hope of a big reward but would prefer a comfortable sure thing.
Big companies with stable employment and paychecks which don't depend entierly on project success exist for a reason. Most people prefer that sort of comfortable stable life. It isn't all about accumulating the biggest bank account but also about knowing you can provide for a family, have free time and safely plan for the future. It isn't just the economic situation that needs to change to make start-ups as prevalent as the author imagines but human nature itself
So this article got me thinking about what it would take to make a program which automatically scans binary software for OS code. I imagine it is possible but it would be an interesting programming problem.
One early thought is that you could scan for matching arithmetic operations. Walk through the assembely and keep a table of register contents/memory contents/constant loads to regenerate algabraic operations. By transforming these operations to some canonical form one could match algabraic operations from the source regardless of compiler optimization or variable renaming.
Of course there are several problems with this approach. First implemented in the obvious fashion it is horribly slow (like N^2M^2 N=binary size M=Source code files). Secondly some programs may do very little explicit algabraic manipulations. Finally common snippets of array bounds logic or pointer arithmetic may trigger false positives.
I wonder if there is a better solution?
Well perhaps in some technical sense but in any realistic sense his would work. For one most copyright holdders aren't going to bother to sue you, Secondly I expect courts will be quite loath to award damages in this case. What the copyright holder intended is now being done so no injunctive relief will be given and it is unlikely you could prove damages for their violation of the GPL.
Finally, one could presumably just assert that it was under the GPL all along and you are just now releasing the code.
The real answer is that these situations are legally questionable. I'm not sure what the result of an actual court case about someone who copies an single algorithm (and not overall code organization) from copyrighting code. However I do know that when companies reimplement another companies copyrighted code they make sure to go out of their way not to hire anyone who has looked at the original just to make sure they have no copyright problems. Even if you are ultimately able to win such a copyright suit would still be very complex and expensive.
Thus if you have seen MS source code it is a bad idea to work on a FOSS product as it gives MS legal leverage against any related code you write. Of course with 5 million dollars in lawyers fees you could probably win the case but that wouldn't be much help to most FOSS products.
On the other hand it is quite unlikely that the GNU or OSDF people are going to go after a commercial product which copied an algorithm but not the code itself. Not only do they not have any incentive (they don't benefit by trying to bankcrupt their competitors) but these organizations are generally opposed to the idea that stealing code ideas (i.e. software patents) should be a crime.
A couple clarifications.
First while surely one can't derive the principle I suggested from pure logic this has not bearing on whether or not it is objectively true. One shouldn't confuse epistemology and metaphysics. All moral theories make a claim to objective truth (well except nihlism/pure relativism) the question is just which one is right.
Also since I tend to believe the average happiness is positive killing off everyone would reduce the future average happiness.
Secondly, I did not mean to imply that any societal deciscion about who is human and who is not is equally good. In particular the entire point is to avoid the suffering of grief and forseen doom/danger. Thus excluding any group who is capable of realizing their exclusion and be upset about it from the societal definition of human is therefore wrong. What perhaps would have been a better way to say what I meant is that the minimum group that needs to not be murdered etc.. are those people capable of understanding and fearing their murder but the extension to children and the severely retarded is just societal convention (though of course in some socities one convention may pragmatically be better than another).
I don't know if I would actually kill said depressed person. Since I'm sure I would feel guilty (irrationally or not) it would probably be on whole a bad thing for me. In any case this is irrelevant, most christians would not actually martyr themselves rather than renounce their faith but this doesn't mean this isn't still the standard they strive towards.
Finally in the Chimera case I suspect the very existance of the Chimera would cause a great deal of distress in many people. Killing it early would alleviate much of this distress because they would not see it acting fully human like (b/c it is killed before fully developed) and this would probably reduce the number of people who get emotionally involved in the question. Moreover, I suspect that people like you would feel some distress at this result but I submit this is alot less intense than the distress disabled people would feel upon learning that someone had been killed just because they are retarded.
Regardless of what you think is right clearly many people would not accord a chimera the full emotional regard they do a human. This means the harm from killing a chimera will be much lower.
Ohh yes I agree completely that we may have to face this question eventually. If for instance instead of growing human brain cells in a sheep they were doing so in higher primates I would not be so unconcerned.
So yes the ethical questions relating to chimeras need to be dealt with at some point. I was just saying this isn't one of those cases. While it does suggestively raise the issue I don't think this is much different than putting human genes in bacteria to synthesize antibodies. In fact the resulting brain will differ from a normal human brain to a much greater extent than the monkey brain differs from ours. So unless you plan to stop all monkey research until we figure out whether they are really just harry 'humans' or are animals it doesn't make sense to worry about this particular case. Our moral worries should concentrate around the animals which are the closest to us and despite this sheep case that will still be monkeys.
---
I was previously avoiding bringing up my own ethical views about retarded people and people in general since I have some very uncommon views. In essence I am a complete and total utilitarian. So I really do think all ethical problems are about encouraging happy experiences and discouraging unhappy ones and that cognitive ability is only ethically relevant insofar as it affects the human/animals experiental enjoyment. In particular while I expect the intensity of experiences correlates fairly well with cognitive ability in species (smarter creaters generally have more of the violitional/attentional/emotive features we think are associated with conciouss experience) they are not one and the same thing. For instance medical procedures which completely sever the frontal lobes from the rest of the brain destroy most of an individuals reasoning and planning ability but leave intact identity, violition and reported subjective states (for a variety of reasons I tend to believe the neural correlate of conciousness is in the thalamus specifically the intra-laminar nucleas but this is just an educated guess by a top neurosurgen/scientist I found compelling).
As a result I think the primary reason for moral prohibitions against killing, withholding medical care, or using them in organ experiments in regards to people but not other mammals is because of their societal impact on humans. Thus while both non-human mamals (and people have experiences and suffer pain (though people perhaps more intensly than animals) the reason it is okay to raise animals for food and not people is the suffering this practice would induce in the friends and relatives of the eaten person. Not to mention the suffering which would inflicted by the knowledge a person had they were being raised for food.
In short the reason it isn't okay just to kill unhappy people (killing happy people would reduce average utility) is because such actions would make their relatives sad and cause other people to fear they too might be killed. Similarly the reason we accord very retarded people all the same protections as other humans is not because of some essential essence of humanity they posses. Rather it is because not according them these rights causes many other people to become upset and worry that they or their loved ones might be treated in a similar manner. Since few people are retarded the cost of caring for them properly and granting them rights is small compared to the amount of distress it would cause if others knew they were being treated as animals.
So in a very real sense who should get the moral protections afforded to living humans is determined by what society regards as a living human. If society really and truly did not regard someone as a person until they were 1 year of age and felt no more emotional attachment to newborns than to fetuses I would have no problem with 'abortion' up until 1 year. Of course as a matter of fact people are much more bothered by the idea of a newborn being killed than a fetus not to mention the strong attachment people start to d
You are missing the point in several ways.
First it just isn't the case that there are no ethical questions experimenting in sheep. Our best evidence indicates sheep have experiences and suffer pain similarly to the way we do and thus we have rules about unnecessary cruelty and using anesthisia and the like in labs.
Secondly you are confusing the question I was answering, will these animals be human like in their cognitive capabilities, with the ethical notion of being human. Likely these are related issues but many people attach ethical import to being born from human parents and other factors for a variety of reasons (avoiding slipperly slopes and the irrational emotional attachment we have to those things we recognize as human).
Finally when I say that it is the whole brain development system which makes someone human it clearly does not mean there is a sharp dividing line and if you have the full system you are human but if you fall one iota short you are not human. I mean it is the whole development system that makes us human in the same sense that it is all the water in the atlantic which makes it an ocean. Sure taking a cup of water out of the atlantic would reduce its oceanness but not very much.
So the point is that being human requires a huge number of factors. Just as missing only a few of these does not make someone totally non-human (retarded people) giving a sheep only one of these does not suddenly make it human.
One of the things I really liked about the movie was that it was nice and cheery unlike the last book in the series. The magic of HG2G is in the lighthearted humor and fun style if they try and copy the depressing last book it would ruin the movies even more than it did the books.
I agree this is an interesting question but human intellectual capacity consists of alot more than one linguistic gene. After all people who have had strokes and suffered severe damage to linguistic ability do not revert to being mere apes and still retain many other complex human tool usage and non-verbal reasoning capabilities.
Of course it is quite possible that the change in linguistic capability was an evolutionary spark that encouraged our other brain changes but it isn't just that on its own. If only one gene was able to make this significant a difference in other primates I would be extremely surprised that they had not evolved this mutation on their own. The state with mice is even more extreme as they likely don't even have a large enough body to hold the sort of brain structure humans have.
Still, your point that the line between humans and animals is not so sharp is valid. If this experiment had been about growing human brain cells in some kind of primate I would be concerned since the developmental triggers might be similar enough to produce some kind of hybrid (assuming you somehow just replaced the cells differentiating into the monkey brain with human ones in some sensible fashion). This still might be impossible, I just don't know, but I would not be so confident in such a case.
This wasn't the question. The question was a purely scientific one of whether this would result in sheep with the reasoning and cognitive abilities of a person.
The question of how subjective mental experience arises is an entierly other matter. In fact I would argue that we have nearly as much evidence that sheep have experiences, suffer pain and otherwise have subjective states as we do for other humans. Of course it is logically possible that sheep have no subjective experiences (as we assume is true for things like tables) but it is also logically possible that other people don't either. Since in both cases we see a strong resembalance in both underlying anatomy (sheep neurons are much like human ones) and observe reactions we associate with subjective experience in ourselves (we can fairly easily recognize when a mammal is in pain) I think there is almost as strong a case that sheep have experiences as there is that other people have experiences.
Thus this issue is pretty much irrelevant to the moral question of whether we should put human brain cells in sheep. Given our state of knowledge we should already be treating those sheep as if they had experiences (though being unaware of abstract concepts like death this does not mean we need to treat them like people).
In short *anything* is logically possible to cause conciouss experience but you don't worry that your computer might be suffering when you install the newest kind of software. If we want to make any choices we need to make the pragmatic assumption that morally relevant states can be deduced from objectively availible information otherwise we could never clip our toenails for fear they would feel pain.
Just injecting some human brain cells into a sheep or even transplanting an early human brain from a fetus is unlikely to produce any kind of human sheep. The human brain doesn't just develop from a genetic blueprint but also requires a huge amount of deveelopmental cues and responds to hormones and signaling molecules (like sonic the hedgehog) to develop properly. Not to mention a host of enviornmental stimuli needed to encourage the brain to wire itself correctly.
In short it isn't just human neurons which make us human but the whole brain development system at work in babies. This isn't the sort of thing which could be duplicated in a sheep without extensive genetic modification or hand controlling all the developmental signals. If this is possible at all it is far beyond our current level of technology.
So don't get freaked out yet people. They are just growing human neurons in sheep at the moment there is no chance we will make a person trapped in a sheep body.
God damn these popular stories can be misleading.
Ohh and in case anyone is curious there is a pretty efficent algorithm to calculate the n-th hexadecimal digit of pi without calculating any of the previous ones
Clearly pi is not random in any absolute sense. One can give an algorithm to calculate it. Of course if you don't have the benefits of a hardware random number generator you are stuck choosing some psuedo-random source but the idea that there is some universal hierarchy of psuedo-random algorithms is just downright idiotic.
Whether a particular psuedo-random algorithm is good or not depends quite heavily on what it is being used for. The requirements for a psuedo-random number for a monte-carlo simulation are much different than those needed for cryptography (in one you want an even distribution over some relevant space while in the other it is usually unpredictability that is desired). Moreover, even within one type of application the specific algorithm being used can make a huge difference. For instance if I am trying to use monte-carlo integration on some nice smooth polynomial the standard psuedo-random technique of feeding back the output of a mod operation to itself will work quite well but if you are trying to integrate a function defined by a similar sort of recursion on mods it may perform horribly.
So there simply can't be any answer of what psuedo-random algorithm is better or worse. Of course one might still ask what would be a good choice for the standard psuedo-random algorithm in an OS given most people's software choices. For this application you want some source that won't displat a pattern relevant to normal computation. Thus, since trigonometric and other pi related functions are very commonly used pi is probably a fairly dangerous choice even if it works well on test cases.
In other words any number or algorithm that like pi pops up in many computer applications is going to be a bad choice. Not because you can run some ridiculous test which will rank its 'randomness' but because it is much more likely to be relevant to the program someone is running. In other words pi is a bad pragmatic choice because someone is alot more likely to use monte-carlo techniques to calculate pi or something to do with pi than with some recurrent mod generator (or something more sophisticated but not so common).
Thus from a pragmatic point of view these tests are just dumb. They would be better off running an array of common software and seeing how it performs. From a mathematical point of view there is nothing to be said as all of these sources (except the hardware one) aren't random at all so the only question is pragmatic.
I dunno about the impact. I expect the iraqi insurgents already have a much better idea of how our security operations work than they can gain through this document. After all they have the opportunity to actually observe this operations and how they react to their attacks.
I expect the information was just concealed as a matter of course in case it might contain something of value to the enemy. Still though your overall point is valid. There is no reason the public really needs to know this stuff and it is better safe than sorry with information which might let them kill more soldiers.
Hmm, really? sendmail can't pass a switch to make sure it gets a response from the authoritative server not just a cache?
I wonder then why my MX records seem to work almost immediatly while other changes take much longer to propogate. Can MX records have different TTL than the domain itself? If so maybe that is the explanation.
Well that sounds like a good explanation for why major ISPs might ignore the TTL. That's alot of wasted lookups and server load. For a large ISP we can expect that someone is going to most Akamai cites every minute so that is at least one lookup per minute per Akamai DNS record.
If other people do this as well ISPs might not have much of a choice but to ignore TTL and maybe there isn't any easy way to force the machine to use a minimum TTL but respect TTL above that limit.
My understanding is that many mail delivery programs do their own recursive resolve and often ignore cached entries. At least when I have updated MX records they have seemed to propogate almost instantly. I presumed this was because sendmail and the like insist on reading the MX from servers authoritative for that domain but perhaps I'm mistaken.
Anyone care to clarify how DNS cache works with MX records?
Here it is with the link fixed
While this is certainly a disturbing development the constitutionality of it is unclear and quite interesting.
On the one hand the first ammendment certainly does not require libraries to provide pornographic magazines or otherwise provide some unbiased representation of viewpoints. In general the first ammendment does not restrict the government from providing some content but not others (except when this infringes on the establishment clause).
However, while library filtering has been deemed constitional (link here) the supreme court has also ruled that libraries must allow adults to bypass the filters. In other words apparently the supreme court has recognized that internet filtering is significantly different than buying library books. The library has legitimate financial constraints in what books it provides but does not in internet filtering.
So the question becomes very unclear in the case of truck stops. Since these are entierly automated they can't very well demand a librarian turn the filtering off. Still, since one does need to be at least 16 to drive and because of the real possibility that by providing enough government internet access filtering could stifle free speech I imagine it would be declared unconstitutional but it is a tough call.
While this is certainly a disturbing development the constitutionality of it is unclear and quite interesting.
On the one hand the first ammendment certainly does not require libraries to provide pornographic magazines or otherwise provide some unbiased representation of viewpoints. In general the first ammendment does not restrict the government from providing some content but not others (except when this infringes on the establishment clause).
However, while library filtering has been deemed constitional the supreme court has also ruled that libraries must allow adults to bypass the filters. In other words apparently the supreme court has recognized that internet filtering is significantly different than buying library books. The library has legitimate financial constraints in what books it provides but does not in internet filtering.
So the question becomes very unclear in the case of truck stops. Since these are entierly automated they can't very well demand a librarian turn the filtering off. Still, since one does need to be at least 16 to drive and because of the real possibility that by providing enough government internet access filtering could stifle free speech I imagine it would be declared unconstitutional but it is a tough call.
Sure, this may be the same thing that happened with Samba but linus never made any claims about the legality. Something which may be technically and legally similar is not necessarily just as good a strategic idea.
In the case of microsoft we had a widely deployed piece of software that the open source community needed to interact with for compatibility reasons. Nothing of the kind is true with BitKeeper. In the case of BitKeeper the open source community could have simply built their own incompatible protocol and not have to worry about being shut out of the market by a BitMover monopoly. On the other hand in the case of microsoft the open source community couldn't simply build a better protocol than Samab but really needed to be compatible.
Secondly, while it's possible I very much doubt that the BitKeeper protocol was being reverse engineered from the expensive pro version. Most likely it being reverse enginered from the free versions (or at least comped versions). Unlike microsoft which needs to keep Samba out there in every windows box BitMover was just allowing this free usage as a donation/PR move and could easily revoke it without comprimising their buisness model.
In short by trying to reverse engineer this protocol it seems that Andrew? gave the impression that the 'price' of donating expensive software to open source projects is to have your market advantages reverse engineered and probably implemented in free projects. So while sure he has just as much right to reverse engineer in samba the first instance is an important blow against a monopoly trying to use propietary protocols to unfairly strangle competition. In this case there was no similar monopolistic pressure (there isn't a strong installed base of BK users who we need to be compatible with) and made it look like there was a steep price for trying to help the open source community.
Regardless of what you think of the deciscion to use BK or the need to reverse engineer this project having someone paid by the SAME organization which is the beneficiary of the free software (or at least appears to be in the media) is surely a bad move politically. It certainly would give me pause if I was a manager at a big corporation thinking of donating some helpful development tools to some open source project.
Yes the key word here is 'comprimised set of device keys'
The way this worked in CSS and probably works similarly here is that at the begining to the disk they encrypt a disk key with many different device keys. Then each device decrypts the disk key using their own device key.
However if you work out the math it simply isn't plausible to include a seperate key for every HD DVD player that might ever be sold (imagine 128 bits for an AES key). Instead each manufacturer, or perhaps even DVD player model in this new system, gets one key. They can then 'revoke' these keys by just refusing to encrypt future DVD keys with these device keys but since each DVD player doesn't have its own key they can't disable movies player by player.
On another point I would find it to be really unlikely that any major DVD player would truly get this penalty imposed against it. It would be a huge loss to be the first movie that doesn't work on sony blah players so no movie company is going to be the one who takes that first step.
Instead this is really a measure to deter manufacturers from 'accidently' making their DVD players ignore copy protection or otherwise violate their rules. Thus it is likely to be used when a player first hits the market or not at all.
Really, so if in the 1800's people had managed to find a high density propelant they should have devoted a significant fraction of their resources to building colonies on the moon out of welded metal and rivets?
No, I think the better response would have been to devot that money to improving the underlying technology and they would actually get self sustaining colonies faster. I am arguing that we might actually make the colonies faster by devoting research money to the underlying technology that might let us make real self sustaining colonies rather than wasting money support people on mars when we know we aren't yet at the point of being able to make self-sustaining colonies.
This is kinda like the rendering dilema. Suppose you have a big scene to render should you immediatly buy a bunch of computers and start to run the job? Will that get it done the fastest supposing you only have Y dollars to spend on the job? Not necessarily if the rendering project would take really long on today's computers sometimes it takes less time just to invest that money and then buy new faster computers in a year which can finish the job before the original computers would have been able to do.
I'm arguing that this is the situation now in spaceflight (though to an even greater extreme since the people doing rendering don't usually have the choice of investing in processor research). It can actually get us self sustaining colonies faster by investing the money now in the underlying technologies than by wasting it sending people into space atop very expensive rockets.