What I want from OS virtualization is to be able to run one guest OS on multiple hosts for redundency.
Then what you want is Marathon style lock-style execution. It's a terribly hard problem because you have to make all software run deterministically (timers and IO events on both machines have to occur in the *exact* same moments in execution).
The chip present in the Thinkpads is a TPM. This chip is basically just a tamper-resistent crypto engine. One of the uses of this is attestation. You can crypto graphically sign your kernel, and the system will make sure to only run a properly signed kernel.
This is hugely useful from a security perspective. Imagine you're machine gets broken into. Any admin knows the only proper thing to do is completely reimage the machine right (there's no guarentee that they didn't install a root kit even if you remove the piece of software that led to the breakin). Well, with attestation, you do now know that the kernel is okay. This lets you start building up trust. If the kernel only loads signed modules, then you know all modules are okay. If the kernel goes further and only lets root run signed software...
This is why the TPMs exist. The entertain industry isn't all that big. There is a *huge* market for security though. Can you use a TPM to enforce DRM? Sure. Honestly though, this is only a problem if you're using DRM crippled media on a closed-source platform. Why the hell are you doing that in the first place?
I'm on my way down to the local public library right now, with my Powerbook and a page scanner.
You have the legal right to check out a book at the library. You have a legal right to scan the book. You do not have the right to access the scanned book after returning the library book.
As long as google employees don't have unfettered access to the scanned books, I think they're okay.
I don't disagree with laws against kiddie porn (which, when you filter , I just disagree with people who think that supply creates demand when supply FILLS demand (economics 101... hell, more basic than that... it's high school.)
Okay, my wording sucks. The point is, the existance of a demand will lead to an increase in supply. One doesn't create "demand" (you could argue that marketing does...).
I honestly don't know enough about what constitutes computer-generated... Couldn't you have something that looked like a child and yet claim that it's really an adult and just happens to look young? The precedent scares me because this is such a gray area.
1. This verdict is absolutely crap. COPYING child porn is not the same as CREATING NEW child porn. No children are harmed by such an act.
This would be an argument that child pornography should not be illegal. From an economic perspective though, if people are consuming child porn, they are creating a demand which is going to increase the supply. This is why child porn is illegal to possess--it indirectly contributes to more children being exploited for it (the general wisdom being that the vast majority of child pornography is exploitative--if not all of it).
2. Submitter -- Why is he a *scumbag* pedophile? People generally don't choose what and who they're attracted to. It is not illegal to be attracted to children. It is only illegal to act on it. Provided that he doesn't, he can still be a good man in my book.
You choose to act upon impulses though. I often have an impulse to smash stupid people's heads in, but I control it. You may say "he's only consuming so it's not destructive..." but then see my response to #1.
3. Laws against pedophiles (not against pedophiliac acts -- pedophiles) are counter-productive. For example, it is illegal to create computer-generated child pornography. Why!? Provided that it gives people who are into such things a release, and no children are harmed, I have no problem with it. Many of today's sex-related laws are based on some twisted idea of morality, and nothing more.
The reasoning behind computer-generated child porn is #1. It creates a demand...
It's certainly true that the current laws are curious. Even the most softcore porn featuring a 17 year old is illegal, yet incredibly hardcore material from the follow day that she turned 18 is legal... Strange.
This judge got out of hand. If the guy has a descent lawyer, it'll be appealed. The guy may end up walking which would be sad.
I'm still wondering how Google accepting China's demands to do business with their country is any different than Amazon bowing to Germany by filtering out Nazi paraphenalia. Each country has its own definition of unacceptable material. The USA (and virtually every other country on earth) believes child porn is taboo. But it's still a choice that was made, to filter and prosecute those who deal with it.
In the US, search engines are *not* required to filter child porn. In fact, various bills that have attempted to make such rules have been ruled unconstitional.
One thing to keep in mind, is that in the case of Germany and the US, the majority of the people believe that these things should be censored. Thankfully, in the US, at least portions of the constitution still trumph popular opinion. In China, it's the *government* that wants to restrict information. It's not the majority of the people.
I noticed this this morning and quickly realized what had happened. I liked things the way they used to be more so I looked in my preferences to see how to disable it. I didn't recall seeing those boxes before but didn't realize that's what you use to disable it.
It would have been nice to have a simple, textual, option like "Disable section stories in main page" or something like that instead of 6 different pairs of oddly shaped boxes.
What's more, now that I disabled the inline section stories, I no longer has "2 more stories.." Perhaps that feature can be added back when section stories are disabled?
It is true that blacks haven't been prevalent among the programmers, but I think isn't explained by racism. This is a very controversial view, but supported along the following reasoning: various things, like height or reaction time are different among different ethnic groups. To the extent that IQ is a hereditable phenomenon, different groups have different average IQs, just as we have different average heights. That explains the outcomes. This is especially true, given all the money the government has poured into trying to get brown and black people to succeed in science (and why didn't they pour money into helping Chinese succeed in basketball?).
While I do have moderator points, a statement like this is so absurd that I just have to respond instead of modding down. It is very simply the product of under education.
The genetic differences (which are very quantifiable) between racial groups is very tiny compared to the variation that occurs within any one racial group. This means, quite simply, that the genetic different between you and "blacks" is equal to (or in many cases, even less than) other "whites". Moreover, IQ has nothing to do with actual success and is actually a pretty poor gauge of intelligence (it is designed to be a test for determining retardation, btw).
So why are most programmers white? Well, I don't know where you work, but that's certainly not the case where I work. There are a large number of people of almost every descent (african--and I mean, from africa, asian, american, european, etc.).
There are many misconceptions about equality when it comes to "black" Americans. This largely has to due with the fact that they are, in fact, a minority (~12%) but tend to live in concentrated areas where they make up more than 50% of the population. This means that the distribution in the general population tends to be even smaller.
The concentration of the population is certainly a relic of 20th century discrimination. I'm not saying discrimation is still rampant today, but it will take a few generations for the population to disburse accordingly.
Think of it this way, wouldn't you want to live among people similiar to you if you were systematically discriminated against?
Please think in the future before you say something like this in public. Not only do you come off sounding like a jack-ass, but you spread a horribly untrue fact that only feeds into negative past behaviors.
The article didn't say much about the damage or how it was ascertained that it was not a postmortem marking from scavangers
A bird of prey swooping at full speed, claw first on the top of the head, is going to suggest it was the source of death. That sort of crushing blow would be hard to deliver if the person was dead (and presumably on the ground). Not to mention, what's the point after death of delivering such a blow?
Yes, its speculation, and can never be proven, but that's the basis of a lot of science.
Re:Does anyone else here actually agree with DRM?
on
A Look at Google DRM
·
· Score: 1
I honestly have no problem with basic (ie non-Sony rootkit) DRM. These people create content in which we consume. In a capitalist society they can choose to deliver the content in whatever means they choose.
No. No. No. No.
DRM should *not* be able to exist legally. It only exists because congress let it exist by introducing the DMCA in the late 90s.
Prior to the DMCA, I could (legally) reverse engineer an iTunes music file, and write a player for Linux. The world rejoyces. I could then purchase iTunes music files, and play them in Linux (which is a fair use of the product I purchase). Say what you want about EULA's, but you cannot sign an illegal contract. I cannot give up my fair use.
The DMCA (by limiting circumvention of copyright mechanisms) prevents me from being able to (legally) do this. It prevents me from exercising my fair use. What's the world without fair use? There's no google (indexing the sites and providing a preview is a fair-use quotation, what happens when we have DRM web pages?). There are no research papers (quotations are a fair-use). There are no satires (again fair use).
DRM is *not* and has *never* been about piracy. If I'm pirating (which is illegal) I have no problem using illegal software (or hardware) to do it. DRM is about making fair-use illegal. It's about locking out competitors (how many players can play iTunes songs? how many players can play mp3s? Funny eh?). DRM is anti-competitive. Competitive economies are necessary for a healthy capitalistic economy. The government's job is to create a healthy economy.
I don't see why we applaud that type of advice when it comes to grades, but heap scorn upon it when it comes to software development.
It really, really, depends here. GPA is widely accepted to not be a good indicator of corporate success (forget about exceptional cases, I'm talking about the average case).
It's somewhat irresponsible to say that GPA doesn't matter (because it does), but it OP's point, I think, is that sometimes one's time can be better spent pursuing other things instead of focusing on getting the highest possible GPA (which I do believe is true).
But if someone was able to create some form of DRM that prohibited truly illegal actions, but allowed all legal ones
What you don't understand here is that this sort of DRM is not possible from a theoritical computer science perspective. You cannot prevent someone from decoding data with technology. The only thing you can do is make it illegal (see the DMCA).
The problem is there are lots of legitimate things one should be able to do that interfere with this. I should be able to reverse engineer any DRM system but that would be illegal.
It is theorised that a "half-dose" of homosexuality might be a positive evolutionary advantage if it leads to extra-caring behaviour that results in more children growing up and providing the next generation. There are stats to show this "uncle" effect, and stats that deny it!
The apparently (perhaps) paradox is that for that half-dose to be passed down would first require copulation. Without treatment for sickle-cell anemia, one will suffer and likely die young (from complications) but will likely still be able to reproduce (although not as prolifically as one without the disorder). If there were a homosexuality gene, one would assume that the likelihood of reproducing at all would be so low that the half-dose advantage would have to be huge.
I'm not saying it isn't genetic, but I just still don't understand...
There is plenty of research demonstrating the likely genetic basis of homosexuality
I find this so hard to understand. If there are genetic reasons for homosexuality, then people with that gene should breed significantly less than people without that gene until such a point in time where that gene is eliminated from the population.
The fact that homosexuality has been observed in so many species would seem to be contrary to the above observation (it should be a rare mutation that really shouldn't enter the main gene pool).
What's worse, this sort of genetic reasoning would suggest that homosexuals are genetically inferior (if they carry a trait that makes them less likely to spread their own genetic material) which really seems like a bad thing to suggest.
Isn't it better for everyone to not pretend like homosexuality is genetic? I know everyone wants everything to be a dieasese or genetic so that they don't have to except the reasonability for it but in this case I think it's dangerous to suggest such things.
which is what my example did, that does look interesting. It also shows why numbered positional parameters are probably a bad idea, if they require you to trade terseness for readability.
You can define the placeholders to have any name you want, but I can't think of a convient scoping construct other than maybe the GNUism of ({}). You could have something like lambda(x, y, {y - x}) which would be a pretty reasonable syntax that would expand (via CPP magic) to ({boost::lambda::placeholder x; boost::lambda::placeholder y; {y - x}}).
Can you imagine what a C++ lambda construct would have to look like?
(lambda (x y) (- y x))
would turn into a mess like:
int lambda(int x, int y) { return y - x; }
Actually, it's a bit more elegant than lisp *gasp*:
_1 - _2
See the Boost Lambda library. Boost is the C++ library working group. Most of the library enhancements have come from Boost. Lambda's have been around for a while but aren't a part of the current standardization effort.
They work by having a set of placeholder variables that act as parasitic types so that the above expressions turns into an instance of a templated type who's function operation will evaluate the given expression. Really cool stuff.
And just how big is the rock that the rest of these people are living under?
I think e-mail is slowly dying actually. A lot of kids don't really use it instead preferring to use instant messenger. If a kid has their own computer with broadband access, that stays on 24x7, why not just use IM?
Spam really hurts the usefulness of e-mail for a lot of users. Personally, I've reduced my personal e-mail account to just notification from various things I'm involved in (school, bills, etc.).
The thing I like so much about IM is that only people who I've explicitly allowed to contact me can actually contact me. This means no spam. With logging, and grep, it's just as useful a communications history as e-mail.
MS doesnt make their money just out of selling games (and I seriously doubt they LOSE money on each Console sale as they claim)
People really don't understand this well at all. Developing the Xbox required a very large up front investment. To justify the investment, Microsoft will analyze how much they expect to sell, and amortize that cost over the consoles and games.
Clearly, there business model is such that if they only sold consoles, and not games, they would not recoop their costs. This makes sense because the relative profit margins on games are going to be much higher than on the consoles. This is how they "lose" money on the Xbox. The sum of the components, however, are not worth more than what they sell it for. Otherwise, people would just buy a bunch and sell it for scrap. If you buy Xbox's, Microsoft does not lose money. It's quite the opposite, every Xbox and game they sell gets them closer to recooping the original investment and eventually making quite a profit.
I was told once that Coca-Cola's CEO has a rule that noone can present a market analysis to him that shows Coca-Cola as having more than %34 of the market. When asked how Coca-Cola could possible grow their customer base he replied, "The average human consumes 64 fluid ounces a day--only 6 ounces of that is a Coca-Cola product."
You can always redefine markets to make a monopolostic company not appear so.
Slashdot Mirror
:-P
What I want from OS virtualization is to be able to run one guest OS on multiple hosts for redundency.
Then what you want is Marathon style lock-style execution. It's a terribly hard problem because you have to make all software run deterministically (timers and IO events on both machines have to occur in the *exact* same moments in execution).
Regards,
Anthony Liguori
The chip present in the Thinkpads is a TPM. This chip is basically just a tamper-resistent crypto engine. One of the uses of this is attestation. You can crypto graphically sign your kernel, and the system will make sure to only run a properly signed kernel.
This is hugely useful from a security perspective. Imagine you're machine gets broken into. Any admin knows the only proper thing to do is completely reimage the machine right (there's no guarentee that they didn't install a root kit even if you remove the piece of software that led to the breakin). Well, with attestation, you do now know that the kernel is okay. This lets you start building up trust. If the kernel only loads signed modules, then you know all modules are okay. If the kernel goes further and only lets root run signed software...
This is why the TPMs exist. The entertain industry isn't all that big. There is a *huge* market for security though. Can you use a TPM to enforce DRM? Sure. Honestly though, this is only a problem if you're using DRM crippled media on a closed-source platform. Why the hell are you doing that in the first place?
I'm on my way down to the local public library right now, with my Powerbook and a page scanner.
You have the legal right to check out a book at the library. You have a legal right to scan the book. You do not have the right to access the scanned book after returning the library book.
As long as google employees don't have unfettered access to the scanned books, I think they're okay.
I don't disagree with laws against kiddie porn (which, when you filter , I just disagree with people who think that supply creates demand when supply FILLS demand (economics 101... hell, more basic than that... it's high school.)
Okay, my wording sucks. The point is, the existance of a demand will lead to an increase in supply. One doesn't create "demand" (you could argue that marketing does...).
Do you agree with this reasoning?
I honestly don't know enough about what constitutes computer-generated... Couldn't you have something that looked like a child and yet claim that it's really an adult and just happens to look young? The precedent scares me because this is such a gray area.
1. This verdict is absolutely crap. COPYING child porn is not the same as CREATING NEW child porn. No children are harmed by such an act.
This would be an argument that child pornography should not be illegal. From an economic perspective though, if people are consuming child porn, they are creating a demand which is going to increase the supply. This is why child porn is illegal to possess--it indirectly contributes to more children being exploited for it (the general wisdom being that the vast majority of child pornography is exploitative--if not all of it).
2. Submitter -- Why is he a *scumbag* pedophile? People generally don't choose what and who they're attracted to. It is not illegal to be attracted to children. It is only illegal to act on it. Provided that he doesn't, he can still be a good man in my book.
You choose to act upon impulses though. I often have an impulse to smash stupid people's heads in, but I control it. You may say "he's only consuming so it's not destructive..." but then see my response to #1.
3. Laws against pedophiles (not against pedophiliac acts -- pedophiles) are counter-productive. For example, it is illegal to create computer-generated child pornography. Why!? Provided that it gives people who are into such things a release, and no children are harmed, I have no problem with it. Many of today's sex-related laws are based on some twisted idea of morality, and nothing more.
The reasoning behind computer-generated child porn is #1. It creates a demand...
It's certainly true that the current laws are curious. Even the most softcore porn featuring a 17 year old is illegal, yet incredibly hardcore material from the follow day that she turned 18 is legal... Strange.
This judge got out of hand. If the guy has a descent lawyer, it'll be appealed. The guy may end up walking which would be sad.
where Peter annexes his next door neighbors pool and he gets letters of praise from Serbia, Iran, Iraq, etc.
I wonder if we'll see anyone resigning at Google in protest...
I'm still wondering how Google accepting China's demands to do business with their country is any different than Amazon bowing to Germany by filtering out Nazi paraphenalia. Each country has its own definition of unacceptable material. The USA (and virtually every other country on earth) believes child porn is taboo. But it's still a choice that was made, to filter and prosecute those who deal with it.
In the US, search engines are *not* required to filter child porn. In fact, various bills that have attempted to make such rules have been ruled unconstitional.
One thing to keep in mind, is that in the case of Germany and the US, the majority of the people believe that these things should be censored. Thankfully, in the US, at least portions of the constitution still trumph popular opinion. In China, it's the *government* that wants to restrict information. It's not the majority of the people.
I noticed this this morning and quickly realized what had happened. I liked things the way they used to be more so I looked in my preferences to see how to disable it. I didn't recall seeing those boxes before but didn't realize that's what you use to disable it.
It would have been nice to have a simple, textual, option like "Disable section stories in main page" or something like that instead of 6 different pairs of oddly shaped boxes.
What's more, now that I disabled the inline section stories, I no longer has "2 more stories.." Perhaps that feature can be added back when section stories are disabled?
There is less genetic difference between dogs, wolves and coyotes, three different species, than there is between one human race and another.
I call your bluff. Show me a resource to support that.
It is true that blacks haven't been prevalent among the programmers, but I think isn't explained by racism. This is a very controversial view, but supported along the following reasoning: various things, like height or reaction time are different among different ethnic groups. To the extent that IQ is a hereditable phenomenon, different groups have different average IQs, just as we have different average heights. That explains the outcomes. This is especially true, given all the money the government has poured into trying to get brown and black people to succeed in science (and why didn't they pour money into helping Chinese succeed in basketball?).
While I do have moderator points, a statement like this is so absurd that I just have to respond instead of modding down. It is very simply the product of under education.
The genetic differences (which are very quantifiable) between racial groups is very tiny compared to the variation that occurs within any one racial group. This means, quite simply, that the genetic different between you and "blacks" is equal to (or in many cases, even less than) other "whites". Moreover, IQ has nothing to do with actual success and is actually a pretty poor gauge of intelligence (it is designed to be a test for determining retardation, btw).
So why are most programmers white? Well, I don't know where you work, but that's certainly not the case where I work. There are a large number of people of almost every descent (african--and I mean, from africa, asian, american, european, etc.).
There are many misconceptions about equality when it comes to "black" Americans. This largely has to due with the fact that they are, in fact, a minority (~12%) but tend to live in concentrated areas where they make up more than 50% of the population. This means that the distribution in the general population tends to be even smaller.
The concentration of the population is certainly a relic of 20th century discrimination. I'm not saying discrimation is still rampant today, but it will take a few generations for the population to disburse accordingly.
Think of it this way, wouldn't you want to live among people similiar to you if you were systematically discriminated against?
Please think in the future before you say something like this in public. Not only do you come off sounding like a jack-ass, but you spread a horribly untrue fact that only feeds into negative past behaviors.
The article didn't say much about the damage or how it was ascertained that it was not a postmortem marking from scavangers
A bird of prey swooping at full speed, claw first on the top of the head, is going to suggest it was the source of death. That sort of crushing blow would be hard to deliver if the person was dead (and presumably on the ground). Not to mention, what's the point after death of delivering such a blow?
Yes, its speculation, and can never be proven, but that's the basis of a lot of science.
I honestly have no problem with basic (ie non-Sony rootkit) DRM. These people create content in which we consume. In a capitalist society they can choose to deliver the content in whatever means they choose.
No. No. No. No.
DRM should *not* be able to exist legally. It only exists because congress let it exist by introducing the DMCA in the late 90s.
Prior to the DMCA, I could (legally) reverse engineer an iTunes music file, and write a player for Linux. The world rejoyces. I could then purchase iTunes music files, and play them in Linux (which is a fair use of the product I purchase). Say what you want about EULA's, but you cannot sign an illegal contract. I cannot give up my fair use.
The DMCA (by limiting circumvention of copyright mechanisms) prevents me from being able to (legally) do this. It prevents me from exercising my fair use. What's the world without fair use? There's no google (indexing the sites and providing a preview is a fair-use quotation, what happens when we have DRM web pages?). There are no research papers (quotations are a fair-use). There are no satires (again fair use).
DRM is *not* and has *never* been about piracy. If I'm pirating (which is illegal) I have no problem using illegal software (or hardware) to do it. DRM is about making fair-use illegal. It's about locking out competitors (how many players can play iTunes songs? how many players can play mp3s? Funny eh?). DRM is anti-competitive. Competitive economies are necessary for a healthy capitalistic economy. The government's job is to create a healthy economy.
I don't see why we applaud that type of advice when it comes to grades, but heap scorn upon it when it comes to software development.
It really, really, depends here. GPA is widely accepted to not be a good indicator of corporate success (forget about exceptional cases, I'm talking about the average case).
It's somewhat irresponsible to say that GPA doesn't matter (because it does), but it OP's point, I think, is that sometimes one's time can be better spent pursuing other things instead of focusing on getting the highest possible GPA (which I do believe is true).
But if someone was able to create some form of DRM that prohibited truly illegal actions, but allowed all legal ones
What you don't understand here is that this sort of DRM is not possible from a theoritical computer science perspective. You cannot prevent someone from decoding data with technology. The only thing you can do is make it illegal (see the DMCA).
The problem is there are lots of legitimate things one should be able to do that interfere with this. I should be able to reverse engineer any DRM system but that would be illegal.
I guess it means specifically that any non-copy-protected content will be DRM free, and copy protected stuff will have DRM. :(
Yup. Google == evil.
Google ships DRM.
DRM is evil.
Therefore Google is doing evil.
Liars.
It is theorised that a "half-dose" of homosexuality might be a positive evolutionary advantage if it leads to extra-caring behaviour that results in more children growing up and providing the next generation. There are stats to show this "uncle" effect, and stats that deny it!
The apparently (perhaps) paradox is that for that half-dose to be passed down would first require copulation. Without treatment for sickle-cell anemia, one will suffer and likely die young (from complications) but will likely still be able to reproduce (although not as prolifically as one without the disorder). If there were a homosexuality gene, one would assume that the likelihood of reproducing at all would be so low that the half-dose advantage would have to be huge.
I'm not saying it isn't genetic, but I just still don't understand...
There is plenty of research demonstrating the likely genetic basis of homosexuality
I find this so hard to understand. If there are genetic reasons for homosexuality, then people with that gene should breed significantly less than people without that gene until such a point in time where that gene is eliminated from the population.
The fact that homosexuality has been observed in so many species would seem to be contrary to the above observation (it should be a rare mutation that really shouldn't enter the main gene pool).
What's worse, this sort of genetic reasoning would suggest that homosexuals are genetically inferior (if they carry a trait that makes them less likely to spread their own genetic material) which really seems like a bad thing to suggest.
Isn't it better for everyone to not pretend like homosexuality is genetic? I know everyone wants everything to be a dieasese or genetic so that they don't have to except the reasonability for it but in this case I think it's dangerous to suggest such things.
which is what my example did, that does look interesting. It also shows why numbered positional parameters are probably a bad idea, if they require you to trade terseness for readability.
You can define the placeholders to have any name you want, but I can't think of a convient scoping construct other than maybe the GNUism of ({}). You could have something like lambda(x, y, {y - x}) which would be a pretty reasonable syntax that would expand (via CPP magic) to ({boost::lambda::placeholder x; boost::lambda::placeholder y; {y - x}}).
Can you imagine what a C++ lambda construct would have to look like?
(lambda (x y) (- y x))
would turn into a mess like:
int lambda(int x, int y) { return y - x; }
Actually, it's a bit more elegant than lisp *gasp*:
_1 - _2
See the Boost Lambda library. Boost is the C++ library working group. Most of the library enhancements have come from Boost. Lambda's have been around for a while but aren't a part of the current standardization effort.
They work by having a set of placeholder variables that act as parasitic types so that the above expressions turns into an instance of a templated type who's function operation will evaluate the given expression. Really cool stuff.
And just how big is the rock that the rest of these people are living under?
I think e-mail is slowly dying actually. A lot of kids don't really use it instead preferring to use instant messenger. If a kid has their own computer with broadband access, that stays on 24x7, why not just use IM?
Spam really hurts the usefulness of e-mail for a lot of users. Personally, I've reduced my personal e-mail account to just notification from various things I'm involved in (school, bills, etc.).
The thing I like so much about IM is that only people who I've explicitly allowed to contact me can actually contact me. This means no spam. With logging, and grep, it's just as useful a communications history as e-mail.
MS doesnt make their money just out of selling games (and I seriously doubt they LOSE money on each Console sale as they claim)
People really don't understand this well at all. Developing the Xbox required a very large up front investment. To justify the investment, Microsoft will analyze how much they expect to sell, and amortize that cost over the consoles and games.
Clearly, there business model is such that if they only sold consoles, and not games, they would not recoop their costs. This makes sense because the relative profit margins on games are going to be much higher than on the consoles. This is how they "lose" money on the Xbox. The sum of the components, however, are not worth more than what they sell it for. Otherwise, people would just buy a bunch and sell it for scrap. If you buy Xbox's, Microsoft does not lose money. It's quite the opposite, every Xbox and game they sell gets them closer to recooping the original investment and eventually making quite a profit.
I was told once that Coca-Cola's CEO has a rule that noone can present a market analysis to him that shows Coca-Cola as having more than %34 of the market. When asked how Coca-Cola could possible grow their customer base he replied, "The average human consumes 64 fluid ounces a day--only 6 ounces of that is a Coca-Cola product."
You can always redefine markets to make a monopolostic company not appear so.