When people's countries are occupied by others who believe they have some ancient right to lebensraum, you can't expect to be able to scold both sides and tell them to play nicely. This is life, not the playground.
The analogy is correct though. In the playground, 10 year olds fight for what they think it belongs to them, just like Israelis and Palestinians fight for what they think it belongs to them.
I do not know if it's a new data structure or not, but while I was searching the definition of "sparse voxel octrees" in order to understand what is the article about, I hit this article from 1993 that speaks about how octrees are one of the best ways to represent spatial data for 3d rendering applications:
I have the impression that sparse voxel octrees is not a new data structure at all, but it has been neglected by the industry because of technological limitations. Now that these issues are resolved, perhaps the next step in 3d graphics are sparse voxel octrees.
Generally operating systems do not support events on the file system (some do, but not all), so on some platforms when a new executable appears, it's the application that has the responsibility to inform the system.
How would you share your everyday reality with a girl that interprets anything through Star Signs or Feng Sui (for example)? you try to start a discussion on some serious topic, and the topic sooner than later is diverted to the shallow things that are in the interest of the girl. When you try to get back on topic, she gets a long face.
I am not saying it's only a girl thing...it might be a guy thing as well. When guys discuss sports and the women want to discuss something serious, then there is a problem.
Both participants of a relationship must be in the same mentality in order for a relationship to work out.
No, I am talking about something simple (and it's amazing people don't understand it!): in order to separate a flat address space to modules that can not touch each other unless the O/S allows it, put a module index in the page descriptor and a module table describing each module's access (by other modules) rights...then one module can't access another module, and the flat address space is not sacrificed.
A 32 bit page descriptor has 20 bits for the page frame, 1 bit for the page-present/non-present information, and 11 bits for anything else. 11 bits is enough to store an 11-bit index, i.e. a number from 0 to 2047. 2048 modules per process are more than enough.
When a program tries to access memory, the CPU checks the module id of the destination page against the module id of the current page. If the module id is the same, nothing happens. Otherwise, it checks the destination module access bits and raises an exception if the specific access (read, write, execute) is not allowed.
A module descriptor can be 4 bits: 3 bits for read/write/execute access, 1 bit custom data. The whole table of 2048 entries takes only 1024 bytes and can be cached fully into the CPU.
Finally, in order to protect from wild jumps, a special jump destination instruction should be the first instruction allowed after a jump. This prevents jumps into the middle of public subroutines.
It's simple, it's efficient, it requires very little electronics, it can be implemented very easily by the CPU designers, it takes 0 time to perform the checks, and it will protect each module from another module.
Different pages can belong to the same module; for example, the video driver can have its code at page FF0000 and can be the only module to access the video ram at address range AA0000 - BB0000.
The module with id = 0 can be the kernel module, i.e. it has access to everything else.
I forgot to add that it will be a commercial enterprise that will advance the state of the art on operating systems, not an open source initiative. I have never seen open source beeing pioneers. Open source usually reimplements what other open source/commercial enterprises have.
There is nothing that prevents Microsoft to adapt to the new reality. They can simply make a new O/S on an advanced programming language like Erlang which is easily parallelizable and provide an emulation layer to run its old O/Ses. The reason that they don't do it is because they don't need to, yet.
The trip to Mars or any other solar system body can easily be achieved if a spaceship is built with the following properties:
1) it uses rotation to simulate gravity. 2) it's big enough to host smaller craft that can visit a planetary body. The spaceship itself does not need to land. 3) it uses nuclear power both for its systems and for propulsion. 4) contains enough space to host a variety of human activities.
If such a thing existed, going to Mars would be very easy. It could take 1 year, but the spaceship would be comfortable and since gravity would be simulated, the astronauts would not have such a hard time. The cost of designing and building such a ship would be big, but not bigger than the cost of one year's money spent on military. Of course, it could be much easier to do if major countries co-operated on it.
The only reason people say "it can't be done" is that no one dares raise the finger against all those people that profit from war.
The other solution for the problems of microkernels which no one looks for is to add in-process modules to CPUs. It's a very cheap solution, the only thing required is to add a map of addresses to modules and the relevant cache (the flat address space need not be sacrificed). I don't understand why it is ignored by CPU designers...
It's the fault of CPU designers. Current ISAs do not provide in-process modules, and therefore O/S designers are forced to work around the problem in various ways.
I've posted a related comment in the recent discussion on Microkernels.
I also sent an email to Tanenbaum asking him to put a request for CPU designers to provide in-process modules (which can be done without sacrificing flat address spaces). He replied that the cost would be prohibitive...I strongly disagree: all that it requires is another cache of translations of memory addresses to modules...
So, in other words, your opinion is that we should all bent down and get fcsked over by those who control the media, right?
So, what you are saying is, in essence, that if a person X has more money/power than person Y then person X can dictate how person Y lives, what person Y says, what person Y thinks etc.
In other words, you are in favor of big brother from 1984, of Stalin, of Franco etc. You think dictatorship is the way to go.
All I can say is 'bravo'. We (as humanity) deserve what is ahead for us.
The most important engineering challenge is to make software correct by writing it. Software is the biggest thing in human history, because it enables God-like powers in us. Our civilization is as good as its software is. The future is dependent on software...
And why extremely inflammatory and derogatory comments should be censored? CNN could have said 'this guy works for us, but his opinions are not our opinions', and case closed. Why CNN had to fire him?
Being able to express any comment I want is my right, even if it is inflammatory and derogatory. If you don't like it, don't listen to me. You don't have the right to censor me whatsoever.
Slashdot Mirror
The analogy is correct though. In the playground, 10 year olds fight for what they think it belongs to them, just like Israelis and Palestinians fight for what they think it belongs to them.
I do not know if it's a new data structure or not, but while I was searching the definition of "sparse voxel octrees" in order to understand what is the article about, I hit this article from 1993 that speaks about how octrees are one of the best ways to represent spatial data for 3d rendering applications:
Volume Graphics
I have the impression that sparse voxel octrees is not a new data structure at all, but it has been neglected by the industry because of technological limitations. Now that these issues are resolved, perhaps the next step in 3d graphics are sparse voxel octrees.
Generally operating systems do not support events on the file system (some do, but not all), so on some platforms when a new executable appears, it's the application that has the responsibility to inform the system.
How would you share your everyday reality with a girl that interprets anything through Star Signs or Feng Sui (for example)? you try to start a discussion on some serious topic, and the topic sooner than later is diverted to the shallow things that are in the interest of the girl. When you try to get back on topic, she gets a long face.
I am not saying it's only a girl thing...it might be a guy thing as well. When guys discuss sports and the women want to discuss something serious, then there is a problem.
Both participants of a relationship must be in the same mentality in order for a relationship to work out.
Oh come on! I played DOOM III 2 years ago and my system was an Athlon 64 2.2 GHz and an GeForce 8!!! this system now has 400$ top.
Thanks. Who are the right people and how do I contact them?
It's not P2P's fault though. It's the people.
No, I am talking about something simple (and it's amazing people don't understand it!): in order to separate a flat address space to modules that can not touch each other unless the O/S allows it, put a module index in the page descriptor and a module table describing each module's access (by other modules) rights...then one module can't access another module, and the flat address space is not sacrificed.
A 32 bit page descriptor has 20 bits for the page frame, 1 bit for the page-present/non-present information, and 11 bits for anything else. 11 bits is enough to store an 11-bit index, i.e. a number from 0 to 2047. 2048 modules per process are more than enough.
When a program tries to access memory, the CPU checks the module id of the destination page against the module id of the current page. If the module id is the same, nothing happens. Otherwise, it checks the destination module access bits and raises an exception if the specific access (read, write, execute) is not allowed.
A module descriptor can be 4 bits: 3 bits for read/write/execute access, 1 bit custom data. The whole table of 2048 entries takes only 1024 bytes and can be cached fully into the CPU.
Finally, in order to protect from wild jumps, a special jump destination instruction should be the first instruction allowed after a jump. This prevents jumps into the middle of public subroutines.
It's simple, it's efficient, it requires very little electronics, it can be implemented very easily by the CPU designers, it takes 0 time to perform the checks, and it will protect each module from another module.
Different pages can belong to the same module; for example, the video driver can have its code at page FF0000 and can be the only module to access the video ram at address range AA0000 - BB0000.
The module with id = 0 can be the kernel module, i.e. it has access to everything else.
I forgot to add that it will be a commercial enterprise that will advance the state of the art on operating systems, not an open source initiative. I have never seen open source beeing pioneers. Open source usually reimplements what other open source/commercial enterprises have.
There is nothing that prevents Microsoft to adapt to the new reality. They can simply make a new O/S on an advanced programming language like Erlang which is easily parallelizable and provide an emulation layer to run its old O/Ses. The reason that they don't do it is because they don't need to, yet.
The trip to Mars or any other solar system body can easily be achieved if a spaceship is built with the following properties:
1) it uses rotation to simulate gravity.
2) it's big enough to host smaller craft that can visit a planetary body. The spaceship itself does not need to land.
3) it uses nuclear power both for its systems and for propulsion.
4) contains enough space to host a variety of human activities.
If such a thing existed, going to Mars would be very easy. It could take 1 year, but the spaceship would be comfortable and since gravity would be simulated, the astronauts would not have such a hard time. The cost of designing and building such a ship would be big, but not bigger than the cost of one year's money spent on military. Of course, it could be much easier to do if major countries co-operated on it.
The only reason people say "it can't be done" is that no one dares raise the finger against all those people that profit from war.
The other solution for the problems of microkernels which no one looks for is to add in-process modules to CPUs. It's a very cheap solution, the only thing required is to add a map of addresses to modules and the relevant cache (the flat address space need not be sacrificed). I don't understand why it is ignored by CPU designers...
I hope we can play DNF before the mass extinction event...
It's the fault of CPU designers. Current ISAs do not provide in-process modules, and therefore O/S designers are forced to work around the problem in various ways.
I've posted a related comment in the recent discussion on Microkernels.
I also sent an email to Tanenbaum asking him to put a request for CPU designers to provide in-process modules (which can be done without sacrificing flat address spaces). He replied that the cost would be prohibitive...I strongly disagree: all that it requires is another cache of translations of memory addresses to modules...
It's good for Microsoft to invent such a machine. Now PCs may run Vista, at last. ...wait a minute...
No, we are at the end of the start times. Computer/IT technology is just leaving the state of puberty.
...simple: shoot it down so as that it lands on Earth.
But if it wasn't for such a case, we wouldn't have the Dukes of Hazzard, would we?
It depends on the name...if it was Tau Booty or Tau Boobtis, I think many slashdotters would relocate...
The best way to eradicate the problem of religion is education.
Communism was like a religion to Stalin and the other USSR leaders.
So, in other words, your opinion is that we should all bent down and get fcsked over by those who control the media, right?
So, what you are saying is, in essence, that if a person X has more money/power than person Y then person X can dictate how person Y lives, what person Y says, what person Y thinks etc.
In other words, you are in favor of big brother from 1984, of Stalin, of Franco etc. You think dictatorship is the way to go.
All I can say is 'bravo'. We (as humanity) deserve what is ahead for us.
You forgot one...Duke Nukem Forever to be released.
The most important engineering challenge is to make software correct by writing it. Software is the biggest thing in human history, because it enables God-like powers in us. Our civilization is as good as its software is. The future is dependent on software...
And why extremely inflammatory and derogatory comments should be censored? CNN could have said 'this guy works for us, but his opinions are not our opinions', and case closed. Why CNN had to fire him?
Being able to express any comment I want is my right, even if it is inflammatory and derogatory. If you don't like it, don't listen to me. You don't have the right to censor me whatsoever.