Better still, a user could modify uber-secure-linux so that it runs both the original uber-secure-signed items PLUS things signed by their own key. This would still not require knowing the uber-secure signing key, and they could still reuse the protection to make sure only their own stuff or uber-secure stuff is run. Or they could edit out the uber-secure key so only their stuff will run. Or they could edit out all the protection so everything will run.
The difference from DRM is that the user can do this. It does not matter if uber-secure-company says "that machine is not secure". They can do that, certainly, but if you are uninterested in whether uber-secure cares, you can still make the modifications.
That's not DRM. You are quite capable of running your modified version of the software without RedHat's keys. Anybody else would be able to run it as well. All it means is that when they look at your version, they can tell it is not RedHat's approved version, which is the entire purpose of their signing.
Now if RedHat made their own Linux box that refused to run the software without the correct signature, and provided this copy of Linux with it, that would be DRM. Under the GPL3 they would have to provide some way for you to sign your own modified copy.
I'm a little confused about this, though. If RedHat made a new Linux box with the Linux software in ROM that was soldered to the board and imbedded in a block of epoxy, it would be pretty much impossible to replace the software then, as well. And nobody sees anything wrong with this as it may be a very sensible and practical construction method. I think the GPL3 is trying to remove artificial restrictions to replacing the software.
I know you think you are making some kind of profound statement, but the terrible state of OSS documentation just reinforces this.
Despite the fact that you can literally cut & paste the code into your application (not allowed with Microsoft's code) it is obvious that availability of source code is almost useless to interoperability. OSS does a great job of implemented documented standard interfaces, such as HTML and network protocols, but it is obvious that the ability of one piece of OSS to talk to another using some new protocol invented by that other one is very limited and slow. Witness the fact that most non-project programs don't install desktop shortcuts or icons on KDE and Gnome, this is due to the inability to figure out how using the code and the lack of documentation (there was documentation for the directory used by Gnome and enough programs used that that Gnome has to be compatable with it, but nothing I can find for how to use their new registry-like stuff).
I would say that the state of OSS proves conclusively that the source cannot be used and documentation is needed. Now I realize that their stuff probably is not documented and they don't want to try to write it (not any more than any OSS project wants to write documentation), so I think the EU and MS are going to have to compromise by them releasing into the public domain sufficient excerpts from their code to make a program that interoperates. That will at least get things to the cut & paste level.
The trick with subpixel rendering is that it is *not* just a 3x higher resolution antialiased image. In fact the hinting has gone through some trouble to arrange things to be 3 subpixels wide. This means that the same number of r,g,b subpixels are turned on as they would with normal antialiasing.
If the user does not percieve the colors equal to what is expected, this would also change their perception of pure white (also made by summing r,g,b), thus changing the entire white balance of the monitor. If this annoys them they would twiddle the color settings to fix it.
You may be worried about non-linear perception of the color levels, with different responses for different colors. However ClearType does not take into account the non-linear sRGB display curve, and this error would swamp any problem in your eyes by an order of magnitude or more, and even that does not seem to be too objectionable.
So basically the above poster is full of it. Don't believe him.
I was under the impression that ClearType is not just the LCD sub-pixel rendering, but the entire "new" antialiasing scheme, designed to replace "Font Smoothing" and useful on both CRTs and LCDs. I don't have any new Windows machines here to test it on, but I remember trying this in a store and the control panel certainly had the ability to turn ClearType on/off that was *seperate* from a control that said whether to use the LCD or not. It is possible this second control has been removed because they have a way to determine this directly from the monitor information. In any case I thought ClearType, even in grayscale, was a big improvement over Font Smoothing (which was also available), and looked more like the antialiasing on OS/X and Linux, probably because it is the same algorithim.
I would certainly expect Microsoft to turn ClearType on at all times if it works the way I think. Although people get programmed to like the way things look and thus complain about "blurry" fonts when they see any gray pixels, the average non-full-time computer user greatly prefers the ClearType output, and so will everybody once they get used to it.
OS/X has antialiasing on at all times, how come nobody complains about it?
No, wrong. The antialiasing selection is done by configurtion files, not the fonts themselves. The fonts contain no inforamtion that says "I can be antialiased but should not be". You may be confused by bitmapped fonts, which cannot be antialiased. Or more likely, Microsoft has a list somewhere of what fonts can be antialiased and you have not found it yet or it is inaccessible through the control panel.
Just want to back up the other responders. I also have never heard "reverse engineering" as being decompilation like you claim. "reverse engineering" is testing the original, inventing new tests to determine things the original tests did not, and producing a copy that outputs the same results as the tests.
Decompilation is genearlly called "cloning" if done legally, maybe cracking/hacking if done illegally. I think legally doing it requires walled-off development groups, such as was done for the BIOS. Decompilation is quite impractical for something the size of Windows, and as you mentioned the MS documentation and many programmers familiar with the platform give a huge head start and make reverse-engineering the far-preferred method.
My god did you even read? I fully admit that the power modes and user-mode drivers *are* improvements to the OS. I was just trying to differentiate between hype and the real thing, something you seem incapable of.
I can guarantee you that if Linux added support for this LCD screen or the remote projectors (something it cannot do without the hardware to test it on) then it will NOT be called Linux 3.0, but still be 2.6. It is NOT a new OS!
Hey I dislike Microsoft as much as the next guy, but saying it is in their interest to make it insecure is silly. If they want to, they can force people to upgrade by adding functionality or changing things so older systems can interface with new systems.
I'm sure they would love to make it secure, as the average person does not know anything wrong with Microsoft except "it's got security bugs". A fixed version would in fact remove 100% of the negative things the average person knows about Microsoft.
- A secondary LCD screen that offers limited, PDA-like functionality while the main OS is in Suspend. How would you do that with XP? Or Linux? It's not anything you can write in user mode.
What the hell are you talking about. That LCD would be a *second screen*. Windows as-is already supports multiple screens, so add a driver for the LCD! I don't consider that a change to the OS. A slight change so that the user can make a program go to that screen without dragging it is, I guess, a change to the OS, but hardly a huge addition. Of course I think MS has gone and made yet another api, totally failing to realize that the existing one was just fine. That is why you seem confused into thinking supporting this new screen is some new OS feature. Sigh.
A little circle appears, kind of like a radar screen, and it figures out which icon you meant to press. Current touchscreens can't do that.
WTF? Last I looked current touchscreens can draw anything they want on it. So I guess they improved the driver, but I suspect they could have put this into XP just as easily.
Power modes
This I would say *is* a feature of Vista as it probably touches lots of different parts of it. It is not a driver like the above two.
The LCD projector thing -- that's not OS?
Um, NO! The LCD projector thing is "hardware in the projector to pick up the wireless signal!" Are you claiming that this could not be done with existing Windows if a hardware manufacturer built such a device?
user-mode drivers
This certainly is a big improvement to the OS. However what I fail to realize is why you thought various drivers were improvements to the OS.
Now where are the real innovations: get rid of text/binary mode in the files!! How about a unified file name space so that open() and readdir() can find everything, like Plan9 had twenty f**king years ago (Linux also sucks in this area, but not as bad as Windows, where you cannot even list the disks without a special api).
/sbin/update would cause power saving to fail on *all* systems, not just some mysterious new systems. This is as though somehow/sbin/update did not use power at all when running (due to some clever hack), and thus there was no reason to be removed from Linux, but when you get this new processor, somehow that new one causes the hack to break and/sbin/update starts using power.
I can't believe the fucking Microsoft apologists here today.
The EU DID NOT ASK FOR SOURCE CODE
Yet here is this idiot saying "So the EU feels the only way for EU companies to compete against Microsoft is to rip off MS source code". Is this part of the your plan on how to spin this? It is MICROSOFT that proposed source code. That is NOT what the EU is asking for.
So, Mr Microsoft apologist, you have just said that Microsoft is incapable of competing by actually making a better product. Very good. You are a genius.
That's insane and a totally bogus arguement. The scientist will observe the faucet and determine that it could be turned up to deliver more water, and thus the water may very well have been flowing faster or slower in the past.
Where did the genes that make bacteria immune to antibiotics come from, then? Oh, they must have all been there, just not expressed, huh? They must be there for every antibiotic in the world, including all the ones that have not been invented, right?
"Intelligent Design" really means *there was no evolution*. It does not mean that God created the universe in a clever way so that evolution happened, it does not mean that god guided evolution. It means God created all the animals and people as they are. Therefore you do *not* believe in ID. It sounds like you believe in God, which is completely different.
Actually the stuff you listed under "redefinition of user" was not put into the GPL3, it was unchagned there.
What most people *are* listing under the disadvantages is the DRM stuff. It is believed by many that it would be impossible to make a DRM decoding device using Linux without giving out enough information to break the DRM, thus preventing a huge potential market for Linux. (I am not convinced, first a non-kernel program could do all the DRM and not be involved in the GPL at all, also the method of reinstalling the software could change the keys so that the machine will no longer decode the "real" data but instead now be able to decode some other data that does not have any entertainment encoded in it).
Your second disadvantage is the big one: I believe it is impossible to change the license on Linux due to Linus removing the "or any later version" text, since this would require agreement from *every* contributor to any parts of Linux to change this. I think because of this it is impossible, and arguing about whether 3 is better or worse than 2 is irrelevant.
I think if the US government said that Google would be put out of business in the US if they didn't comply, Google would have rolled over and complied. Not that I like it, but they would have, just like they did in China.
When asked when the war would started, Gonzales replied "September 11th, 2001"
There were terrorist attacks before that, by the same people. How come those don't count for the start of the war? If there is a bigger attack, does that mean this time now we are not at war, and the war actually starts at the new bigger attack?
I have not seen any documentation that talks about SetAbortProc in any way other than as a function call. I still believe they were totally unaware that it could be parsed from a WMF file. Do you have any pointer to documentation that literally says "the SetAbortProc code is in the file, followed by 4 bytes that are the pointer".
Also I can assure you that I do know C, and on Intel and almost every other modern architecture, a function pointer is a pointer to the first byte of the first instruction in that function.
Slashdot Mirror
Better still, a user could modify uber-secure-linux so that it runs both the original uber-secure-signed items PLUS things signed by their own key. This would still not require knowing the uber-secure signing key, and they could still reuse the protection to make sure only their own stuff or uber-secure stuff is run. Or they could edit out the uber-secure key so only their stuff will run. Or they could edit out all the protection so everything will run.
The difference from DRM is that the user can do this. It does not matter if uber-secure-company says "that machine is not secure". They can do that, certainly, but if you are uninterested in whether uber-secure cares, you can still make the modifications.
That's not DRM. You are quite capable of running your modified version of the software without RedHat's keys. Anybody else would be able to run it as well. All it means is that when they look at your version, they can tell it is not RedHat's approved version, which is the entire purpose of their signing.
Now if RedHat made their own Linux box that refused to run the software without the correct signature, and provided this copy of Linux with it, that would be DRM. Under the GPL3 they would have to provide some way for you to sign your own modified copy.
I'm a little confused about this, though. If RedHat made a new Linux box with the Linux software in ROM that was soldered to the board and imbedded in a block of epoxy, it would be pretty much impossible to replace the software then, as well. And nobody sees anything wrong with this as it may be a very sensible and practical construction method. I think the GPL3 is trying to remove artificial restrictions to replacing the software.
Obviously somebody who never went to college.
I know you think you are making some kind of profound statement, but the terrible state of OSS documentation just reinforces this.
Despite the fact that you can literally cut & paste the code into your application (not allowed with Microsoft's code) it is obvious that availability of source code is almost useless to interoperability. OSS does a great job of implemented documented standard interfaces, such as HTML and network protocols, but it is obvious that the ability of one piece of OSS to talk to another using some new protocol invented by that other one is very limited and slow. Witness the fact that most non-project programs don't install desktop shortcuts or icons on KDE and Gnome, this is due to the inability to figure out how using the code and the lack of documentation (there was documentation for the directory used by Gnome and enough programs used that that Gnome has to be compatable with it, but nothing I can find for how to use their new registry-like stuff).
I would say that the state of OSS proves conclusively that the source cannot be used and documentation is needed. Now I realize that their stuff probably is not documented and they don't want to try to write it (not any more than any OSS project wants to write documentation), so I think the EU and MS are going to have to compromise by them releasing into the public domain sufficient excerpts from their code to make a program that interoperates. That will at least get things to the cut & paste level.
My god, you magically fixed my copy of Firefox, too!
Actually that is a pointer to a reference image showing what it is supposed to look like.
Actually while typing this I realized you probably put that in as a joke. Good one.
The trick with subpixel rendering is that it is *not* just a 3x higher resolution antialiased image. In fact the hinting has gone through some trouble to arrange things to be 3 subpixels wide. This means that the same number of r,g,b subpixels are turned on as they would with normal antialiasing.
If the user does not percieve the colors equal to what is expected, this would also change their perception of pure white (also made by summing r,g,b), thus changing the entire white balance of the monitor. If this annoys them they would twiddle the color settings to fix it.
You may be worried about non-linear perception of the color levels, with different responses for different colors. However ClearType does not take into account the non-linear sRGB display curve, and this error would swamp any problem in your eyes by an order of magnitude or more, and even that does not seem to be too objectionable.
So basically the above poster is full of it. Don't believe him.
I was under the impression that ClearType is not just the LCD sub-pixel rendering, but the entire "new" antialiasing scheme, designed to replace "Font Smoothing" and useful on both CRTs and LCDs. I don't have any new Windows machines here to test it on, but I remember trying this in a store and the control panel certainly had the ability to turn ClearType on/off that was *seperate* from a control that said whether to use the LCD or not. It is possible this second control has been removed because they have a way to determine this directly from the monitor information. In any case I thought ClearType, even in grayscale, was a big improvement over Font Smoothing (which was also available), and looked more like the antialiasing on OS/X and Linux, probably because it is the same algorithim.
I would certainly expect Microsoft to turn ClearType on at all times if it works the way I think. Although people get programmed to like the way things look and thus complain about "blurry" fonts when they see any gray pixels, the average non-full-time computer user greatly prefers the ClearType output, and so will everybody once they get used to it.
OS/X has antialiasing on at all times, how come nobody complains about it?
No, wrong. The antialiasing selection is done by configurtion files, not the fonts themselves. The fonts contain no inforamtion that says "I can be antialiased but should not be". You may be confused by bitmapped fonts, which cannot be antialiased. Or more likely, Microsoft has a list somewhere of what fonts can be antialiased and you have not found it yet or it is inaccessible through the control panel.
Just want to back up the other responders. I also have never heard "reverse engineering" as being decompilation like you claim. "reverse engineering" is testing the original, inventing new tests to determine things the original tests did not, and producing a copy that outputs the same results as the tests.
Decompilation is genearlly called "cloning" if done legally, maybe cracking/hacking if done illegally. I think legally doing it requires walled-off development groups, such as was done for the BIOS. Decompilation is quite impractical for something the size of Windows, and as you mentioned the MS documentation and many programmers familiar with the platform give a huge head start and make reverse-engineering the far-preferred method.
My god did you even read? I fully admit that the power modes and user-mode drivers *are* improvements to the OS. I was just trying to differentiate between hype and the real thing, something you seem incapable of.
I can guarantee you that if Linux added support for this LCD screen or the remote projectors (something it cannot do without the hardware to test it on) then it will NOT be called Linux 3.0, but still be 2.6. It is NOT a new OS!
Hey I dislike Microsoft as much as the next guy, but saying it is in their interest to make it insecure is silly. If they want to, they can force people to upgrade by adding functionality or changing things so older systems can interface with new systems.
I'm sure they would love to make it secure, as the average person does not know anything wrong with Microsoft except "it's got security bugs". A fixed version would in fact remove 100% of the negative things the average person knows about Microsoft.
- A secondary LCD screen that offers limited, PDA-like functionality while the main OS is in Suspend. How would you do that with XP? Or Linux? It's not anything you can write in user mode.
What the hell are you talking about. That LCD would be a *second screen*. Windows as-is already supports multiple screens, so add a driver for the LCD! I don't consider that a change to the OS. A slight change so that the user can make a program go to that screen without dragging it is, I guess, a change to the OS, but hardly a huge addition. Of course I think MS has gone and made yet another api, totally failing to realize that the existing one was just fine. That is why you seem confused into thinking supporting this new screen is some new OS feature. Sigh.
A little circle appears, kind of like a radar screen, and it figures out which icon you meant to press. Current touchscreens can't do that.
WTF? Last I looked current touchscreens can draw anything they want on it. So I guess they improved the driver, but I suspect they could have put this into XP just as easily.
Power modes
This I would say *is* a feature of Vista as it probably touches lots of different parts of it. It is not a driver like the above two.
The LCD projector thing -- that's not OS?
Um, NO! The LCD projector thing is "hardware in the projector to pick up the wireless signal!" Are you claiming that this could not be done with existing Windows if a hardware manufacturer built such a device?
user-mode drivers
This certainly is a big improvement to the OS. However what I fail to realize is why you thought various drivers were improvements to the OS.
Now where are the real innovations: get rid of text/binary mode in the files!! How about a unified file name space so that open() and readdir() can find everything, like Plan9 had twenty f**king years ago (Linux also sucks in this area, but not as bad as Windows, where you cannot even list the disks without a special api).
/sbin/update would cause power saving to fail on *all* systems, not just some mysterious new systems. This is as though somehow /sbin/update did not use power at all when running (due to some clever hack), and thus there was no reason to be removed from Linux, but when you get this new processor, somehow that new one causes the hack to break and /sbin/update starts using power.
I can't believe the fucking Microsoft apologists here today.
The EU DID NOT ASK FOR SOURCE CODE
Yet here is this idiot saying "So the EU feels the only way for EU companies to compete against Microsoft is to rip off MS source code". Is this part of the your plan on how to spin this? It is MICROSOFT that proposed source code. That is NOT what the EU is asking for.
So, Mr Microsoft apologist, you have just said that Microsoft is incapable of competing by actually making a better product. Very good. You are a genius.
No, Office source is not what is wanted, either. What is wanted is documentation, such as how to read the Office file formats.
It's trivial to change an environment variable.
That's insane and a totally bogus arguement. The scientist will observe the faucet and determine that it could be turned up to deliver more water, and thus the water may very well have been flowing faster or slower in the past.
Where did the genes that make bacteria immune to antibiotics come from, then? Oh, they must have all been there, just not expressed, huh? They must be there for every antibiotic in the world, including all the ones that have not been invented, right?
"Intelligent Design" really means *there was no evolution*. It does not mean that God created the universe in a clever way so that evolution happened, it does not mean that god guided evolution. It means God created all the animals and people as they are. Therefore you do *not* believe in ID. It sounds like you believe in God, which is completely different.
Actually the stuff you listed under "redefinition of user" was not put into the GPL3, it was unchagned there.
What most people *are* listing under the disadvantages is the DRM stuff. It is believed by many that it would be impossible to make a DRM decoding device using Linux without giving out enough information to break the DRM, thus preventing a huge potential market for Linux. (I am not convinced, first a non-kernel program could do all the DRM and not be involved in the GPL at all, also the method of reinstalling the software could change the keys so that the machine will no longer decode the "real" data but instead now be able to decode some other data that does not have any entertainment encoded in it).
Your second disadvantage is the big one: I believe it is impossible to change the license on Linux due to Linus removing the "or any later version" text, since this would require agreement from *every* contributor to any parts of Linux to change this. I think because of this it is impossible, and arguing about whether 3 is better or worse than 2 is irrelevant.
Um, Microsoft did do this. I guess you must have missed that news item, huh?
I think if the US government said that Google would be put out of business in the US if they didn't comply, Google would have rolled over and complied. Not that I like it, but they would have, just like they did in China.
When asked when the war would started, Gonzales replied "September 11th, 2001"
There were terrorist attacks before that, by the same people. How come those don't count for the start of the war? If there is a bigger attack, does that mean this time now we are not at war, and the war actually starts at the new bigger attack?
I have not seen any documentation that talks about SetAbortProc in any way other than as a function call. I still believe they were totally unaware that it could be parsed from a WMF file. Do you have any pointer to documentation that literally says "the SetAbortProc code is in the file, followed by 4 bytes that are the pointer".
Also I can assure you that I do know C, and on Intel and almost every other modern architecture, a function pointer is a pointer to the first byte of the first instruction in that function.