There is nothing special about patents. So you cannot write GPL software using a patent you don't own. Big deal. If you don't have a license to the patent, you can't write BSD, public domain, or closed-source software either! There is no difference, but you are trying to make some hand-waving argument to support your fantasy that there is something evil about the GPL.
Sorry. The GPL is somewhere BETWEEN closed source and BSD. It is impossible to find something "wrong" with it that is not also something "wrong" with one of those alternatives. So give it up.
I doubt this would work, at least with the technology available in 1999, and probably not today.
Analog or losing track information means that ripping the file to a usable form is harder, but *not* impossible. This has the unfortunate effect that files available on P2P are *more* valuable than the disk, as this hard work has already been done by somebody. This will backfire extensively.
The only possible technological solutions that would work would be to put decryption right into the speakers and encrypt the music seperatly for each playback device. This would actually make ripping for distribution impossible. It would also allow storage and playback from a Linux machine with open source, which would eliminate the real incentive for cracking the encryption (it is obvious that DVD Jon was motivated by the desire to playback on Linux, as Windows-only pirating only needs to duplicate the data, not decrypt it).
There is a simple fact that is completely ignored here: Microsoft can make their program read/write Open Document.. This means it is impossible for Open Document to be less accessable than Microsoft Office, since you can use Microsoft Office! In fact if Microsoft just stopped acting like babies and added the ability to input/output Open Document, I expect the MAJORITY of such documents will be created with Microsoft Office!
Microsoft will lie and kick and scream and moan, and will spew FUD and do everything they can, to stop Open Document, because it means there can be competitors. The fact that one of the possible competitors is an open-source program with poor support for handicapped users is irrelevant, why don't they prove that they can do better.
Also SHAME on Microsoft for taking advantage of the handicapped to spew their lies. You guys are REALLY low and it is sickening.
I would think it would be easier to put out pre-broken software, perhaps it breaks after a certain date or when something happens on the internet. No need to find a hash collision, and this would work no matter what hashing is used.
The open-source GUIs work far better on Linux than Darwin. This has nothing to do with the quality of the OS's, but due to the fact that virtually all development in the last 10 years has been Linux-specific.
As another poster pointed out you lose all targeting speed because you must then return the mouse to where you were working, which on big screens is just as hard to hit as a menu bar, the longer distance makes up for the larger window size. Also the menu bar requires horizontal positioning. And menubars make point-to-type impossible (menu bars in windows as well if you want point-to-type to work for the widgets inside them). Also the horizontal layout requires limiting the size and number of menu items and makes it difficult to pick one quickly (though that could be solved by putting it vertically on the side of the screen).
The real solution is to have pop-up menus under the mouse, with them popping up with the mouse already pointing at the previously-selected item (or perhaps a "most likely item" chosen intelligently by the application). Perhaps they should be pie shaped. Perhaps they should pop up in response to some keyboard key rather than the right mouse button, as Maya does (since navigating is somewhat difficult while holding the mouse button down). But the big detail is that you should not have to move the mouse at all to pick the most-likely item.
The problem is that users are not used to it. The Mac could have been made in 1984 with a key on the keyboard marked "MENU" and now adays NOBODY would be claiming it was user-unfriendly or confusing, everybody would be very used to pointing at anything on the screen, pushing MENU, and seeing what comes up. There would probably also be very clever mouse designs that made picking from the menu very easy.
We are probably stuck with screen areas being used for menus though, rather than popup, as much as we are stuck with QWERTY or the controls on cars. It will be impossible to get people to try anything else and they will cling to any excuse to say that their accustomed way is "easy". It's only easy because they learned it. Instead people are going to fight over where to waste the space, and come up with more and more elaborate schemes for replacing them (seen how complex toolbar management is in some apps?), probably use mouse hysterisis to try to make them easier to move to, make them auto-hide, and do all kinds of elaborate crap because the real solution is unfamiliar and thus "hard".
That is total bullshit. You can go to your local electronics store and buy $10 worth of parts and solder them together and violate FCC regulations. Yet the electronics store is not illegal, and there is documentation about how all those parts work.
Violating copyright is what is illegal. The GPL is simply a license from the original copyright holder saying "it is ok to violate my copyright if you do it this certain way...".
(none of this has anything to do with whether kernel drivers are illegal. They are only illegal if putting them in actually violates copyright and the license for Linux does not excuse it. Not sure about any of that).
If they release an open source driver for Linux at the moment, then a week later someone could come along and break the API they use. They know they don't have enough man-power to fix this, and they don't gain anything by stopping supporting their driver after a few minor kernel revisions...
You seem to have missed an important fact:
If they released an open source driver, they DONT HAVE TO DO ANYTHING!!!! If some Linux developer breaks the API, and somebody cares about their hardware, then that someone will fix their driver. In fact the cheapest way to handle a driver you have developed is to release it open source and then ignore it.
The only thing cheaper is to not write the driver in the first place. I suspect this is the real reason there are no Linux drivers. Saying "we don't want to release secret information" is just an excuse, when the real reason is that they see not enough benefit in writing any kind of driver or even typing up a document saying what the API does.
It also appears that any driver written for Windows requires an NDA from Microsoft. Otherwise a nearly-free way of supporting Linux would be to open-source the Windows driver and say "you guys figure out how to translate this for Linux".
Linux should support user-space drivers. Probably through FUSE and some other apis. These can then be binary, just like any other appliation. If they crash they will not take the system down. The API is limited but you will be able to open/read/write them. ioctl can be done with Plan9 style names, ie open "/dev/neato_device/volume" and write the desired volume there, etc.
Drivers that need more elaborate API's or need more speed will be stuck with the mutable binary interface and occasional GPL restrictions. Too bad. A lot of interesting drivers do not need this speed. And those that do may force the interface to user-level drivers to be improved until it is usable, which is a very desirable result.
Although it could not really be fixed, couldn't it be changed so that it acts like fgets with a fixed buffer length? This could be very small (256? 1024? 80?) so it is smaller than the buffers most programs use, and will not prevent reading of the typical small text files the function is begin called for. It could also be controlled with a static variable so a programmer could set it larger.
There is a difference. With Linux you *can* fuck up your system by changing the source code. Maybe it will work better, and if so then you might be willing to ignore the fact that Oracle does not officially support it. With Windows you can't try to change it at all, whether or not this will make Oracle dislike you.
However I have a lot more trouble with a certain os that requires "__declspec(dllimport)". When they stop requiring this extra syntax that does not compile on other systems, then I will consider it fair to complain about gcc.
Anybody questioning how this is done has not realized the the "program" is interpreted byte code, not any kind of present day machine language. The interpreter (assummed to have no bugs itself) is incapable of writing over memory not owned by the current process.
The problem is that any widget-level interface is unlikely to match what the toolkit uses, and gluing these together is impossible. How many Linux toolkits ported to Windows use the Win32 api? Probably none except WxWindows and that was designed from the start to do this.
Another problem is complexity. X11 does in fact provide what you want for exactly one widget, the "window border". The problem is that a simple check of any toolkit's code will reveal that it takes perhaps 10x as much code to interact with the window manager as it would take to draw the border and handle drag and resize yourself. The interface to this window widget is extremely baroque, complex, and impossible to implement or use correctly or efficiently, and the delay in communicating changes and lack of synchronization is the main reason X11 looks slow no matter how fast the machine is. You might say that if it was well-designed this would not happen, but the window manager interface was considered well-designed by at least some people when it started out, so I don't have a lot of hopes for some "well-designed" new widget interface.
It is obvious that the problem is that the world is divided into people who prefer antialiased and those who prefer non-antialiased (for small letters at least). Each of these say the other system "looks like crap and it took forever to fix it" and then the other camp says "no my system looks perfect, you are full of shit."
One huge problem is that it is almost impossible to figure out what the person complaining about the fonts wants. If they say "it looks blurry" then they are probably alias-lovers, but otherwise it is hard to tell. I think anybody complaining should be required to send two screenshots, one showing what they call "bad" and one showing what they call "good", so we can tell.
Now I prefer antialiased all the time as the text matches all the photographic information on the screen and the weights of the letter strokes look much more even. I feel that the opponents are so brainwashed by using Windows that any fonts that look different, they don't like. This also explains why Microsoft themselves has been unable to turn ClearType on by default in their system, despite the fact that to the average doofus who does not use a computer all day, but does watch a lot of TV, greatly prefers antialiasing on everything. OS/X uses antialiasing everywhere and Mac users don't seem to complain, but Windows users also complain about blurry fonts on Mac.
Also some terms as I understand them, just to make sure everybody is arguing about the same things:
"Font smoothing" is an obsolete method used by Windows to make fake antialiasing by filtering the aliased image. It did not work for small fonts so they turned it off, though it did a good job of recognizing the slope of larger letter edges and smoothing that.
"Antialiasing" is using information about the actual shape to produce gray scales. The most common method now is to just draw it aliased a lot bigger, like 4x4 bigger, then scale down the boxes of pixels, turning the number of "on" ones into a gray shade. This is exactly the technique used by Windows ClearType, Linux FreeType, and Macintosh.
"Hinting" is adjusting the shape to line up better with the pixels so that either the antialiased or the aliased image looks better. If horizontal and vertical lines land on pixel boundaries the letter will have fewer gray pixels and look sharper. This is totally different than "turn off antialiasing" except that hinting is much harder if there is no antialiasing. Originally this was the big defect in Linux in that the hinting algorithim was broken, but this appears to have been fixed for years now. Hinting is NOT perfect because the distortions in the letters can get really annoying, and scaling a document cannot be done smoothly, instead it appears to wiggle.
"ClearType" in Windows means "turn on antialiasing". It is not subpixel rendering, you can see that is a totally different checkmark that you turn on after turning on cleartype.
"SubPixel rendering" is a trick invented by Microsoft to use the LCD colors as smaller pixels. You render antialiased for a 3x wider image, then use the result grayscale image to control the colors. It must also interact with hinting so that colors don't shift (ie lines that are 2 pixels wide will be some color, so make them 3), but both Linux and Windows abandons this for very small sizes, which seems to indicate that hinting is not as important as once believed.
These problems could be addressed in Linux/X11. The way to do it is to redefine middle-mouse-click as "drop the most recent selection". This just happens to match what xterm/etc do and that many people mistakenly call the cut & paste X function. Because of this X programs are much more likely to be able to get this enhancement because they tend not to have use the middle mouse for anything else.
Because you can rearrange, open/close, and otherwise manipulate windows between the selection and the "drop", this addresses all your concerns about DnD.
What I suggest is that every toolkit or system that implements drag & drop be written so that middle-mouse-click acts exactly as though you dragged from the most recently selected thing and dropped it. For text this is not very hard because it would use the "PRIMARY" selection. For other data this may require programs that select items that can be dragged to also update the primary selection.
I cannot accept MSVC project files for the simple reason that I want to be able to add, delete, and rename the source files in my software. If I distributed a MSVC file I have to remember to update that file. Better to not distribute anything and force the MSVC users to copy & paste the correct list from the Makefile each time.
No, they can't really do this. A teacher intent on teaching Creationism and quoting from the government materials or copying them and handing them out in class is protected by fair use rules.
The letter is simply to show that there is a protest against it. It is also possible that mass reproduction (like for every student in Kansas) may be disallowed.
If these work, this is BIG news. IMHO these make Linux into a completely new operating system, not just a Unix clone. Everything a computer process can think about should be named in a single hierarchial namespace. Until now all operating systems except Plan9 have been bogged down by concepts that think that only blocks of bytes on a disk are eligable for this naming scheme.
Expect vast numbers of FUSE programs. In the future I expect *most* non-trivial programs to actually be FUSE programs.
Slashdot Mirror
Don't be an ass.
There is nothing special about patents. So you cannot write GPL software using a patent you don't own. Big deal. If you don't have a license to the patent, you can't write BSD, public domain, or closed-source software either! There is no difference, but you are trying to make some hand-waving argument to support your fantasy that there is something evil about the GPL.
Sorry. The GPL is somewhere BETWEEN closed source and BSD. It is impossible to find something "wrong" with it that is not also something "wrong" with one of those alternatives. So give it up.
I doubt this would work, at least with the technology available in 1999, and probably not today.
Analog or losing track information means that ripping the file to a usable form is harder, but *not* impossible. This has the unfortunate effect that files available on P2P are *more* valuable than the disk, as this hard work has already been done by somebody. This will backfire extensively.
The only possible technological solutions that would work would be to put decryption right into the speakers and encrypt the music seperatly for each playback device. This would actually make ripping for distribution impossible. It would also allow storage and playback from a Linux machine with open source, which would eliminate the real incentive for cracking the encryption (it is obvious that DVD Jon was motivated by the desire to playback on Linux, as Windows-only pirating only needs to duplicate the data, not decrypt it).
There is a simple fact that is completely ignored here: Microsoft can make their program read/write Open Document.. This means it is impossible for Open Document to be less accessable than Microsoft Office, since you can use Microsoft Office! In fact if Microsoft just stopped acting like babies and added the ability to input/output Open Document, I expect the MAJORITY of such documents will be created with Microsoft Office!
Microsoft will lie and kick and scream and moan, and will spew FUD and do everything they can, to stop Open Document, because it means there can be competitors. The fact that one of the possible competitors is an open-source program with poor support for handicapped users is irrelevant, why don't they prove that they can do better.
Also SHAME on Microsoft for taking advantage of the handicapped to spew their lies. You guys are REALLY low and it is sickening.
I would think it would be easier to put out pre-broken software, perhaps it breaks after a certain date or when something happens on the internet. No need to find a hash collision, and this would work no matter what hashing is used.
The open-source GUIs work far better on Linux than Darwin. This has nothing to do with the quality of the OS's, but due to the fact that virtually all development in the last 10 years has been Linux-specific.
Tog was wrong. Sorry.
As another poster pointed out you lose all targeting speed because you must then return the mouse to where you were working, which on big screens is just as hard to hit as a menu bar, the longer distance makes up for the larger window size. Also the menu bar requires horizontal positioning. And menubars make point-to-type impossible (menu bars in windows as well if you want point-to-type to work for the widgets inside them). Also the horizontal layout requires limiting the size and number of menu items and makes it difficult to pick one quickly (though that could be solved by putting it vertically on the side of the screen).
The real solution is to have pop-up menus under the mouse, with them popping up with the mouse already pointing at the previously-selected item (or perhaps a "most likely item" chosen intelligently by the application). Perhaps they should be pie shaped. Perhaps they should pop up in response to some keyboard key rather than the right mouse button, as Maya does (since navigating is somewhat difficult while holding the mouse button down). But the big detail is that you should not have to move the mouse at all to pick the most-likely item.
The problem is that users are not used to it. The Mac could have been made in 1984 with a key on the keyboard marked "MENU" and now adays NOBODY would be claiming it was user-unfriendly or confusing, everybody would be very used to pointing at anything on the screen, pushing MENU, and seeing what comes up. There would probably also be very clever mouse designs that made picking from the menu very easy.
We are probably stuck with screen areas being used for menus though, rather than popup, as much as we are stuck with QWERTY or the controls on cars. It will be impossible to get people to try anything else and they will cling to any excuse to say that their accustomed way is "easy". It's only easy because they learned it. Instead people are going to fight over where to waste the space, and come up with more and more elaborate schemes for replacing them (seen how complex toolbar management is in some apps?), probably use mouse hysterisis to try to make them easier to move to, make them auto-hide, and do all kinds of elaborate crap because the real solution is unfamiliar and thus "hard".
Read again. "The designers want an operating system that can be tinkered with...."
Not "The users want an operating system that can be tinkered with...."
Thats total bullshit. It is not the manufacturer's responsibility if the Linux authors change things.
That is total bullshit. You can go to your local electronics store and buy $10 worth of parts and solder them together and violate FCC regulations. Yet the electronics store is not illegal, and there is documentation about how all those parts work.
Violating copyright is what is illegal. The GPL is simply a license from the original copyright holder saying "it is ok to violate my copyright if you do it this certain way...".
(none of this has anything to do with whether kernel drivers are illegal. They are only illegal if putting them in actually violates copyright and the license for Linux does not excuse it. Not sure about any of that).
If they release an open source driver for Linux at the moment, then a week later someone could come along and break the API they use. They know they don't have enough man-power to fix this, and they don't gain anything by stopping supporting their driver after a few minor kernel revisions...
You seem to have missed an important fact:
If they released an open source driver, they DONT HAVE TO DO ANYTHING!!!! If some Linux developer breaks the API, and somebody cares about their hardware, then that someone will fix their driver. In fact the cheapest way to handle a driver you have developed is to release it open source and then ignore it.
The only thing cheaper is to not write the driver in the first place. I suspect this is the real reason there are no Linux drivers. Saying "we don't want to release secret information" is just an excuse, when the real reason is that they see not enough benefit in writing any kind of driver or even typing up a document saying what the API does.
It also appears that any driver written for Windows requires an NDA from Microsoft. Otherwise a nearly-free way of supporting Linux would be to open-source the Windows driver and say "you guys figure out how to translate this for Linux".
Linux should support user-space drivers. Probably through FUSE and some other apis. These can then be binary, just like any other appliation. If they crash they will not take the system down. The API is limited but you will be able to open/read/write them. ioctl can be done with Plan9 style names, ie open "/dev/neato_device/volume" and write the desired volume there, etc.
Drivers that need more elaborate API's or need more speed will be stuck with the mutable binary interface and occasional GPL restrictions. Too bad. A lot of interesting drivers do not need this speed. And those that do may force the interface to user-level drivers to be improved until it is usable, which is a very desirable result.
Although it could not really be fixed, couldn't it be changed so that it acts like fgets with a fixed buffer length? This could be very small (256? 1024? 80?) so it is smaller than the buffers most programs use, and will not prevent reading of the typical small text files the function is begin called for. It could also be controlled with a static variable so a programmer could set it larger.
There is a difference. With Linux you *can* fuck up your system by changing the source code. Maybe it will work better, and if so then you might be willing to ignore the fact that Oracle does not officially support it. With Windows you can't try to change it at all, whether or not this will make Oracle dislike you.
Yes it *is* "embrace and extend", you are right.
However I have a lot more trouble with a certain os that requires "__declspec(dllimport)". When they stop requiring this extra syntax that does not compile on other systems, then I will consider it fair to complain about gcc.
J Random Citizen will have a pretty hard time doing that if the Internet is swamped with paid-for political messages.
Anybody questioning how this is done has not realized the the "program" is interpreted byte code, not any kind of present day machine language. The interpreter (assummed to have no bugs itself) is incapable of writing over memory not owned by the current process.
The problem is that any widget-level interface is unlikely to match what the toolkit uses, and gluing these together is impossible. How many Linux toolkits ported to Windows use the Win32 api? Probably none except WxWindows and that was designed from the start to do this.
Another problem is complexity. X11 does in fact provide what you want for exactly one widget, the "window border". The problem is that a simple check of any toolkit's code will reveal that it takes perhaps 10x as much code to interact with the window manager as it would take to draw the border and handle drag and resize yourself. The interface to this window widget is extremely baroque, complex, and impossible to implement or use correctly or efficiently, and the delay in communicating changes and lack of synchronization is the main reason X11 looks slow no matter how fast the machine is. You might say that if it was well-designed this would not happen, but the window manager interface was considered well-designed by at least some people when it started out, so I don't have a lot of hopes for some "well-designed" new widget interface.
It is obvious that the problem is that the world is divided into people who prefer antialiased and those who prefer non-antialiased (for small letters at least). Each of these say the other system "looks like crap and it took forever to fix it" and then the other camp says "no my system looks perfect, you are full of shit."
One huge problem is that it is almost impossible to figure out what the person complaining about the fonts wants. If they say "it looks blurry" then they are probably alias-lovers, but otherwise it is hard to tell. I think anybody complaining should be required to send two screenshots, one showing what they call "bad" and one showing what they call "good", so we can tell.
Now I prefer antialiased all the time as the text matches all the photographic information on the screen and the weights of the letter strokes look much more even. I feel that the opponents are so brainwashed by using Windows that any fonts that look different, they don't like. This also explains why Microsoft themselves has been unable to turn ClearType on by default in their system, despite the fact that to the average doofus who does not use a computer all day, but does watch a lot of TV, greatly prefers antialiasing on everything. OS/X uses antialiasing everywhere and Mac users don't seem to complain, but Windows users also complain about blurry fonts on Mac.
Also some terms as I understand them, just to make sure everybody is arguing about the same things:
"Font smoothing" is an obsolete method used by Windows to make fake antialiasing by filtering the aliased image. It did not work for small fonts so they turned it off, though it did a good job of recognizing the slope of larger letter edges and smoothing that.
"Antialiasing" is using information about the actual shape to produce gray scales. The most common method now is to just draw it aliased a lot bigger, like 4x4 bigger, then scale down the boxes of pixels, turning the number of "on" ones into a gray shade. This is exactly the technique used by Windows ClearType, Linux FreeType, and Macintosh.
"Hinting" is adjusting the shape to line up better with the pixels so that either the antialiased or the aliased image looks better. If horizontal and vertical lines land on pixel boundaries the letter will have fewer gray pixels and look sharper. This is totally different than "turn off antialiasing" except that hinting is much harder if there is no antialiasing. Originally this was the big defect in Linux in that the hinting algorithim was broken, but this appears to have been fixed for years now. Hinting is NOT perfect because the distortions in the letters can get really annoying, and scaling a document cannot be done smoothly, instead it appears to wiggle.
"ClearType" in Windows means "turn on antialiasing". It is not subpixel rendering, you can see that is a totally different checkmark that you turn on after turning on cleartype.
"SubPixel rendering" is a trick invented by Microsoft to use the LCD colors as smaller pixels. You render antialiased for a 3x wider image, then use the result grayscale image to control the colors. It must also interact with hinting so that colors don't shift (ie lines that are 2 pixels wide will be some color, so make them 3), but both Linux and Windows abandons this for very small sizes, which seems to indicate that hinting is not as important as once believed.
These problems could be addressed in Linux/X11. The way to do it is to redefine middle-mouse-click as "drop the most recent selection". This just happens to match what xterm/etc do and that many people mistakenly call the cut & paste X function. Because of this X programs are much more likely to be able to get this enhancement because they tend not to have use the middle mouse for anything else.
Because you can rearrange, open/close, and otherwise manipulate windows between the selection and the "drop", this addresses all your concerns about DnD.
What I suggest is that every toolkit or system that implements drag & drop be written so that middle-mouse-click acts exactly as though you dragged from the most recently selected thing and dropped it. For text this is not very hard because it would use the "PRIMARY" selection. For other data this may require programs that select items that can be dragged to also update the primary selection.
In reality that is what I am doing. We do have MSVC files but I don't maintain them, and usually somebody does fix them after a short while.
I cannot accept MSVC project files for the simple reason that I want to be able to add, delete, and rename the source files in my software. If I distributed a MSVC file I have to remember to update that file. Better to not distribute anything and force the MSVC users to copy & paste the correct list from the Makefile each time.
No, they can't really do this. A teacher intent on teaching Creationism and quoting from the government materials or copying them and handing them out in class is protected by fair use rules.
The letter is simply to show that there is a protest against it. It is also possible that mass reproduction (like for every student in Kansas) may be disallowed.
If these work, this is BIG news. IMHO these make Linux into a completely new operating system, not just a Unix clone. Everything a computer process can think about should be named in a single hierarchial namespace. Until now all operating systems except Plan9 have been bogged down by concepts that think that only blocks of bytes on a disk are eligable for this naming scheme.
Expect vast numbers of FUSE programs. In the future I expect *most* non-trivial programs to actually be FUSE programs.
Did you notice the tooltips on the links were ads as well?