Writing the first copy of a piece of software is very difficult. However, writing all copies of a piece of software after the first copy *is* trivial. This is the huge difference between software and making the second transistor.
The main difference is that if somebody has patented the transistor, you can buy small units (transistors) from them and are still free to invent your phonograph with them. These are small physical objects and you must have them to make your device work, and by far the easiest and cheapest way to aquire them is to exchange money or services for them. Therefore there is a natural and simple way for the patent holder to be directly compensated for the use of their invention.
Now imagine if you could trivially create as many transistors as you wanted in your kitchen sink, using nothing but water, you would have the equivalent of software patents. Suddenly you have zero incentive to actually exchange money for use of the patented device, so suddenly there has to be artificial laws and restrictions imposed on your actions, which is socialism/communism and should actually be violently hated by businesses! More importantly, since they cannot sell anything, the owner of the transistor patent will seek a way to make money by selling something that can make money, such as phonographs, and will have to use or create laws that say that only they can make any inventions using transistors.
Cooking is a much closer analogy to software. Famous chefs either keep their recipies trade secrets or they make them public domain. They can still advertise that they make the "original" or "best" Waldrof salad or whatever, but do not seem exceptionally upset that another copies their invention. This all seems to make sense to the general public, because they can understand cooking. The trick is to make them realize software is the same thing.
The Disney movie Flight of the Navigator probably had the earliest CG spaceship. It was a silver blob shape. I suspect it has the earliest attempt to use CGI to represent a real-world object (rather than an item that is supposed to be computer-generated) in a movie, but I am not sure. Anybody know of earlier examples?
I remember pretty clearly all the initial talk about viruses being for the Mac. The main culprit imho was the "resource" setup of executable files that made it trivial to modify an executable to run any code you wanted, thus hiding the virus entirely in files already on the disk. People would take disks with the infected program, run it on another Mac, and the virus would install itself and infect every executable file on any other disks inserted. (these were 3.5 floppys).
DOS (and Unix) required at least the insertion of some jmp instructions to get your own code added. Everybody thought these were much more clever and possibly more dangerous, but they were at the time almost nonexistent, while Mac virii were everywhere, spread at shocking speed for a system that required disks to be manually carried from one host to another! A typical one would spread over the entire country in a month, and could not be eradicated: cleaning it off and it would be reintroduced soon afterwards.
In fact the word virus was more appropriate to this. Instead of modifying the entire machine, it modified a "cell" (a program) to do it's replication work. The organism (the machine) would eventually be killed. Modern viruses instead seem to hide in the enormous complexity of the file system as extra files, usually of the same size or larger than the other files, so maybe these should be called bacteria instead.
Only item I ever purchased from Fry's was a SCSI CD/RW (this was several years ago). I went through 3 before I found one that worked. Two of them had been ruined by people jamming IDE plugs into the SCSI port, pushing the pins out or breaking them off. So basically I lost a lot of time and Fry's lost a good deal of inventory. It's nice they take returns without question, but I wonder just how clueless the customers and salespeople are, the SCSI drives were 4x as expensive as faster IDE drives, you would think they would ask why and learn that the cheaper one is more likely to work in their machine.
No, the same arguments would have applied if SCO was releasing Linux as public domain, or if they had been selling it to IBM for years as closed source under some agreement where IBM can resell the code, or almost any other "license".
Personally I feel SCO can always claim the code was inserted by some "evil employee out to destroy SCO". The GPL argument I believe is a smoke screen to throw poeple off track. The true argument is paragraph 37: SCO is making no attempt to remedy the copyright infringement by revealing how to remove the infringing code, thus indicating that their copyrights are in fact worthless to them.
I think SCO's claims (that their copyright over part of Linux gives them rights to the whole thing) would be equally bogus if Linux was BSD or public domain. They would even be bogus if Linux was some NDA-based cooperation between SCO and a couple other businesses.
Think about it: if they had released their code public-domain for years, and then claimed "oh that's copyrighted, everybody who used that in anything of theirs owes us money" it would be the same. If they had sold code to IBM for years and then suddendly said "oh that's copyrighted and we didn't give you the copyright, you have to pay us money for everything you used it in" it would be the same.
The best they can claim is that some "evil employee set out to destroy SCO inserted the copyrighted code". But in that case they must identify the code, help in getting it removed (otherwise they are admitting that the copied code is not actually harming them), and then restrict their legal attacks to the "evil employee" (if they want big bucks they can try to prove that IBM paid the evil employee).
So although the GPL is useful, it really has little to do with this case IMHO. I feel IBM is using the GPL argument as a screen around the *real* argument, in order to throw them off the track or get them or even Microsoft to say something stupid. The *real* argument I think is paragraph 37 which points out that SCO is not doing anything active to rememdy the copyright problem, such as identifying what code needs to be changed, thus admitting the copyrights have zero value to them. Then again IBM's lawyers are probably much more clever than that, there may be layers below this already set up to trap them.
SCO's claims would be equally bogus if Linux was completely public domain, or if Linux was some top secret block of closed code shared by two companies, or anything in-between. If you claim copyright on part of a collection of works, that in no way gives you any rights over the other parts of that collection. The best you can do is identify what parts are yours and litigate to get those parts seperated so that your copyrights are no longer violated, and then persue litigation against the party who actually added your parts against your wishes and thus violated your copyrights.
The license has NOTHING to do with the real problem with SCO's claims, there is nothing idealogically wrong with both being against SCO and also not agreeing with the FSF.
PS: personally I don't have any real problem with the FSF, except that the GPL is really simple and their rhetoric makes it sound more complex and dangerous than it really is and thus actually scares people away.
Releasing one thing under the GPL in no way forces you to release everything else you have under the GPL. So if the Darwin kernel was released under the GPL does not mean Apple has to release their GUI which runs atop it under the GPL.
It might be clearer to you if you ignore the FUD about the GPL. Think public domain. If Apple released Darwin public domain (ie anybody can do anything with it) then do you really think they are legally required to release Quartz public domain?
Personally I see no problem with the Apple license. You can see how Darwin works, and you can copy parts into other software.
Yes if you make improvements and distribute your improved version, you have to send the changes to Apple. But really even without those rules people would act the same, and they do on lots of OSS projects. If you made some useful change to Darwin and kept it to yourself, you would have to keep reapplying your patch each time Apple relased a new version of Darwin, if you also wanted whatever improvements are in that. Why not let Apple put the changes into their copy, so you don't have to do this, and so everybody else using Darwin also gets the advantages of your code?
I wasn't saying "tar" should be different (though right now it does nothing if you don't give any switches, so I thought it is ok if that could be interpreted as reading/writing the named tar files).
What I want is a command-line program that does the same thing as double click in the GUI. Why? Mostly so that I can make a program that also presents a representation of files and let the user click them and they do what is expected, without having to link a huge library in. Also it provides a portable way of displaying a url in a browser.
In any case, I should not have to write anything. Gnome and KDE have already written *ALL* the code. What they need to do is make it somehow accessible. I recommend a program called "start" just because that is what cygwin on Windows called it, also because the "open" used by OS/X seems to run some old Unix program.
Most/all modern apps support both the ctrl+* and middle mouse cut & paste.
Middle-mouse is clearer if you think of it as drag and drop with the benifit that you don't have to do the drag, and thus can move and open and close windows before dropping.
XTerm is never going to do ctrl+* because those keystrokes are expected to be sent to the program inside xterm. It would be nice to make a command-line-only "xterm" that does all editing with a gui and does not try to run text editors or anything, though. That would do the job.
One idea I had is that if you select with the middle-mouse button it would do a replace of the text. Still not great because you don't get any chance to fix your selection.
However it now looks like this is being solved in a better way, by realizing that "middle mouse paste" is really drag and drop. For instance Mozilla middle-mouse goes to the url, which is what happens in DND systems if you drag some representation of a url and drop it on the window.
Everybody should stop thinking about this as cut/paste but instead as a better method of drag & drop. Why better? Because you can rearrange the windows before you drop, or open new ones, or do lots of other GUI things that are impossible while dragging something in Windows.
All modern programs support clipboard as well as selection, this uses the ctrl+xcv keys. This also makes more sense if you think of the middle mouse as drag&drop.
I am pretty certain you can make a non-GPL installer for GPL software. You probably should make the original software available, but since it is hard to install you have not reduced any of the demand for your installer.
The real problem about bad installations is not just the initial developer. Anybody else who wants to make an installer will first have to get the software installed so they can figure it out. Once they do that, they don't have any incentive to make an installer.
There absolutely should be a "start" program. KDE and Gnome could provide their own competing ones if wanted, but it is insane that a program cannot perform the "doubleclick a file" action without linking in a huge library.
I used to think that Windows had this, but I was wrong. Still it is possible, you run the program "rundll32.exe url.dll,FileProtocolHandler " and it will open the file (or url) . Despite the massive obscurity of this command (worse than anything I ever saw for Linux), huge numbers of people know it, judging from how easily I located it doing a Web search.
The fact that Linux does not have some kind of "start" command is inexcusable for something that claims to be supporting the command line.
I think your comments about tar are wrong. If I give the name of a tar file, I want to untar it! If I give several, I want to untar them! If I really wanted to tar files up, I would name a directory or a whole lot of files, and put the tar file name at the end! (the -f switch would be supported for back compatability). It really isn't that hard.
Yes, dammit! Linux is full of this sort of stuff. There are obvious solutions to making things more consistent, but some people seem to be unable to see them.
It won't even be incompatable. Make ssh take *both* -p and -P. And in the documentation talk about -P ONLY! (well put -p down on the bottom as "obsolete, means the same as -P").
While you are at it, fix the damn tty driver to take *both* ^H *and* ^? as the erase character. I mean really, this stupid bug has been there in Unix for TWENTY FIVE YEARS and nobody has thought to make two keys do the same thing.
Writing "open source" for some government Department of Software would be a nightmare. I think most people agree here that this would be some huge buearacracy that would stop all innovation, bury everything under techinical requirements, and development and would probably attact real free software much more ferociously than any commercial company can.
Conversely though, if a police department does write some useful software, they should release it, so other police departments (or anybody) can use it. I think that should be argued for.
Several others have said here that such stuff must be BSD, I think I agree. The government is not allowed to release GPL code because they cannot copyright it. Apparently though they can modify GPL code and release that, however, as they have done so several times already.
There is no need to argue licenses. Apparently the government *CANT* use the GPL on anything it writes, because it is not allowed to copyright anything. The GPL is meaningless then, since it is just an exception to the limitations of copyright, and the code would have no limitations on use. Anything the government writes must be public domain (though they can keep it secret, I guess).
Although there are several good reasons why the government should not purposely fund open-source, you have come up with the dumbest argument I have ever heard.
Last time I checked, the government was funding lots of stuff that does not benifit "everyone". That is why there are political parties and disagreement in government.
It is possible that MS is not paying SCO, but that SCO wants people to believe they are.
I think their press statements are carefully designed to make people believe Microsoft is paying them, for instance they say repeatedly that Linux is really dangerous, which is really stupid if you are trying to make money by selling some kind of Linux license. Their statements in no way would make some PHB feel safe even if they did pay for SCO's license, since they imply that there are deep problems with Linux and you are liable from all kinds of sources, so why bother paying SCO when there are all these other problems?
I think 100% of the speculators believe that Microsoft is paying SCO. There is just no chance they are betting on SCO winning any case or collecting license fees. So it is in SCO's interest to have people believe this.
But it is possible that Microsoft is not paying them directly. The main explanation I have is that if Microsoft is caught they are in for a lot of negative publicity, for instance currently they claim that Windows is technically superior to Linux and lower TCO, but that can be easily refuted by "if that was true, why did you have to pay somebody to attack Linux. Obviously that proves Linux is better." It is plausable that SCO set this up to coincide with an already-arranged purchase of that license by Microsoft so it would look like they are doing Microsoft's bidding.
But why doesn't Microsoft refute it? Because that would be admitting that they think the idea that they are funding SCO is serious enough that they have to refute it.
There are also possibilities that this is being funded from some people at Microsoft without official support. Some sort of half-way thing. A $15-million line item for the SCO licensees may have been checked off without question.
No, TrollTech dual-licenses Qt, so you can use it, even though it is not free. Readline is the only useful library I am aware of that actually does not allow closed source code to use it.
Re:i saw it at siggraph last week
on
Walk-thru Fog Screen
·
· Score: 4, Interesting
Saw this as Siggraph as well. It did attract a lot of interest. I walked through it, and most impressive is that you do not feel anything, the airflow is suprisingly low. The biggest problem I had is that the very bright light from the actual projector is visible, I see no way to avoid this and the image certainly looked better the more you are in line with the projector, but that light was blinding.
Also at the same show and more interesting imho:
The most important thing there was a High Dynamic Range display. They placed an LCD in front of a rear-projection display and the combined modulation results in a contrast range of 70000:1. This allows much more realistic images. The images I saw looked like a good slide projector, but could be better in a darker room. There was some registration problems, but they say they are working on using bright white LED's behind the LCD, resulting in a flat screen that is as sharp as an LCD. PS: they patented the idea, which for this I think is ok as long as they actually manufacture an open device, they were a little hesitant to say this, though the current driver is just a dual-headed graphics card and it seems hard to believe you could do much better than that.
Also interesting was a rear-projection globe. It was maybe 6' in diameter and translucent white. This used a single rear projector in the base, reflected off a cone-shaped mirror inside at the top, to project on all sides. They had software showing images of the earth, other planets, continental plate drift. The brightness was suprisingly uniform and the fact that there was a black hole at the north pole was not a problem. A trackball let you spin the globe and the image moves very solidly, indicating the geometry is pretty accurate and they matched it with their image warping software (probably a hardware renderer using texture maps to distort the image correctly). Biggest problem is the room is going to have to be really dark for it to look good.
There was also a demo of those "project on a flat surface" keyboards, and it really works. You can learn to type on it correctly in only a few minutes of practice. Biggest problem I see is that the alignment with the flat surface is critical, the phone manufacturers are going to have to come up with clever folding stands to stand the phone/pda at exactly the right angle. Also it seems obvious to me that the projected image could change, not just to different keyboards, but be used as a display. It requires distortion of the display much like that globe, but even a PDA could do that now.
Perhaps to anybody but only for use in Free software.
That would actually be very clever. IBM could still use the patents for their own non-Free stuff, and could still sell the rights to them to other companies or use them for legal threats. And RedHat could not use the patents for something other than Linux itself or other GPL programs. And if the patent is for some useful interface that is useful for applications or other programs running atop Linux but only if they use stuff in the patent as well, they may be the only ones who can make closed-source applications using that interface, while the existence of that interface in Linux encourages everybody to learn it and use it.
No you are thinking the LGPL. readline is the canonical (and perhaps only) example of the GPL actually harming development by blocking off something that is actually useful for development. Even if you intend to make your new program GPL there is incentive to not use it, as you might change your mind, also the inability to use it also makes you not learn it.
In most cases the GPL is applied to end-user programs where the only reason to link to them them would be to make a closed-source ripoff. Actually useful libraries that make it easier to write programs are almost always LGPL or similar.
Slashdot Mirror
No, Digital Domain keeps Nuke's code a trade secret. Digital Domain has NOT filed any patents.
Writing the first copy of a piece of software is very difficult. However, writing all copies of a piece of software after the first copy *is* trivial. This is the huge difference between software and making the second transistor.
The main difference is that if somebody has patented the transistor, you can buy small units (transistors) from them and are still free to invent your phonograph with them. These are small physical objects and you must have them to make your device work, and by far the easiest and cheapest way to aquire them is to exchange money or services for them. Therefore there is a natural and simple way for the patent holder to be directly compensated for the use of their invention.
Now imagine if you could trivially create as many transistors as you wanted in your kitchen sink, using nothing but water, you would have the equivalent of software patents. Suddenly you have zero incentive to actually exchange money for use of the patented device, so suddenly there has to be artificial laws and restrictions imposed on your actions, which is socialism/communism and should actually be violently hated by businesses! More importantly, since they cannot sell anything, the owner of the transistor patent will seek a way to make money by selling something that can make money, such as phonographs, and will have to use or create laws that say that only they can make any inventions using transistors.
Cooking is a much closer analogy to software. Famous chefs either keep their recipies trade secrets or they make them public domain. They can still advertise that they make the "original" or "best" Waldrof salad or whatever, but do not seem exceptionally upset that another copies their invention. This all seems to make sense to the general public, because they can understand cooking. The trick is to make them realize software is the same thing.
The Disney movie Flight of the Navigator probably had the earliest CG spaceship. It was a silver blob shape. I suspect it has the earliest attempt to use CGI to represent a real-world object (rather than an item that is supposed to be computer-generated) in a movie, but I am not sure. Anybody know of earlier examples?
I remember pretty clearly all the initial talk about viruses being for the Mac. The main culprit imho was the "resource" setup of executable files that made it trivial to modify an executable to run any code you wanted, thus hiding the virus entirely in files already on the disk. People would take disks with the infected program, run it on another Mac, and the virus would install itself and infect every executable file on any other disks inserted. (these were 3.5 floppys).
DOS (and Unix) required at least the insertion of some jmp instructions to get your own code added. Everybody thought these were much more clever and possibly more dangerous, but they were at the time almost nonexistent, while Mac virii were everywhere, spread at shocking speed for a system that required disks to be manually carried from one host to another! A typical one would spread over the entire country in a month, and could not be eradicated: cleaning it off and it would be reintroduced soon afterwards.
In fact the word virus was more appropriate to this. Instead of modifying the entire machine, it modified a "cell" (a program) to do it's replication work. The organism (the machine) would eventually be killed. Modern viruses instead seem to hide in the enormous complexity of the file system as extra files, usually of the same size or larger than the other files, so maybe these should be called bacteria instead.
Only item I ever purchased from Fry's was a SCSI CD/RW (this was several years ago). I went through 3 before I found one that worked. Two of them had been ruined by people jamming IDE plugs into the SCSI port, pushing the pins out or breaking them off. So basically I lost a lot of time and Fry's lost a good deal of inventory. It's nice they take returns without question, but I wonder just how clueless the customers and salespeople are, the SCSI drives were 4x as expensive as faster IDE drives, you would think they would ask why and learn that the cheaper one is more likely to work in their machine.
Personally I feel SCO can always claim the code was inserted by some "evil employee out to destroy SCO". The GPL argument I believe is a smoke screen to throw poeple off track. The true argument is paragraph 37: SCO is making no attempt to remedy the copyright infringement by revealing how to remove the infringing code, thus indicating that their copyrights are in fact worthless to them.
I think SCO's claims (that their copyright over part of Linux gives them rights to the whole thing) would be equally bogus if Linux was BSD or public domain. They would even be bogus if Linux was some NDA-based cooperation between SCO and a couple other businesses.
Think about it: if they had released their code public-domain for years, and then claimed "oh that's copyrighted, everybody who used that in anything of theirs owes us money" it would be the same. If they had sold code to IBM for years and then suddendly said "oh that's copyrighted and we didn't give you the copyright, you have to pay us money for everything you used it in" it would be the same.
The best they can claim is that some "evil employee set out to destroy SCO inserted the copyrighted code". But in that case they must identify the code, help in getting it removed (otherwise they are admitting that the copied code is not actually harming them), and then restrict their legal attacks to the "evil employee" (if they want big bucks they can try to prove that IBM paid the evil employee).
So although the GPL is useful, it really has little to do with this case IMHO. I feel IBM is using the GPL argument as a screen around the *real* argument, in order to throw them off the track or get them or even Microsoft to say something stupid. The *real* argument I think is paragraph 37 which points out that SCO is not doing anything active to rememdy the copyright problem, such as identifying what code needs to be changed, thus admitting the copyrights have zero value to them. Then again IBM's lawyers are probably much more clever than that, there may be layers below this already set up to trap them.
The license has NOTHING to do with the real problem with SCO's claims, there is nothing idealogically wrong with both being against SCO and also not agreeing with the FSF.
PS: personally I don't have any real problem with the FSF, except that the GPL is really simple and their rhetoric makes it sound more complex and dangerous than it really is and thus actually scares people away.
It might be clearer to you if you ignore the FUD about the GPL. Think public domain. If Apple released Darwin public domain (ie anybody can do anything with it) then do you really think they are legally required to release Quartz public domain?
Personally I see no problem with the Apple license. You can see how Darwin works, and you can copy parts into other software.
Yes if you make improvements and distribute your improved version, you have to send the changes to Apple. But really even without those rules people would act the same, and they do on lots of OSS projects. If you made some useful change to Darwin and kept it to yourself, you would have to keep reapplying your patch each time Apple relased a new version of Darwin, if you also wanted whatever improvements are in that. Why not let Apple put the changes into their copy, so you don't have to do this, and so everybody else using Darwin also gets the advantages of your code?
I wasn't saying "tar" should be different (though right now it does nothing if you don't give any switches, so I thought it is ok if that could be interpreted as reading/writing the named tar files).
What I want is a command-line program that does the same thing as double click in the GUI. Why? Mostly so that I can make a program that also presents a representation of files and let the user click them and they do what is expected, without having to link a huge library in. Also it provides a portable way of displaying a url in a browser.
In any case, I should not have to write anything. Gnome and KDE have already written *ALL* the code. What they need to do is make it somehow accessible. I recommend a program called "start" just because that is what cygwin on Windows called it, also because the "open" used by OS/X seems to run some old Unix program.
Most/all modern apps support both the ctrl+* and middle mouse cut & paste.
Middle-mouse is clearer if you think of it as drag and drop with the benifit that you don't have to do the drag, and thus can move and open and close windows before dropping.
XTerm is never going to do ctrl+* because those keystrokes are expected to be sent to the program inside xterm. It would be nice to make a command-line-only "xterm" that does all editing with a gui and does not try to run text editors or anything, though. That would do the job.
However it now looks like this is being solved in a better way, by realizing that "middle mouse paste" is really drag and drop. For instance Mozilla middle-mouse goes to the url, which is what happens in DND systems if you drag some representation of a url and drop it on the window.
Everybody should stop thinking about this as cut/paste but instead as a better method of drag & drop. Why better? Because you can rearrange the windows before you drop, or open new ones, or do lots of other GUI things that are impossible while dragging something in Windows.
All modern programs support clipboard as well as selection, this uses the ctrl+xcv keys. This also makes more sense if you think of the middle mouse as drag&drop.
The real problem about bad installations is not just the initial developer. Anybody else who wants to make an installer will first have to get the software installed so they can figure it out. Once they do that, they don't have any incentive to make an installer.
I used to think that Windows had this, but I was wrong. Still it is possible, you run the program "rundll32.exe url.dll,FileProtocolHandler " and it will open the file (or url) . Despite the massive obscurity of this command (worse than anything I ever saw for Linux), huge numbers of people know it, judging from how easily I located it doing a Web search.
The fact that Linux does not have some kind of "start" command is inexcusable for something that claims to be supporting the command line.
I think your comments about tar are wrong. If I give the name of a tar file, I want to untar it! If I give several, I want to untar them! If I really wanted to tar files up, I would name a directory or a whole lot of files, and put the tar file name at the end! (the -f switch would be supported for back compatability). It really isn't that hard.
It won't even be incompatable. Make ssh take *both* -p and -P. And in the documentation talk about -P ONLY! (well put -p down on the bottom as "obsolete, means the same as -P").
While you are at it, fix the damn tty driver to take *both* ^H *and* ^? as the erase character. I mean really, this stupid bug has been there in Unix for TWENTY FIVE YEARS and nobody has thought to make two keys do the same thing.
Conversely though, if a police department does write some useful software, they should release it, so other police departments (or anybody) can use it. I think that should be argued for.
Several others have said here that such stuff must be BSD, I think I agree. The government is not allowed to release GPL code because they cannot copyright it. Apparently though they can modify GPL code and release that, however, as they have done so several times already.
There is no need to argue licenses. Apparently the government *CANT* use the GPL on anything it writes, because it is not allowed to copyright anything. The GPL is meaningless then, since it is just an exception to the limitations of copyright, and the code would have no limitations on use. Anything the government writes must be public domain (though they can keep it secret, I guess).
Last time I checked, the government was funding lots of stuff that does not benifit "everyone". That is why there are political parties and disagreement in government.
I think their press statements are carefully designed to make people believe Microsoft is paying them, for instance they say repeatedly that Linux is really dangerous, which is really stupid if you are trying to make money by selling some kind of Linux license. Their statements in no way would make some PHB feel safe even if they did pay for SCO's license, since they imply that there are deep problems with Linux and you are liable from all kinds of sources, so why bother paying SCO when there are all these other problems?
I think 100% of the speculators believe that Microsoft is paying SCO. There is just no chance they are betting on SCO winning any case or collecting license fees. So it is in SCO's interest to have people believe this.
But it is possible that Microsoft is not paying them directly. The main explanation I have is that if Microsoft is caught they are in for a lot of negative publicity, for instance currently they claim that Windows is technically superior to Linux and lower TCO, but that can be easily refuted by "if that was true, why did you have to pay somebody to attack Linux. Obviously that proves Linux is better." It is plausable that SCO set this up to coincide with an already-arranged purchase of that license by Microsoft so it would look like they are doing Microsoft's bidding.
But why doesn't Microsoft refute it? Because that would be admitting that they think the idea that they are funding SCO is serious enough that they have to refute it.
There are also possibilities that this is being funded from some people at Microsoft without official support. Some sort of half-way thing. A $15-million line item for the SCO licensees may have been checked off without question.
No, TrollTech dual-licenses Qt, so you can use it, even though it is not free. Readline is the only useful library I am aware of that actually does not allow closed source code to use it.
Saw this as Siggraph as well. It did attract a lot of interest. I walked through it, and most impressive is that you do not feel anything, the airflow is suprisingly low. The biggest problem I had is that the very bright light from the actual projector is visible, I see no way to avoid this and the image certainly looked better the more you are in line with the projector, but that light was blinding.
Also at the same show and more interesting imho:
The most important thing there was a High Dynamic Range display. They placed an LCD in front of a rear-projection display and the combined modulation results in a contrast range of 70000:1. This allows much more realistic images. The images I saw looked like a good slide projector, but could be better in a darker room. There was some registration problems, but they say they are working on using bright white LED's behind the LCD, resulting in a flat screen that is as sharp as an LCD. PS: they patented the idea, which for this I think is ok as long as they actually manufacture an open device, they were a little hesitant to say this, though the current driver is just a dual-headed graphics card and it seems hard to believe you could do much better than that.
Also interesting was a rear-projection globe. It was maybe 6' in diameter and translucent white. This used a single rear projector in the base, reflected off a cone-shaped mirror inside at the top, to project on all sides. They had software showing images of the earth, other planets, continental plate drift. The brightness was suprisingly uniform and the fact that there was a black hole at the north pole was not a problem. A trackball let you spin the globe and the image moves very solidly, indicating the geometry is pretty accurate and they matched it with their image warping software (probably a hardware renderer using texture maps to distort the image correctly). Biggest problem is the room is going to have to be really dark for it to look good.
There was also a demo of those "project on a flat surface" keyboards, and it really works. You can learn to type on it correctly in only a few minutes of practice. Biggest problem I see is that the alignment with the flat surface is critical, the phone manufacturers are going to have to come up with clever folding stands to stand the phone/pda at exactly the right angle. Also it seems obvious to me that the projected image could change, not just to different keyboards, but be used as a display. It requires distortion of the display much like that globe, but even a PDA could do that now.
That would actually be very clever. IBM could still use the patents for their own non-Free stuff, and could still sell the rights to them to other companies or use them for legal threats. And RedHat could not use the patents for something other than Linux itself or other GPL programs. And if the patent is for some useful interface that is useful for applications or other programs running atop Linux but only if they use stuff in the patent as well, they may be the only ones who can make closed-source applications using that interface, while the existence of that interface in Linux encourages everybody to learn it and use it.
In most cases the GPL is applied to end-user programs where the only reason to link to them them would be to make a closed-source ripoff. Actually useful libraries that make it easier to write programs are almost always LGPL or similar.
I agree with you completely.
Best tool for the job, and the GPL is neither always the best tool, nor is it the antichrist.