I'm not sure what you are getting at, as focus-follows-mouse helps considerably in your scenario of multiple documents.
However Gimp (and anything wanting working floating tool bars) is broken due to slavish copying of Windows and abandonment of stuff that was done intelligently in older versions of X window managers.
First of all, make point-to-type the default. It is better. No argument from anybody who seriously tries it. Anybody who argues against it has not used it.
Second, and more serious, but easy to fix, is STOP RAISING WINDOWS ON CLICK! This makes overlapping windows completely useless except as decoration. I want to look at an item in the top window and manipulate an item in the lower window. You can raise windows when the user clicks the border, or the PROGRAM CAN RAISE IT (this second solution is SO trivial to do that I cannot believe anybody thinks a window manager should raise on click, yet they ALL do, most likely because "Windoze does that so by definition it is user friendly".
Third, and harder, is to change the communication to programs so the program is just told "hey they are trying to raise this window" and the program can do what it wants, such as raise other windows. Right now about all that can be done are complex arrangements of marking windows as belonging to an application or marking "child" or "modal" windows. None of this allows a working floating toolbar that stays above more than one document window, while this, again, TRIVIAL, api would allow any such thing and would make Gimp work.
There are some complaints about voter id, but that is not one of them. The poll workers check your name against the voter registration rolls, so they know who you are (unless you lie, but that will cause trouble if the real voter shows up, and is really what voter id would solve).
The problem with that is that it allows vote buying (ie "prove to me that you voted for so and so, or you are fired!" (or divorced, or dead, or kicked out of the union, or whatever).
Though I doubt vote buying in the modern day will lead to much effect on the election (anything large enough would be detected because somebody would talk), it does mean violence and threats and thus is a bad thing.
There are schemes using cryptography that would allow you to prove that your vote was counted correctly yet be unable to prove how you voted, but they are far too complex for the average voter to understand and believe in.
Not in California, and I don't remember seeing this in Massachusetts either.
It sounds like a bad idea. If there was a single checkbox, for small offices the democrat/republican slate votes would overwhelm all the individual votes by people who actually had an opinion about the candidates. Though I'm sure there are plenty of votors who check every Democrat (or Republican) whether they heard of them or not, but lets not make it too easy for that to happen...
It does seem like image analysis would work. There is no need for real OCR, all that has to be done is recognize that the image *is* text. Since there is no reason to send an image of text other than that it should be a very high indication of spam.
That's just stupid. Here is how it *SHOULD* work, something which everybody (Linux, Windows, AND OS/X) seem to be ignoring:
1. User inserts CD. Window showing contents of CD automatically opens, but nothing automatically runs!
2. User clicks on the word processor in the CD window.
2a. THE PROGRAM IS RUNNING! It's not "installing". IT RUNS! IE the program is *EXECUTING* NOW. Got it. Please look up "executing" or "running" in a computer terminology dictionary if you are still confused by this concept. It does not involve modifying the operating system or any disks, okay?
3. If the user likes the program and does not want to have to stick the disk in, they drag & drop the program onto another location to copy it. They can then double-click it there instead.
3a. If the user dislikes the program they can eject the disk and their system is the same as before (except if they saved a document they now have a file that if they double-click it will not work because the program cannot be found).
I'm sorry, but what part of "this isn't going to work very well if it is some kind of mass market software where you want to get paid individually for thousands of identical copies" did you not understand?
The main complaint I have is that people really are underestimating how much software actually sells only one copy, by design.
But "Rick Santorum" and "Senator Santorum" and "Santorum republican" and "Santorum (r)" work fine. I think this proves my point. I doubt too many people use only the last name when searching, and I think you will be unable to find any reference to him that consists of only his last name. So only unusual things, such as the last name only, can be google bombed.
I put the GPL (actually the LGPL) on my code for entirely greedy, selfish reasons. It is so people cannot take advantage of my code without me being able to see the modifications they do to it. However, me, being a greedy selfish capitalist, *can* do anything I want with my code such as make closed-source forks of it.
Communism is where the government makes you do something for the good of everybody. If the government forced *ALL* code to be GPL, that would be communism.
People give food to their children often without getting paid for it. That could be called "communist", too. However they may very well only be interested in having their children care for them when they are old or some other selfish reason.
Wrong. Your customer does NOT in any way have the option to get the software for free. You can refuse to give them your modification of the GPL software unless they give you money. Because you are not distributing it unless they pay you, you are not violating the copyright, and thus the GPL has no say in what you do.
The GPL does say the customer can then give the software away for free or try to charge for it, or to modify it themselves and sell that. Because of this it is true that this isn't going to work very well if it is some kind of mass market software where you want to get paid individually for thousands of identical copies.
However it works quite well if nobody but your customer is interested in the code you sell them (ie it is custom software like 90% of the software people are employed to write), or if your customer has an interest in keeping the code secret.
There is no problem at all with that. You can run GPL3 programs all you want on any copy of Linux, and distribute them on the same disk as Linux. Not sure how you got the idea there was a problem.
I do have some problems with it. Not with RMS writing a license with those DRM restrictions. But with making that be the GPL3. It perhaps could be called the GPL4.
The GPL3 apparently contains a lot of cleanup of the writing, clarifications, and fixes for international use. In particular the text is clarified so that the LGPL is a small "exception" added to the GPL, rather than an entire seperate document. All of these have results identical to the intentions of the GPL2 or are slight relaxations of the requirements. All of this is good and everybody likes it, and I would like to use it.
However it also has this stuff that most people here are calling the "DRM restriction". I actually have reasons to not want it:
First I feel it is bad as it will reduce usage of GPL software in devices. Knowing how the device works is still extremely useful, including knowing the reason why you can't change the software. The GPL forces the company into allowing people to know how the device works. Stallman originally wanted to fix a printer *driver*, not the code in the printer! His attempt to make sure he can change the code in the printer may result in being unable to write the driver again, which is completely counter-productive. Knowing how the device works means you can probably communicate with it and emulate it and make competing products. (yes I know DRM can keep unauthorized things from communicating with it, but the GPL3 does not prevent that type of DRM anyway, as has been pointed out about six thousand times to anti-GPL trolls here).
Second, my own software already contains an exception (to the LGPL), which is intended to make the LGPL work the way I think makes more sense. Basically you are allowed to link the unmodified software with your code and do anything you want with the result, such as sell it as closed-source. However if you *modify* the software, you must release the modifications (and then you can link with the modified version and release that any way you want). The purpose is so that the algorithims and code cannot be "stolen" but can be used by as many people as possible. You can remove the exception in your own version, so you can merge in GPL/LGPL code, though we can't accept any such changes. As far as I can tell, this exception makes the "DRM restriction" nullified, though I guess you can't build the DRM into the derived version of the library, it must be in your program.
Like many people I would very much like to get the cleaned up and internationalized language of the GPL3. However I don't want the DRM stuff, as I disagree with it somewhat, and my exception probably nullifies it, so I don't want to confuse people. Unfortunatly my code says "GPL2 or any later version" and lots of others have contributed to it so I can't change that. So I am stuck, the only way to get the cleaned up language is for it to be in something the FSF calls a "later version of the GPL". So I would really like them to provide this option. This does not mean they have to back off on their DRM stuff. Just put that in a "GPL4" and let people choose. It would be no worse than the current situation where people who don't want the DRM stuff will stay at GPL2. (future changes would have to be called "GPL3.1" and "GPL4.1", etc, with rules that increasing any number is a "later version", so you can change 3.1 to 3.2 or 4.1, but cannot change 4.1 to 3.2 or 3.2 to 4.1).
It also appears, as others have pointed out, that the DRM stuff (and perhaps the Patent stuff) is an "additional restriction" which means you are not allowed to modify code from saying "GPL2 and later" to saying "GPL3 and later". This kind of means the GPL3 can never be enforced unless the code is written from scratch. This could be another reason to make a GPL3 and a DRM-restriction GPL4.
Unlike linking an uncommon word like "miserable failure" to the main page for a famous person, this is the reverse, an attempt to link the person's name to the uncommon word.
I don't think this is really going to work. There are too many "correct" links for that famous person.
I went and tried "George Bush" and "GW Bush" and lots of other variations. All came up with news articles and the whitehouse web page and right-wing blogs. Only after I started getting very negative and adding words like "George Monkey Bush" did I get anything that would not be Republican material.
I tried "John Kerry". He isn't even in the news anymore, and his site was at the top. Again you had to type in things like "John Swift Boat Kerry" to get stuff not written by Democrats.
I tried a local politician, Bill Rosendahl (city councilman where I am, and he is liberal and gay, which should lead to some attacks) and the top thing was a city council report where he introduced a motion, followed by his web site. In this case the 5th item sounded like an attack ("the miseducation of Bill Rosendahl"), but turned out to be something from the LA Weekly praising him for some affordable housing stuff.
The problem is that any kind of verification like that allows vote buying. I think this is less of a worry for actually throwing an election nowadays, but it can lead to threats and violence and should not be allowed.
There are crypto schemes that would allow you to verify that your vote was counted correctly without being able to show how you voted, but they are far too complex for an average voter to understand.
Well then the Republicans are fucking idiots if they don't question each and every single electronic voting district that went Democratic, especially unexpectedly. Really, are you saying the Republicans are that stupid?
My biggest worry about these machines is that unless the election is a landslide one way or another, the loser is going to cry that the election was stolen, and you literally will not be able to prove them wrong. This is going to suck no matter who wins and if you actually believe it is going to make a difference which side does the complaining you have a lot to learn.
What is stolen is each voters ability to influence the election. Unlike the music company that still has the music and can still try to sell it, the voter really has lost something they had. So yes, this can be considered stealing.
More importantly, "scale *any* image bigger and smaller without any loss in image quality" is pretty much means they are NOT using vector graphics. "Image" generally means bitmapped image. If they said "scale the new icons bigger and smaller without any loss in image quality" I would say that is vector graphics.
It is sounding like the fact that this is done by hardware that is given a rectangle to texturemap and the edges of the rectangle are vectors is making these people call what everybody else would call raster graphics "vector". This is really sad.
Neither this or OS/X are really using vector graphics. Sure the API supports vectors, but it is not actively used. For real vector graphics, take a look at the large scalable shapes in Flash. Also the IRIX desktop in 1990 or so did it, in that the icons were vector images (not that they looked like anything acceptable today).
I suspect you have no idea what you are talking about.
Possibly you are confusing the fact that the icon images can be scaled by the hardware? Ie they are texture-mapped onto a rectangle shape and the GPU does nice filtering of it when scaling? The edges of that shape are vectors, I guess, but that is hardly what anybody calls vector graphics.
Yep, I think the copy of Mandrake was old when I installed it, and that was 2 years ago. I actually bought a new one (since getting the NVidia drivers to work was such a pain, I paid for them) but really have not gotten around to installing it.
In any case, I thought it was perhaps a bug. It did not fail with any missing libraries, and the version 7 plugin produced sound, so perhaps they are calling things wrong in this new version?
On my Mandrake machine. I got no sound from YouTube, and sound works in the FlashPlayer7.
Notes:
Biggest problem is no sound from YouTube (or probably from anywhere). Sound works for me with FlashPlayer7 and switching back to that makes it work without any restarting (so it did not permanently mess up sound, like some programs can). This is a Mandrake machine, 2.4.22-10mdkenterprise, I really have no idea how I have sound set up, but it works for me in most software.
Yes it fixed places that check for the version number of the flash player.
Popping up the menu with the right button (which I did to check that it reported 9 or 7) would cause Firefox to crash somewhat later. Does not seem to happen with 7. May indicate an overflow of some malloc'd data buffer.
To use, put libflashplayer9.so into ~/.mozilla/plugins and don't rename it. Apparently if it exists it will be loaded in preference to libflashplayer.so. (I wasted some time making a flashplayer.so symbolic link that switched between 7 and 9 before I finally figured out that 9 was being used no matter how I set it. Instead, to switch back to 7, rename libflashplayer9.so to libflashplayer9.so.hidden).
Removal instructions in the readme.txt say to remove libflashplayer.so, not the correct file of libflashplayer9.so.
ldd shows it links in far more libraries than 7 did, lots of gtk stuff. I suspect this is due to Pango (which does I18N text layout) using the gobject library, not because any gtk widgets are being used. This has also been complained about on Cairo (which is supposed to be a drawing library *used* by toolkits like gtk, but because good font layout requires Pango, there is a circular dependency back to gtk!)
No, you cannot distribute your zillion-line program without source if it contains 300 lines of GPL code.
However, YOU CAN REMOVE THAT 300 LINES and then distribute the rest just fine. This is NOT what SCO is claiming.
Please look up the what the sentence "REMOVE THE 300 LINES" means. It means ONLY 300 lines. It does not mean all your code. Your code is not "infected" one tiny little bit.
Perhaps I have managed to "get the facts" into your little pea brain. But I doubt it.
The error with this argument is that by *default* you are not allowed to copy anything, as it violates copyright. Any user seeing a "(C) somebody" in the help box should assumme they cannot copy anything, and thus they will not violate the GPL. I think showing the GPL (thus explaining that you can copy it if you follow certain rules) is probably counter-productive because it will confuse most people, such as you, into thinking it is a EULA. If the person wants his friend to have a copy they can send the download URL and not violate anything.
Uh, no, normal users do not use root. Even modern Lindows does not come that way.
The problem is that the user may be fooled into running this cool-new-graphics-program that uses this to get root access. I would not be suprised if the exploit could be run by a Windows program running under Wine. In fact, we use NVidia on both Windows and Linux here and many bugs are cross-platform, has anybody checked if this exploit happens in Windows, too?
Microsoft could finally prove they are not assholes. Release it with complete specs and sample output, don't require Windows libraries, allo anybody to read and write it with any software using any license, and PDF will be dead in a few months.
It does sound better: it is output-only (which is really all we care about in PDF), it uses XML, and it supports alpha compositing like SVG does. Unfortunatly doing anything correctly means Microsoft has to admit that Open Source is not an evil cancer. Don't know if they can bear to do this, or if they are even capable of doing it.
Slashdot Mirror
I'm not sure what you are getting at, as focus-follows-mouse helps considerably in your scenario of multiple documents.
However Gimp (and anything wanting working floating tool bars) is broken due to slavish copying of Windows and abandonment of stuff that was done intelligently in older versions of X window managers.
First of all, make point-to-type the default. It is better. No argument from anybody who seriously tries it. Anybody who argues against it has not used it.
Second, and more serious, but easy to fix, is STOP RAISING WINDOWS ON CLICK! This makes overlapping windows completely useless except as decoration. I want to look at an item in the top window and manipulate an item in the lower window. You can raise windows when the user clicks the border, or the PROGRAM CAN RAISE IT (this second solution is SO trivial to do that I cannot believe anybody thinks a window manager should raise on click, yet they ALL do, most likely because "Windoze does that so by definition it is user friendly".
Third, and harder, is to change the communication to programs so the program is just told "hey they are trying to raise this window" and the program can do what it wants, such as raise other windows. Right now about all that can be done are complex arrangements of marking windows as belonging to an application or marking "child" or "modal" windows. None of this allows a working floating toolbar that stays above more than one document window, while this, again, TRIVIAL, api would allow any such thing and would make Gimp work.
What the hell are you talking about?
There are some complaints about voter id, but that is not one of them. The poll workers check your name against the voter registration rolls, so they know who you are (unless you lie, but that will cause trouble if the real voter shows up, and is really what voter id would solve).
The problem with that is that it allows vote buying (ie "prove to me that you voted for so and so, or you are fired!" (or divorced, or dead, or kicked out of the union, or whatever).
Though I doubt vote buying in the modern day will lead to much effect on the election (anything large enough would be detected because somebody would talk), it does mean violence and threats and thus is a bad thing.
There are schemes using cryptography that would allow you to prove that your vote was counted correctly yet be unable to prove how you voted, but they are far too complex for the average voter to understand and believe in.
Not in California, and I don't remember seeing this in Massachusetts either.
It sounds like a bad idea. If there was a single checkbox, for small offices the democrat/republican slate votes would overwhelm all the individual votes by people who actually had an opinion about the candidates. Though I'm sure there are plenty of votors who check every Democrat (or Republican) whether they heard of them or not, but lets not make it too easy for that to happen...
It does seem like image analysis would work. There is no need for real OCR, all that has to be done is recognize that the image *is* text. Since there is no reason to send an image of text other than that it should be a very high indication of spam.
That's just stupid. Here is how it *SHOULD* work, something which everybody (Linux, Windows, AND OS/X) seem to be ignoring:
1. User inserts CD. Window showing contents of CD automatically opens, but nothing automatically runs!
2. User clicks on the word processor in the CD window.
2a. THE PROGRAM IS RUNNING! It's not "installing". IT RUNS! IE the program is *EXECUTING* NOW. Got it. Please look up "executing" or "running" in a computer terminology dictionary if you are still confused by this concept. It does not involve modifying the operating system or any disks, okay?
3. If the user likes the program and does not want to have to stick the disk in, they drag & drop the program onto another location to copy it. They can then double-click it there instead.
3a. If the user dislikes the program they can eject the disk and their system is the same as before (except if they saved a document they now have a file that if they double-click it will not work because the program cannot be found).
I'm sorry, but what part of "this isn't going to work very well if it is some kind of mass market software where you want to get paid individually for thousands of identical copies" did you not understand?
The main complaint I have is that people really are underestimating how much software actually sells only one copy, by design.
But "Rick Santorum" and "Senator Santorum" and "Santorum republican" and "Santorum (r)" work fine. I think this proves my point. I doubt too many people use only the last name when searching, and I think you will be unable to find any reference to him that consists of only his last name. So only unusual things, such as the last name only, can be google bombed.
I think the Communism label is completely wrong.
I put the GPL (actually the LGPL) on my code for entirely greedy, selfish reasons. It is so people cannot take advantage of my code without me being able to see the modifications they do to it. However, me, being a greedy selfish capitalist, *can* do anything I want with my code such as make closed-source forks of it.
Communism is where the government makes you do something for the good of everybody. If the government forced *ALL* code to be GPL, that would be communism.
People give food to their children often without getting paid for it. That could be called "communist", too. However they may very well only be interested in having their children care for them when they are old or some other selfish reason.
Wrong. Your customer does NOT in any way have the option to get the software for free. You can refuse to give them your modification of the GPL software unless they give you money. Because you are not distributing it unless they pay you, you are not violating the copyright, and thus the GPL has no say in what you do.
The GPL does say the customer can then give the software away for free or try to charge for it, or to modify it themselves and sell that. Because of this it is true that this isn't going to work very well if it is some kind of mass market software where you want to get paid individually for thousands of identical copies.
However it works quite well if nobody but your customer is interested in the code you sell them (ie it is custom software like 90% of the software people are employed to write), or if your customer has an interest in keeping the code secret.
That quote was from Brian Reid.
Here, I'll help you attack RMS if that is all you want:
"Richard Stallman eats small defenseless kittens for breakfast" - Bill Spitzak
There you go, another quote to attack him with, just as good as that first one.
There is no problem at all with that. You can run GPL3 programs all you want on any copy of Linux, and distribute them on the same disk as Linux. Not sure how you got the idea there was a problem.
I do have some problems with it. Not with RMS writing a license with those DRM restrictions. But with making that be the GPL3. It perhaps could be called the GPL4.
The GPL3 apparently contains a lot of cleanup of the writing, clarifications, and fixes for international use. In particular the text is clarified so that the LGPL is a small "exception" added to the GPL, rather than an entire seperate document. All of these have results identical to the intentions of the GPL2 or are slight relaxations of the requirements. All of this is good and everybody likes it, and I would like to use it.
However it also has this stuff that most people here are calling the "DRM restriction". I actually have reasons to not want it:
First I feel it is bad as it will reduce usage of GPL software in devices. Knowing how the device works is still extremely useful, including knowing the reason why you can't change the software. The GPL forces the company into allowing people to know how the device works. Stallman originally wanted to fix a printer *driver*, not the code in the printer! His attempt to make sure he can change the code in the printer may result in being unable to write the driver again, which is completely counter-productive. Knowing how the device works means you can probably communicate with it and emulate it and make competing products. (yes I know DRM can keep unauthorized things from communicating with it, but the GPL3 does not prevent that type of DRM anyway, as has been pointed out about six thousand times to anti-GPL trolls here).
Second, my own software already contains an exception (to the LGPL), which is intended to make the LGPL work the way I think makes more sense. Basically you are allowed to link the unmodified software with your code and do anything you want with the result, such as sell it as closed-source. However if you *modify* the software, you must release the modifications (and then you can link with the modified version and release that any way you want). The purpose is so that the algorithims and code cannot be "stolen" but can be used by as many people as possible. You can remove the exception in your own version, so you can merge in GPL/LGPL code, though we can't accept any such changes. As far as I can tell, this exception makes the "DRM restriction" nullified, though I guess you can't build the DRM into the derived version of the library, it must be in your program.
Like many people I would very much like to get the cleaned up and internationalized language of the GPL3. However I don't want the DRM stuff, as I disagree with it somewhat, and my exception probably nullifies it, so I don't want to confuse people. Unfortunatly my code says "GPL2 or any later version" and lots of others have contributed to it so I can't change that. So I am stuck, the only way to get the cleaned up language is for it to be in something the FSF calls a "later version of the GPL". So I would really like them to provide this option. This does not mean they have to back off on their DRM stuff. Just put that in a "GPL4" and let people choose. It would be no worse than the current situation where people who don't want the DRM stuff will stay at GPL2. (future changes would have to be called "GPL3.1" and "GPL4.1", etc, with rules that increasing any number is a "later version", so you can change 3.1 to 3.2 or 4.1, but cannot change 4.1 to 3.2 or 3.2 to 4.1).
It also appears, as others have pointed out, that the DRM stuff (and perhaps the Patent stuff) is an "additional restriction" which means you are not allowed to modify code from saying "GPL2 and later" to saying "GPL3 and later". This kind of means the GPL3 can never be enforced unless the code is written from scratch. This could be another reason to make a GPL3 and a DRM-restriction GPL4.
Unlike linking an uncommon word like "miserable failure" to the main page for a famous person, this is the reverse, an attempt to link the person's name to the uncommon word.
I don't think this is really going to work. There are too many "correct" links for that famous person.
I went and tried "George Bush" and "GW Bush" and lots of other variations. All came up with news articles and the whitehouse web page and right-wing blogs. Only after I started getting very negative and adding words like "George Monkey Bush" did I get anything that would not be Republican material.
I tried "John Kerry". He isn't even in the news anymore, and his site was at the top. Again you had to type in things like "John Swift Boat Kerry" to get stuff not written by Democrats.
I tried a local politician, Bill Rosendahl (city councilman where I am, and he is liberal and gay, which should lead to some attacks) and the top thing was a city council report where he introduced a motion, followed by his web site. In this case the 5th item sounded like an attack ("the miseducation of Bill Rosendahl"), but turned out to be something from the LA Weekly praising him for some affordable housing stuff.
The problem is that any kind of verification like that allows vote buying. I think this is less of a worry for actually throwing an election nowadays, but it can lead to threats and violence and should not be allowed.
There are crypto schemes that would allow you to verify that your vote was counted correctly without being able to show how you voted, but they are far too complex for an average voter to understand.
Well then the Republicans are fucking idiots if they don't question each and every single electronic voting district that went Democratic, especially unexpectedly. Really, are you saying the Republicans are that stupid?
My biggest worry about these machines is that unless the election is a landslide one way or another, the loser is going to cry that the election was stolen, and you literally will not be able to prove them wrong. This is going to suck no matter who wins and if you actually believe it is going to make a difference which side does the complaining you have a lot to learn.
What is stolen is each voters ability to influence the election. Unlike the music company that still has the music and can still try to sell it, the voter really has lost something they had. So yes, this can be considered stealing.
More importantly, "scale *any* image bigger and smaller without any loss in image quality" is pretty much means they are NOT using vector graphics. "Image" generally means bitmapped image. If they said "scale the new icons bigger and smaller without any loss in image quality" I would say that is vector graphics.
It is sounding like the fact that this is done by hardware that is given a rectangle to texturemap and the edges of the rectangle are vectors is making these people call what everybody else would call raster graphics "vector". This is really sad.
Neither this or OS/X are really using vector graphics. Sure the API supports vectors, but it is not actively used. For real vector graphics, take a look at the large scalable shapes in Flash. Also the IRIX desktop in 1990 or so did it, in that the icons were vector images (not that they looked like anything acceptable today).
I suspect you have no idea what you are talking about.
Possibly you are confusing the fact that the icon images can be scaled by the hardware? Ie they are texture-mapped onto a rectangle shape and the GPU does nice filtering of it when scaling? The edges of that shape are vectors, I guess, but that is hardly what anybody calls vector graphics.
Yep, I think the copy of Mandrake was old when I installed it, and that was 2 years ago. I actually bought a new one (since getting the NVidia drivers to work was such a pain, I paid for them) but really have not gotten around to installing it.
In any case, I thought it was perhaps a bug. It did not fail with any missing libraries, and the version 7 plugin produced sound, so perhaps they are calling things wrong in this new version?
On my Mandrake machine. I got no sound from YouTube, and sound works in the FlashPlayer7.
Notes:
Biggest problem is no sound from YouTube (or probably from anywhere). Sound works for me with FlashPlayer7 and switching back to that makes it work without any restarting (so it did not permanently mess up sound, like some programs can). This is a Mandrake machine, 2.4.22-10mdkenterprise, I really have no idea how I have sound set up, but it works for me in most software.
Yes it fixed places that check for the version number of the flash player.
Popping up the menu with the right button (which I did to check that it reported 9 or 7) would cause Firefox to crash somewhat later. Does not seem to happen with 7. May indicate an overflow of some malloc'd data buffer.
To use, put libflashplayer9.so into ~/.mozilla/plugins and don't rename it. Apparently if it exists it will be loaded in preference to libflashplayer.so. (I wasted some time making a flashplayer.so symbolic link that switched between 7 and 9 before I finally figured out that 9 was being used no matter how I set it. Instead, to switch back to 7, rename libflashplayer9.so to libflashplayer9.so.hidden).
Removal instructions in the readme.txt say to remove libflashplayer.so, not the correct file of libflashplayer9.so.
ldd shows it links in far more libraries than 7 did, lots of gtk stuff. I suspect this is due to Pango (which does I18N text layout) using the gobject library, not because any gtk widgets are being used. This has also been complained about on Cairo (which is supposed to be a drawing library *used* by toolkits like gtk, but because good font layout requires Pango, there is a circular dependency back to gtk!)
I hope you are joking, but I suspect not.
No, you cannot distribute your zillion-line program without source if it contains 300 lines of GPL code.
However, YOU CAN REMOVE THAT 300 LINES and then distribute the rest just fine. This is NOT what SCO is claiming.
Please look up the what the sentence "REMOVE THE 300 LINES" means. It means ONLY 300 lines. It does not mean all your code. Your code is not "infected" one tiny little bit.
Perhaps I have managed to "get the facts" into your little pea brain. But I doubt it.
The error with this argument is that by *default* you are not allowed to copy anything, as it violates copyright. Any user seeing a "(C) somebody" in the help box should assumme they cannot copy anything, and thus they will not violate the GPL. I think showing the GPL (thus explaining that you can copy it if you follow certain rules) is probably counter-productive because it will confuse most people, such as you, into thinking it is a EULA. If the person wants his friend to have a copy they can send the download URL and not violate anything.
Uh, no, normal users do not use root. Even modern Lindows does not come that way.
The problem is that the user may be fooled into running this cool-new-graphics-program that uses this to get root access. I would not be suprised if the exploit could be run by a Windows program running under Wine. In fact, we use NVidia on both Windows and Linux here and many bugs are cross-platform, has anybody checked if this exploit happens in Windows, too?
Microsoft could finally prove they are not assholes. Release it with complete specs and sample output, don't require Windows libraries, allo anybody to read and write it with any software using any license, and PDF will be dead in a few months.
It does sound better: it is output-only (which is really all we care about in PDF), it uses XML, and it supports alpha compositing like SVG does. Unfortunatly doing anything correctly means Microsoft has to admit that Open Source is not an evil cancer. Don't know if they can bear to do this, or if they are even capable of doing it.