The problem is that this is equally likely (some say more so) in a closed-source system.
I think what you describe would be extremely difficult.
The biggest scare is not where a terrorist is a Microsoft employee or an open-source contributer and has to disguise his code. It is where the terrorist actually is in complete control of the closed source, such as actually owning the company. Now the military is not stupid enough to put closed-source into a missle (at least I hope not) but the government and many corporations will all too gladly put in a system, such as a 911 or traffic or infrastructure control or voting machines or bank databases or some other terrorist target that is closed source.
I think the original poster meant "wait for both a mutex to be released or a socket to become readable, returning which ever one comes first." Posix+pthreads does not offer this ability. Only solution I know of is to use a pipe between the threads, and though that has the advantage of being a communication medium as well, I kind of distrust how much overhead it has.
I feel that Linux could clobber any Windows API advantages quite easily by adding "fds" that are easily created for any other kernel object, such as a thread or mutex. All these do is provide another thing to send to poll(), you cannot read/write data to most of them. The advantage of these fd's over pipes is that they hopefully are very tiny, perhaps requiring zero extra data over the object itself, and they "close" when the object they are connected to is destroyed.
Also poll() has to be replaced with a/dev/poll or other arrangement that does not require extreme overhead on each call. I would also like to see a libc-level "call this when x happens" and "wait for the next thing to happen" interface, right now these are tied into the GUI or other toolkits, making it impossible to interoperate.
in Win32 you get told when and don't have to hang around waiting for it to happen
Huh? As far as I know you have to call WaitForMultipleObjects or a similar function. This is identical to select() or poll(), and the calling thread is "waiting around for it to happen".
The only alternative is to have the system actually spawn threads when events happen (not sure if any systems have tried this) or the Unix signal method where a call to the function is pretty much forced onto the stack of an existing thread (this is pretty much a mess as anybody programming signals will tell you).
While they did some things right like having the exit of other threads and condition flags be waited for by the same call (on Unix I have to make pipes just so a thread can tell another that something is happening, which is silly since the thread object exists), they also did some wrong stuff, such as making it impossible to wait for stdin/out and popen() and a number of other Unix-style interfaces. This seems to be a deliberate attempt to make portability difficult, as I don't know why else they would do this.
I don't think they would blatently sabotage your software like that. If they wanted to a more likely attack would be to modify Word so that the saved files are no longer read correctly by your softwar. The other common attack was to add something to Windows that Word can call that makes it work better or faster (but this is probably impossible for a word processor nowadays).
The whole point of the article is silly. Microsoft would have a much easier time releasing a "cheap windows+office" combination than trying to make a Linux release.
The problem is that the potential is equally there, or even greater, in closed source. The whole problem with this FUD argument is that the fact that the code is open source makes no difference. The potential is in ALL software.
The "compiler backdoor" was a theoretical argument. There were no working implementations.
However just yesterday I gave up on a searching in MSDN after most of the day and started doing Google to locate an explanation for a Windows problem. Google revealed in 1/2 hour that I could not write what I wanted (even this information was difficult to locate, but at least somebody finally complained and others in mailing lists confirmed this).
[the problem I was having was trying to figure out if I could get it so that _fopen() took a string in UTF-8. This is really useful if you are processing text in UTF-8 and want to call a closed-source library that may want a filename to open. Although GetACP() can return CP_UTF8 and after reading MSDN I was about 25% sure that if I can set it that way it will do what I want, Google revealed that many people agreed with my conclusion and told me the only way to set this is to change something as Administrator and reboot (!) the computer. Well thanks a lot, Mr Gates, for coming up with a new way to piss me off]
I have many times given up on MSDN and gone to Google for Windows problems. For instance finding a comprehensive list of bitblt codes, or sample source code that actually works. Granted the MSDN pages are much more useful for finding out what the arguments to a function are. Google is much more useful for finding out the name of a function.
Such a change to the GPL is not possible, as the GPL is simply a license that allows you to violate the copyright on the code if you obey certain conditions. Since it is only a license and not a contract, it cannot restrict you from doing things that you can normally legally do, such as run the software.
Microsoft is not threatened by Linux on the desktop. None of those arguments you present have anything to do with Linux use in servers, which is what Microsoft wants to fight, and they have pretty much lost any argument other than FUD there.
Customers who purchased the product with the copyright infringing code would be entitled to sue M$ upon the basis that they were denied the opportunity to download it.
I don't think so. Only the copyright holder can actually bring charges. Certainly in all the GPL cases so far that is what has happened. Users of some infringing router can complain but the purpose of this is to get the attention of a copyright holder, and to generate bad publicity for the infringer. It apparently serves no legal purpose, as it always is true that the router manufacturers have never done anything until they actually heard from a copyright holder (then they capitulate immediately).
A minor bit of sampling of an old Rolling Stones tune "The Last Time", took greed to new heights. Ex Stones manager Allen Klein owns the rights to the tune, so he stirred up enough shit to capture 100% of the band's royalties from the single.
It sounds like he got the royalties from the song, which is not the same as the IP. He could not release his own recording of the song, but he did in effect get a monetary award equal to the amount of money they made from it.
In the case you described, M$ would be guilty of copyright violation. They would have to cease distributing the infringing material. This may mean they would be unable to distribute Windows, at least until they clean-room engineer a replacement for the GPL part.
In no way would it affect the copyright on other parts of Windows. It would not affect existing copies, since that would only punish the end users, who are not guilty of copyright violations.
M$ would also be legally liable as a copyright violator and may be forced to pay damages. There is no precedence for copyright violators being forced to give up IP, and the fact that the copyrighted information allowed infringement throught the GPL license does not change this. In fact even if they at that point willingly GPL'd the entire source code for Windows, it would not remove the copyright violation or their liability for the past behavior.
Now it is possible that Microsoft could offer to release all the source code, and the original copyright holder would consider this a reasonable settlement and would sign an agreement to not sue them further. However it is also possible that Bill Gates could agree to be the author's personal servant for life, and this would be considered the reasonable settlement. Neither of these results are any more legally likely than the other.
You are correct that "release all your source code" is really a settlement offer. In fact releasing the source does not get you out of trouble if the person who's copyright you violated does not accept it as a settlement (if it did you could get out of shoplifting by agreeing to pay the price of the item only if you were caught).
I really doubt that if a GPL copyright violation went to court the actual judgement would force release of source code. There is no precedent for that in any copyright case in history. I would expect monetary damages to be awarded instead, and that the magnitude of the damage to be approximately the amount the copyright violator benifited from the action plus what the original author lost. Plus of course a cease & desist of the copyright violations.
It is true that the potential monetary damages, plus the cost of re-engineering or abandoning a product, is sufficient that almost always the violator agrees out of court to release the source code. But I really doubt it is billions. In fact are there not cases where companies have pulled products and done nothing else? That would indicate the expected monetary damages is zero.
Yes, you are entitled to the source code, and it is available from their site.
You can also redistribute it. Make sure you remove all the copyrighted material such as the SuSE logos and the installation program and help files, however. And make sure you remove any and all non-GPL stuff that you don't have a right to redistribute, such as Acrobat or any other such included programs. And you better recompile everything from scratch so you are sure their is nothing in the binaries that you don't have rights to redistribute. There are probably a lot of other rules, too.
"violators of the GPL" are actually copyright infringers and are subject to the same punishments as copyright infringers. In all copyright cases in history, the maximum punishment has been cease & desist making the illegal copies, and monetary damages.
I have never heard of a copyright infringer being forced to lose rights to other IP of their own. I very much doubt anybody will ever be forced to open source code. This would be like saying the New York Times has to give away all copies of their paper from now on because one of their columns was plagarized. Such ideas are total nonsense, but are always brought up by the enemies of the GPL.
One part of confusion is that the infringer may choose to obey the GPL in exchange for getting the lawsuit threat dropped and to be able to continue distributing their product. But they were not "forced" by the GPL to do this. In fact, legally, it does not in any way get them out of their liability for the previous copyright violations (otherwise you could violate the GPL for years and then release the source code at the end as a "get out of jail free" card).
Okay, all I know is that SPF is a good deal simpler than SenderID and much more popular, due to the simple text format verses the use of XML.
However can somebody please clearly explain what (if any) differences there are between what they do. I mean after the data is decoded, is one of the superior to the other, or a superset of the other? Or are they totally independent checks, or are they slightly intersecting checks?
Honestly I can say I am extremely happy to see Microsoft adopting a standard that was not proposed by them. They should learn from this, the amount of good feelings they engender by doing this and resulting increses in sales of their other products and increased cooperation by other programmers probably outweighs any monetary gain from a proprietary solution by a hundred fold or more.
I have to second the common opinion here that something is screwed up in these estimates. I have NEVER seen a machine with Linux pre-installed in a store.
My last two machines were blank, built by a computer shop, and I installed Linux on them. I assumme that Microsoft did not get their tax for these, but it is really hard to tell.
Anyway, it is vastly easier to get a no-os machine than a Linux-preinstalled machine. Anybody intending to install a pirate copy of Windows would get one of these. Yet this report seems to indicate sales of no-os machines is zero? WTF?
The Lindows machines sold by Walmart are obviosly junk, available only by mail order, and if you look on the same web page you will see cheaper no-os machines, and I would not be suprised if they have better hardware because they don't have to worry about Linux compatability. They certainly sell much better-quality no-os machines (much faster and much larger hard disks) for only slightly more than the Lindows box.
Linux fans might fear admitting it, but I suspect even a Linux user who buys a Lindows machine purposely to add to the sales figures will wipe it and reinstall it as a dual-boot with Windows configuration. I did, even though I don't ever use the Windows system, it is nice to know I can boot to it if I need to run something (yes it is a pirated copy, so sue me Mr Gates).
The problem is not that it "renders wrong" but that the page does not come up at all unless the browser identifies itself as IE.
I'm not sure how much this is skewing the results. About 3 years ago I was forced to change the browser ID to IE to get sites to work and I would estimate about 50% of the Mozilla users did. However nowadays I don't bother, and I have really not had any trouble (I'm using Konqueror). Since it is a pain to change, and does not seem as necessary, I suspect the claims that this is skewing down the Linux results is exaggerated. Also the newer browsers make "identify as IE only to this site" possible which would fix Google's statistics.
Maybe some kind of accurate estimate of system usage can be made by somehow comparing Firefox usage on Windows to Linux and to IE usage on Windows, but this is definately complex statistical math. Certainly though I am a Linux user, I use IE whenever I have to use a Windows machine, not Firefox.
A solution to this problem that allows for applications to be downloaded from webpages and installed directly would remedy the situations that occurred during the usability study
I also don't like this. Unfortunatly Win32 has convinced too many people that an appliation has to be "installed".
Here is what I would like to see:
1. User goes to web page, sees kool new application, and clicks "download".
2. This downloads a thing that looks like a file, exactly like they decided to download a.pdf or piece of music.
3. They can double-click on the file (or the browser can offer to run it). This RUNS the appliation. It does not "install" it.
4. If they hate the appliation they can quit, and they throw the file in the trash, and it is GONE!!!
5. If they like it they can drag and drop it in a place so they can run it from a menu. They can drag it to a place so that all users can run it from the menu, this will require the root password. To remove it from the menus they drag it out of there, either into the trash, or to somewhere else to keep it.
6. Any application that needs root privledges to run will fake it's operation as well as possible, so you can at least see what it does, and offer to install itself. This is done by calling sudo and asking for the root password. If the user does this the icon is moved to the root location. They can still "uninstall" by throwing the icon in the trash, if they have permission to do this.
How to accomplish? This would be done with something similar to Mac "app" directories, though I recommend that no extension be used. Probably these are detected by having a particular file inside them. File browser would act exactly like these are single files. Running them would run a fixed-name program inside them (main?) and set argv[0] to the full pathname, and add the directory to the start of the LD_LIBRARY_PATH. All the information used in.desktop files would be placed inside these directories so just dropping it on the menu makes it work.
Typing "ls -a" is even better. You can learn lots about how to get the desktop to do what you want by looking at directories whose names start with a period.
Has anybody considered making home *be* the "desktop", rather than using a subdirectory called Desktop? That would really match the original design of Unix. It could just hide anything starting with period. Yes existing users would see a lot of clutter, but they may learn to stick it all in folders on the "desktop".
The term "window" to refer to an area of a screen display dates at least back to 1979. More likely it dates back to 1960 or so. Notice that Microsoft's previous system was call "MSDOS", not "DOS" or "Disk" or "Operating System" or "Files", because there is no way they could have gotten away with such naming.
Did you notice that 2/3 of your complaints are not user interface complaints??? Probably not.
One huge problem with GUI is that all fixes to it are considered minor. It does not help when people who think they are complaining about the GUI complain about other things (such as speed and compatability) and don't even realize it.
If your "speed' problem could be fixed by "oh the user just has to go into the registry and add/hkey/gofaster" then probably it would not be addressed, since it is "possible" to fix it. Unfortunatly most GUI complaints are also dismissed as well.
I can guarantee you that any OO developer reading your rant would decide to work #1 on speed, and #2 on the Office compatability. Your complaint about tables of contents is probably invisible, or will be sent to the "toc guy" who will click the menu items he designed and says "works for me"
One thing that might work is to complain specificly and accurately about the GUI. Say "the secretaries always hit button x thinking it would make their table of contents". And don't say anything that will distract the programmer like "hey it's also dog slow". Unfortunately observing exactly what is wrong with the GUI is probably far harder than fixing it.
Slashdot Mirror
The problem is that this is equally likely (some say more so) in a closed-source system.
I think what you describe would be extremely difficult.
The biggest scare is not where a terrorist is a Microsoft employee or an open-source contributer and has to disguise his code. It is where the terrorist actually is in complete control of the closed source, such as actually owning the company. Now the military is not stupid enough to put closed-source into a missle (at least I hope not) but the government and many corporations will all too gladly put in a system, such as a 911 or traffic or infrastructure control or voting machines or bank databases or some other terrorist target that is closed source.
I think the original poster meant "wait for both a mutex to be released or a socket to become readable, returning which ever one comes first." Posix+pthreads does not offer this ability. Only solution I know of is to use a pipe between the threads, and though that has the advantage of being a communication medium as well, I kind of distrust how much overhead it has.
/dev/poll or other arrangement that does not require extreme overhead on each call. I would also like to see a libc-level "call this when x happens" and "wait for the next thing to happen" interface, right now these are tied into the GUI or other toolkits, making it impossible to interoperate.
I feel that Linux could clobber any Windows API advantages quite easily by adding "fds" that are easily created for any other kernel object, such as a thread or mutex. All these do is provide another thing to send to poll(), you cannot read/write data to most of them. The advantage of these fd's over pipes is that they hopefully are very tiny, perhaps requiring zero extra data over the object itself, and they "close" when the object they are connected to is destroyed.
Also poll() has to be replaced with a
in Win32 you get told when and don't have to hang around waiting for it to happen
Huh? As far as I know you have to call WaitForMultipleObjects or a similar function. This is identical to select() or poll(), and the calling thread is "waiting around for it to happen".
The only alternative is to have the system actually spawn threads when events happen (not sure if any systems have tried this) or the Unix signal method where a call to the function is pretty much forced onto the stack of an existing thread (this is pretty much a mess as anybody programming signals will tell you).
While they did some things right like having the exit of other threads and condition flags be waited for by the same call (on Unix I have to make pipes just so a thread can tell another that something is happening, which is silly since the thread object exists), they also did some wrong stuff, such as making it impossible to wait for stdin/out and popen() and a number of other Unix-style interfaces. This seems to be a deliberate attempt to make portability difficult, as I don't know why else they would do this.
I don't think they would blatently sabotage your software like that. If they wanted to a more likely attack would be to modify Word so that the saved files are no longer read correctly by your softwar. The other common attack was to add something to Windows that Word can call that makes it work better or faster (but this is probably impossible for a word processor nowadays).
The whole point of the article is silly. Microsoft would have a much easier time releasing a "cheap windows+office" combination than trying to make a Linux release.
The problem is that the potential is equally there, or even greater, in closed source. The whole problem with this FUD argument is that the fact that the code is open source makes no difference. The potential is in ALL software.
The "compiler backdoor" was a theoretical argument. There were no working implementations.
I'm pretty sure the article is biased.
However just yesterday I gave up on a searching in MSDN after most of the day and started doing Google to locate an explanation for a Windows problem. Google revealed in 1/2 hour that I could not write what I wanted (even this information was difficult to locate, but at least somebody finally complained and others in mailing lists confirmed this).
[the problem I was having was trying to figure out if I could get it so that _fopen() took a string in UTF-8. This is really useful if you are processing text in UTF-8 and want to call a closed-source library that may want a filename to open. Although GetACP() can return CP_UTF8 and after reading MSDN I was about 25% sure that if I can set it that way it will do what I want, Google revealed that many people agreed with my conclusion and told me the only way to set this is to change something as Administrator and reboot (!) the computer. Well thanks a lot, Mr Gates, for coming up with a new way to piss me off]
I have many times given up on MSDN and gone to Google for Windows problems. For instance finding a comprehensive list of bitblt codes, or sample source code that actually works. Granted the MSDN pages are much more useful for finding out what the arguments to a function are. Google is much more useful for finding out the name of a function.
Such a change to the GPL is not possible, as the GPL is simply a license that allows you to violate the copyright on the code if you obey certain conditions. Since it is only a license and not a contract, it cannot restrict you from doing things that you can normally legally do, such as run the software.
Microsoft is not threatened by Linux on the desktop. None of those arguments you present have anything to do with Linux use in servers, which is what Microsoft wants to fight, and they have pretty much lost any argument other than FUD there.
Customers who purchased the product with the copyright infringing code would be entitled to sue M$ upon the basis that they were denied the opportunity to download it.
I don't think so. Only the copyright holder can actually bring charges. Certainly in all the GPL cases so far that is what has happened. Users of some infringing router can complain but the purpose of this is to get the attention of a copyright holder, and to generate bad publicity for the infringer. It apparently serves no legal purpose, as it always is true that the router manufacturers have never done anything until they actually heard from a copyright holder (then they capitulate immediately).
Bit of Googling found several quotes like this:
A minor bit of sampling of an old Rolling Stones tune "The Last Time", took greed to new heights. Ex Stones manager Allen Klein owns the rights to the tune, so he stirred up enough shit to capture 100% of the band's royalties from the single.
It sounds like he got the royalties from the song, which is not the same as the IP. He could not release his own recording of the song, but he did in effect get a monetary award equal to the amount of money they made from it.
No.
In the case you described, M$ would be guilty of copyright violation. They would have to cease distributing the infringing material. This may mean they would be unable to distribute Windows, at least until they clean-room engineer a replacement for the GPL part.
In no way would it affect the copyright on other parts of Windows. It would not affect existing copies, since that would only punish the end users, who are not guilty of copyright violations.
M$ would also be legally liable as a copyright violator and may be forced to pay damages. There is no precedence for copyright violators being forced to give up IP, and the fact that the copyrighted information allowed infringement throught the GPL license does not change this. In fact even if they at that point willingly GPL'd the entire source code for Windows, it would not remove the copyright violation or their liability for the past behavior.
Now it is possible that Microsoft could offer to release all the source code, and the original copyright holder would consider this a reasonable settlement and would sign an agreement to not sue them further. However it is also possible that Bill Gates could agree to be the author's personal servant for life, and this would be considered the reasonable settlement. Neither of these results are any more legally likely than the other.
Yea you are right. If the binary needed to be recompiled and was based on GPL code then SuSE has actually violated the GPL.
You are correct that "release all your source code" is really a settlement offer. In fact releasing the source does not get you out of trouble if the person who's copyright you violated does not accept it as a settlement (if it did you could get out of shoplifting by agreeing to pay the price of the item only if you were caught).
I really doubt that if a GPL copyright violation went to court the actual judgement would force release of source code. There is no precedent for that in any copyright case in history. I would expect monetary damages to be awarded instead, and that the magnitude of the damage to be approximately the amount the copyright violator benifited from the action plus what the original author lost. Plus of course a cease & desist of the copyright violations.
It is true that the potential monetary damages, plus the cost of re-engineering or abandoning a product, is sufficient that almost always the violator agrees out of court to release the source code. But I really doubt it is billions. In fact are there not cases where companies have pulled products and done nothing else? That would indicate the expected monetary damages is zero.
You are joking, right?
Yes, you are entitled to the source code, and it is available from their site.
You can also redistribute it. Make sure you remove all the copyrighted material such as the SuSE logos and the installation program and help files, however. And make sure you remove any and all non-GPL stuff that you don't have a right to redistribute, such as Acrobat or any other such included programs. And you better recompile everything from scratch so you are sure their is nothing in the binaries that you don't have rights to redistribute. There are probably a lot of other rules, too.
"violators of the GPL" are actually copyright infringers and are subject to the same punishments as copyright infringers. In all copyright cases in history, the maximum punishment has been cease & desist making the illegal copies, and monetary damages.
I have never heard of a copyright infringer being forced to lose rights to other IP of their own. I very much doubt anybody will ever be forced to open source code. This would be like saying the New York Times has to give away all copies of their paper from now on because one of their columns was plagarized. Such ideas are total nonsense, but are always brought up by the enemies of the GPL.
One part of confusion is that the infringer may choose to obey the GPL in exchange for getting the lawsuit threat dropped and to be able to continue distributing their product. But they were not "forced" by the GPL to do this. In fact, legally, it does not in any way get them out of their liability for the previous copyright violations (otherwise you could violate the GPL for years and then release the source code at the end as a "get out of jail free" card).
Okay, all I know is that SPF is a good deal simpler than SenderID and much more popular, due to the simple text format verses the use of XML.
However can somebody please clearly explain what (if any) differences there are between what they do. I mean after the data is decoded, is one of the superior to the other, or a superset of the other? Or are they totally independent checks, or are they slightly intersecting checks?
Honestly I can say I am extremely happy to see Microsoft adopting a standard that was not proposed by them. They should learn from this, the amount of good feelings they engender by doing this and resulting increses in sales of their other products and increased cooperation by other programmers probably outweighs any monetary gain from a proprietary solution by a hundred fold or more.
I have to second the common opinion here that something is screwed up in these estimates. I have NEVER seen a machine with Linux pre-installed in a store.
My last two machines were blank, built by a computer shop, and I installed Linux on them. I assumme that Microsoft did not get their tax for these, but it is really hard to tell.
Anyway, it is vastly easier to get a no-os machine than a Linux-preinstalled machine. Anybody intending to install a pirate copy of Windows would get one of these. Yet this report seems to indicate sales of no-os machines is zero? WTF?
The Lindows machines sold by Walmart are obviosly junk, available only by mail order, and if you look on the same web page you will see cheaper no-os machines, and I would not be suprised if they have better hardware because they don't have to worry about Linux compatability. They certainly sell much better-quality no-os machines (much faster and much larger hard disks) for only slightly more than the Lindows box.
Linux fans might fear admitting it, but I suspect even a Linux user who buys a Lindows machine purposely to add to the sales figures will wipe it and reinstall it as a dual-boot with Windows configuration. I did, even though I don't ever use the Windows system, it is nice to know I can boot to it if I need to run something (yes it is a pirated copy, so sue me Mr Gates).
The problem is not that it "renders wrong" but that the page does not come up at all unless the browser identifies itself as IE.
I'm not sure how much this is skewing the results. About 3 years ago I was forced to change the browser ID to IE to get sites to work and I would estimate about 50% of the Mozilla users did. However nowadays I don't bother, and I have really not had any trouble (I'm using Konqueror). Since it is a pain to change, and does not seem as necessary, I suspect the claims that this is skewing down the Linux results is exaggerated. Also the newer browsers make "identify as IE only to this site" possible which would fix Google's statistics.
Maybe some kind of accurate estimate of system usage can be made by somehow comparing Firefox usage on Windows to Linux and to IE usage on Windows, but this is definately complex statistical math. Certainly though I am a Linux user, I use IE whenever I have to use a Windows machine, not Firefox.
Maybe 2.6 should have been called LinuxME.
A solution to this problem that allows for applications to be downloaded from webpages and installed directly would remedy the situations that occurred during the usability study
.pdf or piece of music.
.desktop files would be placed inside these directories so just dropping it on the menu makes it work.
I also don't like this. Unfortunatly Win32 has convinced too many people that an appliation has to be "installed".
Here is what I would like to see:
1. User goes to web page, sees kool new application, and clicks "download".
2. This downloads a thing that looks like a file, exactly like they decided to download a
3. They can double-click on the file (or the browser can offer to run it). This RUNS the appliation. It does not "install" it.
4. If they hate the appliation they can quit, and they throw the file in the trash, and it is GONE!!!
5. If they like it they can drag and drop it in a place so they can run it from a menu. They can drag it to a place so that all users can run it from the menu, this will require the root password. To remove it from the menus they drag it out of there, either into the trash, or to somewhere else to keep it.
6. Any application that needs root privledges to run will fake it's operation as well as possible, so you can at least see what it does, and offer to install itself. This is done by calling sudo and asking for the root password. If the user does this the icon is moved to the root location. They can still "uninstall" by throwing the icon in the trash, if they have permission to do this.
How to accomplish? This would be done with something similar to Mac "app" directories, though I recommend that no extension be used. Probably these are detected by having a particular file inside them. File browser would act exactly like these are single files. Running them would run a fixed-name program inside them (main?) and set argv[0] to the full pathname, and add the directory to the start of the LD_LIBRARY_PATH. All the information used in
Typing "ls -a" is even better. You can learn lots about how to get the desktop to do what you want by looking at directories whose names start with a period.
Has anybody considered making home *be* the "desktop", rather than using a subdirectory called Desktop? That would really match the original design of Unix. It could just hide anything starting with period. Yes existing users would see a lot of clutter, but they may learn to stick it all in folders on the "desktop".
I just have one thing to say:
#ifndef DDImage_API
# ifdef _WIN32
# if defined(DDImage_EXPORTS) || defined(DDImage_beta_EXPORTS)
# define DDImage_API __declspec(dllexport)
# else
# define DDImage_API __declspec(dllimport)
# endif
# else
# define DDImage_API
# endif
#endif
And go stick "DDImage_API" in front of everything in your header files.
Thank you Microsoft for your "slick" development tools.
The term "window" to refer to an area of a screen display dates at least back to 1979. More likely it dates back to 1960 or so. Notice that Microsoft's previous system was call "MSDOS", not "DOS" or "Disk" or "Operating System" or "Files", because there is no way they could have gotten away with such naming.
Did you notice that 2/3 of your complaints are not user interface complaints??? Probably not.
/hkey/gofaster" then probably it would not be addressed, since it is "possible" to fix it. Unfortunatly most GUI complaints are also dismissed as well.
One huge problem with GUI is that all fixes to it are considered minor. It does not help when people who think they are complaining about the GUI complain about other things (such as speed and compatability) and don't even realize it.
If your "speed' problem could be fixed by "oh the user just has to go into the registry and add
I can guarantee you that any OO developer reading your rant would decide to work #1 on speed, and #2 on the Office compatability. Your complaint about tables of contents is probably invisible, or will be sent to the "toc guy" who will click the menu items he designed and says "works for me"
One thing that might work is to complain specificly and accurately about the GUI. Say "the secretaries always hit button x thinking it would make their table of contents". And don't say anything that will distract the programmer like "hey it's also dog slow". Unfortunately observing exactly what is wrong with the GUI is probably far harder than fixing it.