Cairo uses FreeType and thus will render the same as Xft. It may be able to take advantage of sub-pixel character widths, but since Windows has this problem in any GDI32-based program, it probably is not what you are complaining about.
Generally "turning on hinting", which often involves compiling or copying a new binary from somewhere other than a distribution, means enabling the code that is covered by Apple's patents, so I suspect you did this. Personally I have seen no improvement over the modern autohinters.
My personal preference is to turn on AA on everything, I think the defaults for both Windows and Linux look like crap. However my preference is called "blurry" by huge numbers of people. For some reason people don't complain about OS/X being "blurry", yet quick comparisons with the same (Windows) font shows that OS/X, Linux with FreeType AA, and Windows with "cleartype" (not "font smoothing") turned on all look nearly identical, but somehow Linux and Windows generate complaints, perhaps because people are not used to it.
Turning off antialiasing does reveal deficiencies in the FreeType rendering (who knows about OS/X as I don't know if it can be turned off), but I would prefer to get people to accept antialiased text more, rather than try to reproduce Windows, if in fact the problem is that people are just not used to it. I also don't see how Windows (or anybody) can do the "3D desktop" through texturemaps unless they accept antialiasing on all the fonts.
Okay, I've had enough of this. Post somewhere two damn screen shots. Showing the SAME font copied from Windows, and showing "good" Windows rendering verses "bad" Linux rendering. I fully expect they are different and you have some legitimate complaint, but it is impossible to tell what it is. Some people don't like the antialiasing, yet when you turn it off some people of the opposite persuasion complain. Maybe you have the Linux version compiled without the hinting. It really is difficult to tell what is going on without a screen shot from both, perhaps with indicators showing exactly what differences are the objectionable ones, and when there are differences, which example you think is better.
Efficient support for very small files would allow a lot of crap on Windows/Linux (such as the Registry and the Gnome copy of it) to be eliminated, and allow all "metadata" (things like the artist in a song) to be stored as files. Read up on the ReiserFS plans. Ideally 99.99% of the files on a disk would be less than 100 bytes.
However this is best solved by writing the filesystem to put these small files into the blocks with other data, such as many small files together, or inserted directly into the directory. Correctly done you would get a bunch of small files at once with a single read, and since use of this would probably need to look at many at once, this could be very efficient. In any case larger sectors are harmless.
Though I agree that working only to pass the Acid test is probably not a good idea, it is somewhat different than the MP3 encoding test. The MP3 encoding is lossy.
The Acid test is more like a test for a lossless encoding format. If the lossless encoder fails the test then the encoder is not working, because it is not lossless. However if the programmers worked on the Acid test exclusivley it is quite likely the result will encode that fine but fail on other data (or be really slow on that other data).
Non-compliant HTML is perhaps like music that exceeds the dynamic range that the lossless encoder does, or something like that (I know nothing about sound). The lossless encoder should do as well as it can with that, but there is no reason why that should make it fail to be lossless with legal data.
A correct browser would render the Acid test perfectly, no matter whether the programmers paid attention to it or not.
"Liberal" should mean something more like what Libertarians believe. I mean look at the name "Libertarian"! There is an obvious relationship.
The stuff normally labelled "Liberal" and hated by the right can more accurately be called "Socialism" (I probably fall into this mindset so I am not trying to insult it).
"Conservative" is probably the right term for the neo-cons. It means they don't want to change things, or want to undo changes made recently.
free people + free markets = libertarian
free people + controlled markets = socialism / democrats
controlled people + free markets = conservative / republicans
controlled people + controlled markets = communism (at least in practice)
If root screwed up/etc/passwd (or any of dozens of other files such as the init.rc stuff) and then logged out, you would be just as hosed. I really don't see how sudo is worse.
You seem to be missing the fact that the GPL would allow Apple to release their own version under exactly the same terms, reusing all that MicroSoft+IBM work.
I suspect you are thinking that "being able to see the source code" == "GPL". If in fact Microsoft+IBM gave away the software for free and did not give anybody else the ability to give away the same software for free (ie by either keeping it closed source, or releasing the source with normal copyright and no GPL exceptions to the copyright restrictions), then Apple would have a complaint.
All I know is that our Windows developer had to replace the assert.h with his own version that does an assembler INT instruction (12?). The normal assert causes the program to exit in such a way that no stack trace was available.
I'm not sure all your comments are very accurate, perhaps this is just a troll.
I would agree with lots of the other posters here is that lack of stdin/stdout and sockets that are the same thing as normal fd's is probably the biggest problem with developing Windows software. You would think they would at least make it so printf would output into the IDE's debugger output window, or make assert(), exit(), and abort() work (if these interacted with the IDE so that you could continue the program it would be a huge improvement over Unix, but instead they don't work at all). All I can guess is that they are scared to death of allowing people to write software that can be ported to Unix and they will cripple their own stuff just to avoid this.
It would help if you described exactly what went wrong. What "precaution" did the instructions not say to do? Can you explain exactly what alternative to GRUB will allow dual-boot and is less dangerous? Can you list anything else? Can you at least give a HINT as to what alternative you would expect that would somehow be better?
Can you please cut out a piece of those Novell screen shots that shows a "non antialiased letter". I looked everywhere. There were circular radio buttons that were not antialiased, but all letters were.
As a totally seperate thing, I would appreciate a screen shot cut out of there showing what you think looks "bad", along with similar size text off of OS/X and Windows that you think looks "good". I do think there are some spacing issues, perhaps that is what you are complaining about.
I think Linux and OS/X, and the newer ClearType antialiasing on Windows all look virtually identical, all using the same algorithims. Normally the complaints about the antialiasing seem to be from Windows users that are used to "font smoothing", these are the people who complain that antialising (including Cleartype) looks "fuzzy". However those people also complain about OS/X, while you seem to praise OS/X, so I'm am somewhat stumped as to what you are getting at.
Also it is not PDF, but Quartz, that is responsible for the antialiasing in OS/X. I believe they also have a nicer selection of fonts, both Linux and Windows can be improved by copying the OS/X fonts.
Why don't Windows and Linux get rid of all that initialization code, delay building structures until first use, so that the damn window shows up within a half second of when you launched the program? This used to work on 1Mhz X terminals so it certainly should be possible on our 2Ghz processors today! Then there would not be any worry about this focus stealing.
Of course this is probably too hard for modern programmers to do, huh.
This is exactly what I think is happening with these exit polls. It is not fraud.
Lets say for instance that 2% of all Bush voters feel guilty about it and claim they voted for Kerry, while there are no other lies on the exit polls.
In Texas with (I don't know, lets guess) 75% for Bush, the exit polls would then say 73.5% for Bush. Everybody looks at this and says it is well within any error. In California (again with a guess number) 75% for Kerry, the exit polls say 75.5% for Kerry, again well within any error. But in Ohio, with it 50.5% for Bush, the exit polls say it is 49.5% for Bush and 50.5 for Kerry, and the exit polls now disagree!
Even though the error is consistent all across the country, nobody will think there is anything wrong except in Ohio. It seems amplified in the close states, making it look like there is some kind of planned fraud in the close states.
If the "or other" clause was included, somebody could make copy and say that their changes are GPL3 only. If their changes were popular enough (ie if they were from the core developers) they would effectively change Linux to GPL3, since the only GPL2 code would be useless outdated stuff that does not interoperate.
In fact this is exactly the same as why BSD code can be "stolen" by a company, despite claims to the contrary by many posters here. Anybody who thinks the GPL3 is dangerous should have to agree that BSD is even more dangerous.
It certainly is VM. Memory mapping of files is pretty much part of the VM system. Predictive loading of memory mapped files thus would also be part of the VM system, in that it would not work at all if the VM system did not somehow know it had been done. Since it is likely based on what things were memory mapped before, it is pretty obvious that the VM system would also be in charge of remembering this information and doing it a second time.
Not really sure what you are getting at and why you seem so hurt that somebody says "its part of the VM system".
Yes, that makes sense. Certainly there are reasons for a company to not want to use the GPL. However the reason is NOT "we might accidentally GPL our code". The reason is "we can't redistribute the program unless we GPL the code or we rewrite it to not use the GPL code". There is a HUGE difference, in that the first implies that just using GPL code could be a dangerous and once you do it there is no way to avoid the danger. In fact it is trivial to avoid the danger. This difference in perception is why I call this FUD.
If they used GPL code in project A, and then distributed project A to external customers wihout the source code, then they are guilty of copyright violation. It does not mean they have to distribute the code for A.
It is true that it is likely the copyright owner will agree to not sue if they start distributing the code for A. However nothing in the law says this. Deciding on their own to distribute the code does not get them out of it (otherwise it would be a huge GPL loophole, as a company could wait a few years and then distribute the code when it is no longer worth anything, and suddenly all their previous violations are cancelled). And the courts cannot force them to do it, they can only force them to stop distribution and to pay monetary compensation to the copyright holder.
Yes I agree. Sorry if my post made it look like DRM is actually possible.
I do believe that a true open source DRM would result in a design such that it is extremely hard to determine the key, for instance where it is written such that the key itself is never stored in the resulting program, and other hacks. But it will never be impossible, for the simple reasons you state.
No your company is a bunch of idiots who have bought the FUD from Microsoft.
they don't want to inadvertantly make all of thier work go under the GPL if it's distributed
This is WRONG. It DOES NOT HAPPEN, it is a lie from Microsoft and one of the biggest and hardest to eradicate.
If you distribute without obeying the GPL, you are VIOLATING COPYRIGHT. You are then required to cease distribution, and you may be liable for monetary damages. You are NOT "forced to give away your source code". In fact even if you put every piece of source code you have in the public domain, it does not mean you did not violate copyright, so you are still liable. So not only are you not forced to give away the code, doing so does not get you out of anything, so this lie is doubly untrue!
Think hard. If the NY Tiimes was accused of plagarism, would this cause all their articles to now be the property of the person who wrote the one copied article?
You deleted the very next paragraph which describes the licenses rules for this "not precisely defined by law" statement:
If such an object file uses only numerical parameters, data structure layouts and accessors, and small macros and small inline functions (ten lines or less in length), then the use of the object file is unrestricted, regardless of whether it is legally a derivative work. (Executables containing this object code plus portions of the Library will still fall under Section 6.)
And you deleted the very next paragraph, which says that all this is only describing things where the LGPL is totally irrelevant because it is not a copyright violation. Otherwise the LGPL is explicitly granting you rights to violate the copyright:
Otherwise, if the work is a derivative of the Library, you may distribute the object code for the work under the terms of Section 6. Any executables containing that work also fall under Section 6, whether or not they are linked directly with the Library itself.
You also deleted all of "section 6" which describes how you can distribute without source code.
Now personally I think this section has some nasty problems in that RMS claims the end user should be able to change the library and relink, which is seriously out of touch with how modern software works, it would only be possible by distributing (perhaps obfuscated) source code to your program, making the difference between the LGPL and GPL nearly meaningless. I add an exception to my LGPL code to explicitly state that static linking, or any other use of the code that does not modify the code itself, it allowed. Adding such an exception is very common and I wish there was an official version of this, or the LGPL was fixed in this way.
It is quite possible to provide the source code so that anybody can make a music player that uses "Open DRM". The only problem is that the resulting music player won't play any commercial music. The secret is kept in a key which is different than the test ones provided with the source code. In fact this would allow the engineers to work on debugging the machine and adding features while only working with a set of "test" music thus making all the info they have worthless for breaking the DRM.
If the code is open-source, I'm quite certain that it would be analyzed so completely by thousands of people trying to prove they are smart by fixing it, that it would be quite impossible to get a key out of a working version, a level of security that the RIAA and MS could only dream of.
Don't be a moron. Dell has no responsibility to support software it does not sell with the machine. They are selling RedHat version something-or-other. This does not mean they have to support anything else that happens to have the letters L,I,N,U, and X in it's name.
Slashdot Mirror
Cairo uses FreeType and thus will render the same as Xft. It may be able to take advantage of sub-pixel character widths, but since Windows has this problem in any GDI32-based program, it probably is not what you are complaining about.
Generally "turning on hinting", which often involves compiling or copying a new binary from somewhere other than a distribution, means enabling the code that is covered by Apple's patents, so I suspect you did this. Personally I have seen no improvement over the modern autohinters.
My personal preference is to turn on AA on everything, I think the defaults for both Windows and Linux look like crap. However my preference is called "blurry" by huge numbers of people. For some reason people don't complain about OS/X being "blurry", yet quick comparisons with the same (Windows) font shows that OS/X, Linux with FreeType AA, and Windows with "cleartype" (not "font smoothing") turned on all look nearly identical, but somehow Linux and Windows generate complaints, perhaps because people are not used to it.
Turning off antialiasing does reveal deficiencies in the FreeType rendering (who knows about OS/X as I don't know if it can be turned off), but I would prefer to get people to accept antialiased text more, rather than try to reproduce Windows, if in fact the problem is that people are just not used to it. I also don't see how Windows (or anybody) can do the "3D desktop" through texturemaps unless they accept antialiasing on all the fonts.
Trying to make MS-supporters look silly by posting fake stuff like this does not work.
Okay, I've had enough of this. Post somewhere two damn screen shots. Showing the SAME font copied from Windows, and showing "good" Windows rendering verses "bad" Linux rendering. I fully expect they are different and you have some legitimate complaint, but it is impossible to tell what it is. Some people don't like the antialiasing, yet when you turn it off some people of the opposite persuasion complain. Maybe you have the Linux version compiled without the hinting. It really is difficult to tell what is going on without a screen shot from both, perhaps with indicators showing exactly what differences are the objectionable ones, and when there are differences, which example you think is better.
Entertainment and media centre functionality are part of the OS when you consider Digital Restrictions Management.
If this article is accurate, almost certainly the problems are with DRM, not with getting a picture to show up on the monitor.
Efficient support for very small files would allow a lot of crap on Windows/Linux (such as the Registry and the Gnome copy of it) to be eliminated, and allow all "metadata" (things like the artist in a song) to be stored as files. Read up on the ReiserFS plans. Ideally 99.99% of the files on a disk would be less than 100 bytes.
However this is best solved by writing the filesystem to put these small files into the blocks with other data, such as many small files together, or inserted directly into the directory. Correctly done you would get a bunch of small files at once with a single read, and since use of this would probably need to look at many at once, this could be very efficient. In any case larger sectors are harmless.
Though I agree that working only to pass the Acid test is probably not a good idea, it is somewhat different than the MP3 encoding test. The MP3 encoding is lossy.
The Acid test is more like a test for a lossless encoding format. If the lossless encoder fails the test then the encoder is not working, because it is not lossless. However if the programmers worked on the Acid test exclusivley it is quite likely the result will encode that fine but fail on other data (or be really slow on that other data).
Non-compliant HTML is perhaps like music that exceeds the dynamic range that the lossless encoder does, or something like that (I know nothing about sound). The lossless encoder should do as well as it can with that, but there is no reason why that should make it fail to be lossless with legal data.
A correct browser would render the Acid test perfectly, no matter whether the programmers paid attention to it or not.
Seems to me that the labels are all messed up.
"Liberal" should mean something more like what Libertarians believe. I mean look at the name "Libertarian"! There is an obvious relationship.
The stuff normally labelled "Liberal" and hated by the right can more accurately be called "Socialism" (I probably fall into this mindset so I am not trying to insult it).
"Conservative" is probably the right term for the neo-cons. It means they don't want to change things, or want to undo changes made recently.
free people + free markets = libertarian
free people + controlled markets = socialism / democrats
controlled people + free markets = conservative / republicans
controlled people + controlled markets = communism (at least in practice)
If root screwed up /etc/passwd (or any of dozens of other files such as the init.rc stuff) and then logged out, you would be just as hosed. I really don't see how sudo is worse.
You seem to be missing the fact that the GPL would allow Apple to release their own version under exactly the same terms, reusing all that MicroSoft+IBM work.
I suspect you are thinking that "being able to see the source code" == "GPL". If in fact Microsoft+IBM gave away the software for free and did not give anybody else the ability to give away the same software for free (ie by either keeping it closed source, or releasing the source with normal copyright and no GPL exceptions to the copyright restrictions), then Apple would have a complaint.
All I know is that our Windows developer had to replace the assert.h with his own version that does an assembler INT instruction (12?). The normal assert causes the program to exit in such a way that no stack trace was available.
Alt+Shift+Tab works fine on Windows.
I'm not sure all your comments are very accurate, perhaps this is just a troll.
I would agree with lots of the other posters here is that lack of stdin/stdout and sockets that are the same thing as normal fd's is probably the biggest problem with developing Windows software. You would think they would at least make it so printf would output into the IDE's debugger output window, or make assert(), exit(), and abort() work (if these interacted with the IDE so that you could continue the program it would be a huge improvement over Unix, but instead they don't work at all). All I can guess is that they are scared to death of allowing people to write software that can be ported to Unix and they will cripple their own stuff just to avoid this.
It would help if you described exactly what went wrong. What "precaution" did the instructions not say to do? Can you explain exactly what alternative to GRUB will allow dual-boot and is less dangerous? Can you list anything else? Can you at least give a HINT as to what alternative you would expect that would somehow be better?
Whoa! The clue train has left the building!
The $100 is the cost of manufacture. This is so that they can be subsidized and given away for free by a government or a charitable organization.
Can you please cut out a piece of those Novell screen shots that shows a "non antialiased letter". I looked everywhere. There were circular radio buttons that were not antialiased, but all letters were.
As a totally seperate thing, I would appreciate a screen shot cut out of there showing what you think looks "bad", along with similar size text off of OS/X and Windows that you think looks "good". I do think there are some spacing issues, perhaps that is what you are complaining about.
I think Linux and OS/X, and the newer ClearType antialiasing on Windows all look virtually identical, all using the same algorithims. Normally the complaints about the antialiasing seem to be from Windows users that are used to "font smoothing", these are the people who complain that antialising (including Cleartype) looks "fuzzy". However those people also complain about OS/X, while you seem to praise OS/X, so I'm am somewhat stumped as to what you are getting at.
Also it is not PDF, but Quartz, that is responsible for the antialiasing in OS/X. I believe they also have a nicer selection of fonts, both Linux and Windows can be improved by copying the OS/X fonts.
Why don't Windows and Linux get rid of all that initialization code, delay building structures until first use, so that the damn window shows up within a half second of when you launched the program? This used to work on 1Mhz X terminals so it certainly should be possible on our 2Ghz processors today! Then there would not be any worry about this focus stealing.
Of course this is probably too hard for modern programmers to do, huh.
This is exactly what I think is happening with these exit polls. It is not fraud.
Lets say for instance that 2% of all Bush voters feel guilty about it and claim they voted for Kerry, while there are no other lies on the exit polls.
In Texas with (I don't know, lets guess) 75% for Bush, the exit polls would then say 73.5% for Bush. Everybody looks at this and says it is well within any error. In California (again with a guess number) 75% for Kerry, the exit polls say 75.5% for Kerry, again well within any error. But in Ohio, with it 50.5% for Bush, the exit polls say it is 49.5% for Bush and 50.5 for Kerry, and the exit polls now disagree!
Even though the error is consistent all across the country, nobody will think there is anything wrong except in Ohio. It seems amplified in the close states, making it look like there is some kind of planned fraud in the close states.
If the "or other" clause was included, somebody could make copy and say that their changes are GPL3 only. If their changes were popular enough (ie if they were from the core developers) they would effectively change Linux to GPL3, since the only GPL2 code would be useless outdated stuff that does not interoperate.
In fact this is exactly the same as why BSD code can be "stolen" by a company, despite claims to the contrary by many posters here. Anybody who thinks the GPL3 is dangerous should have to agree that BSD is even more dangerous.
It certainly is VM. Memory mapping of files is pretty much part of the VM system. Predictive loading of memory mapped files thus would also be part of the VM system, in that it would not work at all if the VM system did not somehow know it had been done. Since it is likely based on what things were memory mapped before, it is pretty obvious that the VM system would also be in charge of remembering this information and doing it a second time.
Not really sure what you are getting at and why you seem so hurt that somebody says "its part of the VM system".
Yes, that makes sense. Certainly there are reasons for a company to not want to use the GPL. However the reason is NOT "we might accidentally GPL our code". The reason is "we can't redistribute the program unless we GPL the code or we rewrite it to not use the GPL code". There is a HUGE difference, in that the first implies that just using GPL code could be a dangerous and once you do it there is no way to avoid the danger. In fact it is trivial to avoid the danger. This difference in perception is why I call this FUD.
No, I really meant what I said.
If they used GPL code in project A, and then distributed project A to external customers wihout the source code, then they are guilty of copyright violation. It does not mean they have to distribute the code for A.
It is true that it is likely the copyright owner will agree to not sue if they start distributing the code for A. However nothing in the law says this. Deciding on their own to distribute the code does not get them out of it (otherwise it would be a huge GPL loophole, as a company could wait a few years and then distribute the code when it is no longer worth anything, and suddenly all their previous violations are cancelled). And the courts cannot force them to do it, they can only force them to stop distribution and to pay monetary compensation to the copyright holder.
Yes I agree. Sorry if my post made it look like DRM is actually possible.
I do believe that a true open source DRM would result in a design such that it is extremely hard to determine the key, for instance where it is written such that the key itself is never stored in the resulting program, and other hacks. But it will never be impossible, for the simple reasons you state.
No your company is a bunch of idiots who have bought the FUD from Microsoft.
they don't want to inadvertantly make all of thier work go under the GPL if it's distributed
This is WRONG. It DOES NOT HAPPEN, it is a lie from Microsoft and one of the biggest and hardest to eradicate.
If you distribute without obeying the GPL, you are VIOLATING COPYRIGHT. You are then required to cease distribution, and you may be liable for monetary damages. You are NOT "forced to give away your source code". In fact even if you put every piece of source code you have in the public domain, it does not mean you did not violate copyright, so you are still liable. So not only are you not forced to give away the code, doing so does not get you out of anything, so this lie is doubly untrue!
Think hard. If the NY Tiimes was accused of plagarism, would this cause all their articles to now be the property of the person who wrote the one copied article?
Way to go with the selective quoting. Bravo.
You deleted the very next paragraph which describes the licenses rules for this "not precisely defined by law" statement:
If such an object file uses only numerical parameters, data structure layouts and accessors, and small macros and small inline functions (ten lines or less in length), then the use of the object file is unrestricted, regardless of whether it is legally a derivative work. (Executables containing this object code plus portions of the Library will still fall under Section 6.)
And you deleted the very next paragraph, which says that all this is only describing things where the LGPL is totally irrelevant because it is not a copyright violation. Otherwise the LGPL is explicitly granting you rights to violate the copyright:
Otherwise, if the work is a derivative of the Library, you may distribute the object code for the work under the terms of Section 6. Any executables containing that work also fall under Section 6, whether or not they are linked directly with the Library itself.
You also deleted all of "section 6" which describes how you can distribute without source code.
Now personally I think this section has some nasty problems in that RMS claims the end user should be able to change the library and relink, which is seriously out of touch with how modern software works, it would only be possible by distributing (perhaps obfuscated) source code to your program, making the difference between the LGPL and GPL nearly meaningless. I add an exception to my LGPL code to explicitly state that static linking, or any other use of the code that does not modify the code itself, it allowed. Adding such an exception is very common and I wish there was an official version of this, or the LGPL was fixed in this way.
It is quite possible to provide the source code so that anybody can make a music player that uses "Open DRM". The only problem is that the resulting music player won't play any commercial music. The secret is kept in a key which is different than the test ones provided with the source code. In fact this would allow the engineers to work on debugging the machine and adding features while only working with a set of "test" music thus making all the info they have worthless for breaking the DRM.
If the code is open-source, I'm quite certain that it would be analyzed so completely by thousands of people trying to prove they are smart by fixing it, that it would be quite impossible to get a key out of a working version, a level of security that the RIAA and MS could only dream of.
Don't be a moron. Dell has no responsibility to support software it does not sell with the machine. They are selling RedHat version something-or-other. This does not mean they have to support anything else that happens to have the letters L,I,N,U, and X in it's name.