You seem to have missed the idea of dual licensing. The limitations of the GPL make other licenses valuable and thus the original author can sell these other licenses. Meanwhile the GPL release allows the author's solution to become popular, standardized, and expected by users, making the sale of it more valuable. You can be certain Qt would sell nothing if they did not also have the GPL version. In addition it appears GPL code is very useful for advertising your abilities as a programmer for getting jobs.
GPL is actually *better* for professionals to get paid than public domain, totally opposite of your argument. Of course the reason is not something Stallman wants, but it is true.
For small companies and people, the GPL is the only way they can advertise and get their ideas and standards used by others. It flattens the playing field so that it the design of computers is not 100% controlled by whoever has the brand name recognition and advertising budget. This is why Microsoft fears it, not because it thinks it will be forced to open-source their own stuff or that software will all become zero-cost.
Code that is Windows-only is not cross-platform, stupid.
What I want to see from Microsoft is modifications so it is easier to port code to Windows without rewriting it. This does not involve releasing any kind of source, what it requires is Posix/Linux compatibility, and not a huge deal of it. Just using forward slashes in filenames would be a huge help (yes they do accept them but still insist on copying/pasting backslashes between applications). Accept UTF-8 without the leading 3 bytes, and make all programs accept files without ^M characters before the ^J characters. Don't add underscores to the starts of function names with some bogus excuse about Posix chapters. Add strlcpy from BSD. Make "/A:/" be the same as "A:/" so rooted filenames can always start with a slash. Just show some sign that you are going to fix the piles and piles of crap that is the real reason everybody here hates them.
And if they want to release source code I want it to be something other than modifications to make a piece of OSS work on Windows, I want code that is cross platform, like actually interesting algorithims.
The voter gets the paper record, looks at it, then puts it into a locked ballot box. The ballots in that box are counted later for the official count, with discrepencies with any electronic count triggering a careful audit. The voter does not take anything with them out of the election area.
There have been proposals for systems by which a voter can take something with them that proves that their vote was counted exactly once but cannot show how they voted, but all of them are complex and impossible for average voters to understand.
That name was purposly chosen to be as confusing as possible with "Open Office", so that is not a good example.
But the real reason Microsoft does not use "open" in the names of things is because of Sun using it way too much. It's not because they hate open source.
Dont be such a fool. They *required* the "vendor specific fields" to be filled in with the "vendor specific data" or your Windows machine would not authenticate. This was done purposely so that you had to use a Windows server if you had Windows desktops. It was also pretty clear that any requirements for correct data being in those fields was against the spec.
The fact that a Unix desktop could log into a Windows server did not concern them, as they already controlled the desktop market, and were trying to use it to take over the server market. So don't give the lame excuse that you could authenticate a RedHat machine or whatever.
If you believe otherwise then you are a totally blind idiot. This was incredibly blatent, even for Microsoft.
My immediate impression is this is somebody trying to blame M$ for their own failings.
How well does this work on Linux (with the various filesystems) and OS/X? Is Vista really doing something stupid, or is it being blamed for the same mistake as everybody else? What about XP?
Other thing is I remember the disk-thrashing bug in Linux Ubuntu. I have it and have to run a startup program to turn off the hard disk power savings to stop the head-park every half second. I did a lot of searching of the web, looking for an explanation of why XP works, and the only real experiments I found indicated that XP just kept reading the disk, so often that it *never* parked the heads. Thus Linux's reduced (but non-zero) use of the disk made things worse. All other tests seemed to indicate they left the power saving settings the same and I never saw any other explanation. This does sound like it might be related to the SSD problems, but those tests were certainly with XP and not Vista-only. Anybody know anything about this?
Or add a moderation that means "I disagree with this but it is interesting". I use "interesting" for that but I'm afraid far too many moderators use "overrated" for anything they disagree with.
Can't overrated/underrated be replaced by a scheme where the moderator says "I think it should have this number". Their vote moves it one in the direction of that number, if it is already there or past it it does not change. If several vote at the same time for the same number it will still only move to that number.
I would also like to see as a new set of modification types:
"I agree and this is a good post" (mods up)
"I disagree but this is a good post" (mods up)
"I agree but this is a stupid post" (mods down)
"I disagree and this is a stupid post" (mods down)
"Flame/troll by somebody I disagree with" (mods down, this indicates a fake post by an opponent to the moderator, but pretending to support their opinion but with incorrect or easily refuted strawman arguments)
"Flame/troll by somebody I agree with" (mods down, this indicates a fake post by somebody pretending to argue against the moderator's position, but with incorrect or easily refuted strawman arguemnts).
"Funny" (which really means off-topic but the moderator thinks people should see it anyway)
It's not like the drive letters and backslashes would stop working.
The changes are pretty simple:
1. Make/C:/ mean the same as C:/ and make readdir("/") return C: as one of the results. They should also make readdir("//") return the network neighborhood, but that is a minor extra detail, but then no api other than readdir() is needed to enumerate all the available files.
2. Make apis that return pathnames return them with forward slashes.
3. Ignore the text/binary switch when writing files and make them only write \n between lines.
Old programs would continue to work. If these actually broke programs they are pretty badly written.
Though I don't really agree with the grandparent, your counterexample is wrong. Toyota's plants in the US are producing cars that are then sold in the US, they are not being shipped back to Japan for sale there. If a US company opened a service center in India for selling to Indian customers I don't think anybody here would have any complaint.
I agree "copyright infingement" is the correct term. Or "violating copyright" or if you want a stronger-sounding term.
"Piracy" should only mean copyright infringement for financial gain (ie they sell the copies). I think the RIAA/etc are shooting themselves in the foot by trying to expand this more evil-sounding term to every copyright violator, instead of their plan of making everybody sound more evil they have instead defused the "pirate" term so that even profiteers sound less evil.
"Plagiarism" is when you violate copyright and also try to pretend the work is your own or otherwise hide or deny credit to the original creator. Usually not a problem for music/movies, but this does come up when people try to compare GPL violations to this.
They are both "wrong" and there are some serious attempts to fix this.
But when they don't conflict, they both make incredibly accurate predictions that have been tested over and over and over again. You just now used a computer to type something, it was designed and implemented by applying the very "wrong" theory of Quantum mechanics, it is proving itself to make useful predictions trillions of times a second right there on your desk! Yet it really is "wrong" because the theory does not match reality in some situations.
The very fact that they can be "wrong" is EXACTLY what makes them useful theories. This is exactly what "falsifiability" means. But you of course don't get it, do you. Very sad.
You are ignoring the fact that it is the *successful* adaptations that are used in the next "experiment". The failures are thrown away because they are wiped out by the more successful gene. Your math is based on the assumption that there are just a whole lot of parallel sets of random mutations.
I think you are looking it, like a lot of people, this way: if there is a population of P and G generations, and 50% chance of a mutation being benificial, then there are P*G/2 benificial mutations. This number can be quite big, but is obviosly small compared to how many genes there are.
What really you get is P^(G/2) benificial mutations, which is an astronomical number, greater than the number of genes in every single organism in total. Well not quite that number, but what you get is the best result of all these possibilities, because at each generation the P is mostly used by only the ones that had the benificial mutation last generation.
What's very sad now is people like you thinking the "majority" thought he was innocent.
I certainly did not think he was innocent. Take a look at my posting history, I'm pretty certain I said something about this. And I doubt I'm the only one.
I don't even think the majority of people who posted in response to these stories thought he was innocent. Go read them again.
Your "stracpy" is actually standardized, it is called strdup().
I'm not sure how useful the stracat one is, as most code I see strcat's in a loop, or has some if statements around some of the strings, or otherwise makes this difficult. What would have to be done instead is to make some sort of list which stores pointers to all the strings you want to concatenate. You then call a function at the end which measures all of them, and allocates and concatenates into it. I suspect most C programmers would write this themselves.
Also for standard C the variable argument list for stracat would have to have a null terminator.
One thing I would like to see in posix is a "is_string_constant(char*)" that returns a bool. It is true if it points to a string constant and false if it points to allocated memory or local memory on the stack. This would allow "efficient" C++ style strings that don't malloc copies of string constants, and would probably remove most of the objections of C programmers to using a string library.
This is an annoying incompatability with other compilers, but I don't think it has anything to do with the ability to recompile code from 32 to 64 and vice-versa. In fact either solution makes different software portable and it is hard to say which is better.
Microsoft's main incompatability was that they use a new keyword (something like "__int64") to mean the 64-bit quantity, when *EVERY OTHER COMPILER IN THE WORLD* used "long long". It does seem they did this on purpose to make it hard to write portable software as you cannot make a macro so that "long long" can turn into their need, so instead you have to rewrite all the source code to replace the "long long" with a macro. Still the two-word "long long" is a pretty stupid solution so you can also blame whoever thought of that originally (DEC?)
The reason to make "long" be 32 bits is to be compatible with a lot of software dating back to 16-bit ints where "long" was used to indicate 32 bit fields in data structures.
The reason to make "long" be 64 bits is to be compatible with a lot of software that also dates to the same period where "long" meant "big enough to hold a pointer, or the difference between pointers".
Personally I think the C++ standard should stop trying for some noble purity and add a bunch of keywords: "intN" where N is a number means an integer of N bits or bigger, "unsigned intN" means an unsigned one, and "ptrdiff_t" or something means the type when you subtract two pointers, and "unsigned ptrdiff_t" means "size_t".
You seem to think all the computers running Microsoft operating systems would instantaneously disappear if Microsoft disappeared? That every company selling software to work on those systems would suddenly stop being able to sell anything?
You have a seriously twisted view on how the world works.
Here is what would really happen if Microsoft went bankrupt:
1. Sales by companies that make software that works on Windows would *increase* as everybody would stop waiting for Microsoft to produce the product and would be willing to pay to keep their existing systems working.
2. Windows/Vista/etc would not magically stop working. People needing compatability would continue to install it for a long, long time.
3. An alternative, probably a clone written by some of those 80,000 former employees, would appear in *weeks*. It would likely use stolen IP but there would be nobody who would care. It might use a Linux or BSD kernel but there is no need to make any technical predictions. This company would get a massive influx of money as computer manufacturers would need them to supply their machines.
3. Whoever produced that clone would become filthy rich and would turn into an evil monopoly.
Slashdot Mirror
You seem to have missed the idea of dual licensing. The limitations of the GPL make other licenses valuable and thus the original author can sell these other licenses. Meanwhile the GPL release allows the author's solution to become popular, standardized, and expected by users, making the sale of it more valuable. You can be certain Qt would sell nothing if they did not also have the GPL version. In addition it appears GPL code is very useful for advertising your abilities as a programmer for getting jobs.
GPL is actually *better* for professionals to get paid than public domain, totally opposite of your argument. Of course the reason is not something Stallman wants, but it is true.
For small companies and people, the GPL is the only way they can advertise and get their ideas and standards used by others. It flattens the playing field so that it the design of computers is not 100% controlled by whoever has the brand name recognition and advertising budget. This is why Microsoft fears it, not because it thinks it will be forced to open-source their own stuff or that software will all become zero-cost.
Code that is Windows-only is not cross-platform, stupid.
What I want to see from Microsoft is modifications so it is easier to port code to Windows without rewriting it. This does not involve releasing any kind of source, what it requires is Posix/Linux compatibility, and not a huge deal of it. Just using forward slashes in filenames would be a huge help (yes they do accept them but still insist on copying/pasting backslashes between applications). Accept UTF-8 without the leading 3 bytes, and make all programs accept files without ^M characters before the ^J characters. Don't add underscores to the starts of function names with some bogus excuse about Posix chapters. Add strlcpy from BSD. Make "/A:/" be the same as "A:/" so rooted filenames can always start with a slash. Just show some sign that you are going to fix the piles and piles of crap that is the real reason everybody here hates them.
And if they want to release source code I want it to be something other than modifications to make a piece of OSS work on Windows, I want code that is cross platform, like actually interesting algorithims.
This is probably a troll, but:
The voter gets the paper record, looks at it, then puts it into a locked ballot box. The ballots in that box are counted later for the official count, with discrepencies with any electronic count triggering a careful audit. The voter does not take anything with them out of the election area.
There have been proposals for systems by which a voter can take something with them that proves that their vote was counted exactly once but cannot show how they voted, but all of them are complex and impossible for average voters to understand.
Because of projectors. There are lots of them that have only VGA plugs.
Also I think it is very rare for there to be a projector or display that has DVI but does not also have VGA.
The government is only *pretending* to be incompetent! It's all part of the plan, to try to discredit the conspiracy theories, can't you see?
That name was purposly chosen to be as confusing as possible with "Open Office", so that is not a good example.
But the real reason Microsoft does not use "open" in the names of things is because of Sun using it way too much. It's not because they hate open source.
Dont be such a fool. They *required* the "vendor specific fields" to be filled in with the "vendor specific data" or your Windows machine would not authenticate. This was done purposely so that you had to use a Windows server if you had Windows desktops. It was also pretty clear that any requirements for correct data being in those fields was against the spec.
The fact that a Unix desktop could log into a Windows server did not concern them, as they already controlled the desktop market, and were trying to use it to take over the server market. So don't give the lame excuse that you could authenticate a RedHat machine or whatever.
If you believe otherwise then you are a totally blind idiot. This was incredibly blatent, even for Microsoft.
My immediate impression is this is somebody trying to blame M$ for their own failings.
How well does this work on Linux (with the various filesystems) and OS/X? Is Vista really doing something stupid, or is it being blamed for the same mistake as everybody else? What about XP?
Other thing is I remember the disk-thrashing bug in Linux Ubuntu. I have it and have to run a startup program to turn off the hard disk power savings to stop the head-park every half second. I did a lot of searching of the web, looking for an explanation of why XP works, and the only real experiments I found indicated that XP just kept reading the disk, so often that it *never* parked the heads. Thus Linux's reduced (but non-zero) use of the disk made things worse. All other tests seemed to indicate they left the power saving settings the same and I never saw any other explanation. This does sound like it might be related to the SSD problems, but those tests were certainly with XP and not Vista-only. Anybody know anything about this?
Or add a moderation that means "I disagree with this but it is interesting". I use "interesting" for that but I'm afraid far too many moderators use "overrated" for anything they disagree with.
Can't overrated/underrated be replaced by a scheme where the moderator says "I think it should have this number". Their vote moves it one in the direction of that number, if it is already there or past it it does not change. If several vote at the same time for the same number it will still only move to that number.
I would also like to see as a new set of modification types:
"I agree and this is a good post" (mods up)
"I disagree but this is a good post" (mods up)
"I agree but this is a stupid post" (mods down)
"I disagree and this is a stupid post" (mods down)
"Flame/troll by somebody I disagree with" (mods down, this indicates a fake post by an opponent to the moderator, but pretending to support their opinion but with incorrect or easily refuted strawman arguments)
"Flame/troll by somebody I agree with" (mods down, this indicates a fake post by somebody pretending to argue against the moderator's position, but with incorrect or easily refuted strawman arguemnts).
"Funny" (which really means off-topic but the moderator thinks people should see it anyway)
"Off topic" (means off-topic but bad, mods down)
It's not like the drive letters and backslashes would stop working.
The changes are pretty simple:
1. Make /C:/ mean the same as C:/ and make readdir("/") return C: as one of the results. They should also make readdir("//") return the network neighborhood, but that is a minor extra detail, but then no api other than readdir() is needed to enumerate all the available files.
2. Make apis that return pathnames return them with forward slashes.
3. Ignore the text/binary switch when writing files and make them only write \n between lines.
Old programs would continue to work. If these actually broke programs they are pretty badly written.
I typed "OS/X Leapord box" into Google and the ad on the right shows $107.98, right ON the results page!
Clicking the Amazon ad link above that shows $109.99.
This took approximately 4 seconds. Does not look like you Googled very hard...
Please look up "sarcasm" in the dictionary next time.
Though I don't really agree with the grandparent, your counterexample is wrong. Toyota's plants in the US are producing cars that are then sold in the US, they are not being shipped back to Japan for sale there. If a US company opened a service center in India for selling to Indian customers I don't think anybody here would have any complaint.
The sudden appearance of an animal with many vastly different traits would disprove evolution.
I agree "copyright infingement" is the correct term. Or "violating copyright" or if you want a stronger-sounding term.
"Piracy" should only mean copyright infringement for financial gain (ie they sell the copies). I think the RIAA/etc are shooting themselves in the foot by trying to expand this more evil-sounding term to every copyright violator, instead of their plan of making everybody sound more evil they have instead defused the "pirate" term so that even profiteers sound less evil.
"Plagiarism" is when you violate copyright and also try to pretend the work is your own or otherwise hide or deny credit to the original creator. Usually not a problem for music/movies, but this does come up when people try to compare GPL violations to this.
No because rabbits inherited traits from other species. The appearance of them before the species with similar traits would disprove evolution.
They are both "wrong" and there are some serious attempts to fix this.
But when they don't conflict, they both make incredibly accurate predictions that have been tested over and over and over again. You just now used a computer to type something, it was designed and implemented by applying the very "wrong" theory of Quantum mechanics, it is proving itself to make useful predictions trillions of times a second right there on your desk! Yet it really is "wrong" because the theory does not match reality in some situations.
The very fact that they can be "wrong" is EXACTLY what makes them useful theories. This is exactly what "falsifiability" means. But you of course don't get it, do you. Very sad.
You are ignoring the fact that it is the *successful* adaptations that are used in the next "experiment". The failures are thrown away because they are wiped out by the more successful gene. Your math is based on the assumption that there are just a whole lot of parallel sets of random mutations.
I think you are looking it, like a lot of people, this way: if there is a population of P and G generations, and 50% chance of a mutation being benificial, then there are P*G/2 benificial mutations. This number can be quite big, but is obviosly small compared to how many genes there are.
What really you get is P^(G/2) benificial mutations, which is an astronomical number, greater than the number of genes in every single organism in total. Well not quite that number, but what you get is the best result of all these possibilities, because at each generation the P is mostly used by only the ones that had the benificial mutation last generation.
What's very sad now is people like you thinking the "majority" thought he was innocent.
I certainly did not think he was innocent. Take a look at my posting history, I'm pretty certain I said something about this. And I doubt I'm the only one.
I don't even think the majority of people who posted in response to these stories thought he was innocent. Go read them again.
Your "stracpy" is actually standardized, it is called strdup().
I'm not sure how useful the stracat one is, as most code I see strcat's in a loop, or has some if statements around some of the strings, or otherwise makes this difficult. What would have to be done instead is to make some sort of list which stores pointers to all the strings you want to concatenate. You then call a function at the end which measures all of them, and allocates and concatenates into it. I suspect most C programmers would write this themselves.
Also for standard C the variable argument list for stracat would have to have a null terminator.
One thing I would like to see in posix is a "is_string_constant(char*)" that returns a bool. It is true if it points to a string constant and false if it points to allocated memory or local memory on the stack. This would allow "efficient" C++ style strings that don't malloc copies of string constants, and would probably remove most of the objections of C programmers to using a string library.
This is an annoying incompatability with other compilers, but I don't think it has anything to do with the ability to recompile code from 32 to 64 and vice-versa. In fact either solution makes different software portable and it is hard to say which is better.
Microsoft's main incompatability was that they use a new keyword (something like "__int64") to mean the 64-bit quantity, when *EVERY OTHER COMPILER IN THE WORLD* used "long long". It does seem they did this on purpose to make it hard to write portable software as you cannot make a macro so that "long long" can turn into their need, so instead you have to rewrite all the source code to replace the "long long" with a macro. Still the two-word "long long" is a pretty stupid solution so you can also blame whoever thought of that originally (DEC?)
The reason to make "long" be 32 bits is to be compatible with a lot of software dating back to 16-bit ints where "long" was used to indicate 32 bit fields in data structures.
The reason to make "long" be 64 bits is to be compatible with a lot of software that also dates to the same period where "long" meant "big enough to hold a pointer, or the difference between pointers".
Personally I think the C++ standard should stop trying for some noble purity and add a bunch of keywords: "intN" where N is a number means an integer of N bits or bigger, "unsigned intN" means an unsigned one, and "ptrdiff_t" or something means the type when you subtract two pointers, and "unsigned ptrdiff_t" means "size_t".
Instant spelling error detection was a big deal in Word *BEFORE* Windows. I know as I was working at another company and we were competing with it.
Personally, I think there must be some recent real improvements to Word but Belial6's attempt to come up with one is a hilariously massive failure.
You seem to think all the computers running Microsoft operating systems would instantaneously disappear if Microsoft disappeared? That every company selling software to work on those systems would suddenly stop being able to sell anything?
You have a seriously twisted view on how the world works.
Here is what would really happen if Microsoft went bankrupt:
1. Sales by companies that make software that works on Windows would *increase* as everybody would stop waiting for Microsoft to produce the product and would be willing to pay to keep their existing systems working.
2. Windows/Vista/etc would not magically stop working. People needing compatability would continue to install it for a long, long time.
3. An alternative, probably a clone written by some of those 80,000 former employees, would appear in *weeks*. It would likely use stolen IP but there would be nobody who would care. It might use a Linux or BSD kernel but there is no need to make any technical predictions. This company would get a massive influx of money as computer manufacturers would need them to supply their machines.
3. Whoever produced that clone would become filthy rich and would turn into an evil monopoly.
As opposed to Linux, where I have to enter my password about 39,000 times a day?
Huh? Can you explain what you are talking about?