A combination of GPL'ed and your code is a derived work because (duh) it includes GPL'ed code. Since your code is unusable without GPL'ed one, even if you distribute it separately it's still a part of derived work. You can use it internally but you can't distribute it -- a part of derived work is still a derived work.
If you modify your code so it does not use anything GPL'ed (by means other than adding a mangled version of GPL'ed code, of course), it becomes usable on its own, therefore you can do with it whatever you please as long as you don't violate licenses of other pieces that you use in it. No one would care if in the process of development you made intermediate versions that would be illegal to distribute as long as you didn't actually distribute them.
IANAL, however this is really, really simple -- this is why people told you to just fix your product. GPL isn't designed to prevent all possible kinds of freeloading by commercial entities or to exterminate non-free code, it specifically prevents one thing -- converting free code into non-free one. If your code never was free to begin with, it does not hurt you.
Developing code that is supposed to be redistributed with a GPL-incompatible license around GPL'ed libraries in a hope that eventually they can be replaced by something else is stupid because such a replacement usually involves a switch to a different interface provided by a different product, or reimplementation by people who may have trouble inventing a code that is different from GPL'ed source that they have seen. However this is not something that one can do accidentally. It's much more likely that someone "accidentally" will include restrictively licensed sample code from a book, or perform a modification/redistribution of some product commercially distributed in a source form but with closed license -- in those situations license terms can be unknown or misunderstood by the developer.
And you must admit that it's possible things could get uglier. For example, an IP holding company could buy the copyright to the code and start suing.
Yeah! Someone will buy a BSD-licensed piece of code just to be able to file nuisance lawsuits that will at most award him a fraction of his legal costs and your announcement that you credit him. He can't retroactively add stuff to the license, so all he can do is to pester people who managed to run afoul of the attribution clause, and only if it wasn't already revoked like it happened with most of BSD-licensed software anyway. If he indeed has nothing better to do, it would be easier to just buy a building with your office to evict your company.
This is one of the "BUT GIMP DOES NOT WORK WITH CMYK!!!!" (actually it does) arguments that Windows fanboys repeat constantly. You are slightly better than the rest of them because you have bothered to find out completely useless for "business user" scenarios where OO is slightly slower or uses different (though better justified) syntax. Congratulations. Now shut up and open a bunch of.xls files someone emailed to you -- there is a possibility that they aren't full of viruses.
The only situation when Microsoft Office documents (or ODF documents if it matters) should ever leave the computer where they are created is when someone else is supposed to modify them. This should never happen when a document crosses boundaries of organization.
Otherwise dollar is 500 times cheaper, because the volume of production is physically limited by two people hunting goats on Tobago, the only situation when this example can be possibly relevant.
But the economics of repair in space are just not worth it, unless the repair (or refuel-don't count fuel mass,it's not wasted mass, if it can be done) satellite mass is a small fraction of the destination satellite.
This is very important. Satellite design and launch are costly -- design takes time and launch requires large amount of fuel and other materials. Building the satellite is nothing compared to those costs. So if the alternatives are to replace the satellite with its exact copy plus/minus some error fixed, or to launch a repair satellite, then assuming that repair satellite design is also a fixed cost (one repair satellite can fix various satellites), the cost to build can be seen as negligible, and only the difference in launch cost should be important.
On one end of the scale are small, single-function satellites -- communications relays, GPS/navigation, weather. To repair them you need a satellite much heavier than themselves. You just launch an updated satellite, and if it's in GSO, it's a good idea to deorbit the old one to avoid crowding.
On the other end are massive space stations -- even if there were completely automated ones, used as platforms to carry multiple devices, "park" other satellites, etc., it's obvious that failure of, say, a communications relay would warrant a replacement, especially if that heavy station is regularly refueled anyway.
If you get even larger, station can simply carry its own repair robots capable of fixing anything that is attached to it. At a GSO it probably will make more sense to keep a station with hundreds of "attachments" and a bunch of robots than to send people or a special repair satellite there. Launching people, life support and supplies for them every few months to GSO would require huge amount of fuel, and keeping living space for them would increase the amount of fuel for the initial launch, however refueling and occasionally replacing the robots would be much cheaper than maintaining hundreds of individual satellites that provide the same functionality.
Repair satellite that is sitting in orbit and maneuvers between other satellites still has to burn a large amount of fuel to change its orbit, so unless broken satellites are in a "lucky" location, it's still too much fuel compared to replacement.
It also bothers me that EVERY RECENT LARGE-SCALE DECISION IN US SPACE PROGRAM IS THE MOST OBVIOUS DECISION MADE BY A LAYMAN WHO DOESN'T UNDERSTAND THE PROBLEM. Send people to orbit? Build an airplane with boosters and a large fuel tank! Make space program more useful for military? Put a missile launcher there! So it will shoot missiles DOWN, man! Establish colonies on Moon and Mars? Ignore the details, but let politicians declare that it will be done somehow! You need a heavy launcher for that? Oh, it's too expensive, we thought, you can reach Mars on a Shuttle! Expensive satellites only live for 20 years? Make a repair-bot that is even more expensive to launch, and its overcomplicated mechanism will likely break in a year!
No, he would feather his friends' nests and concentrate on more important things -- in case of Bill Gates it's convincing himself that he is not intellectually inferior to people who developed better software, something that seems to be his elusive lifelong goal.
Solution: Make sure that housing is no longer overpriced, so it's worth $20-30K, that would be a fair price for construction of most modern houses, plus worthless land that they are built on. Of course, that will also mean that you won't be able to take hundreds of thousands in equity loans just because you own a dry wall shack in the middle of nowhere, what causes massive inflation for the rest of us.
Long division IS a primitive set of instructions. This system still wouldn't work with it because at the age when kids study it, the problem is more likely inability to understand and follow clearly explained operations, boredom, distractions or sloppy writing -- something that would be very difficult to correctly recognize without a human teacher.
When a real knowledge is taught, the only way to make sure that student understands it is to make him write his own explanation, something that no program in a foreseeable future is going to understand, leave alone find holes in.
And yet in USSR, where state-sponsored education was among few properly implemented things, I have got a far better education than most of American students -- in both private and public schools.
We definitely didn't have social studies (aka mind-numbing propaganda of government's ideology in both countries) at that age, though it was an unavoidable part of "modern history" -- thankfully introduced much later, after students developed sufficient critical thinking skills.
Spelling was a part of Russian, however it was more a study of numerous rules of the language, with memorization of relatively few exceptions. Russian language is probably easier to study in a more systematic manner because words have many easy to derive forms, and usually proper spelling can be derived from a set of rules -- English seems to be more random, and it provides less possibilities to verify the spelling through comparison of different form of the word.
Some kind of "penmanship" was a part of the first-year study of letters, but it was expected that after that students are supposed to be able to write everything in cursive. Except foreign language classes (in my case English, introduced at the 4th year, now 5th) where cursive wasn't used at all.
Algebra and Physics started at 6th year (now 7th), Geometry, I think, two years before that. Chemistry and parts of Biology are tied together with Physics, so material taught in one course was used in subsequent study of other courses -- knowledge about atoms and molecules, physical properties of various substances from Physics were used in Chemistry, and knowledge of inorganic and some organic compounds used in later parts of Biology. In its turn Physics course relied on Math (Algebra/Geometry/Trigonometry/beginning of Calculus), but connection was mostly noticeable in 8th-10th (now 9th-11th) years when Physics went through a more math-heavy study of mechanics/heat/electricity mentioned earlier.
History started from 4th (now 5th) year as Ancient history, what was more or less free of ideological idiocy, and along with actual history mentioned ancient art and mythology. Later it switched to Middle Ages, Modern history, USSR history (as Russia and then USSR itself) and at the very end 20th century in general. Obviously anything about late 19th and the whole 20th century was heavily peppered with Communist ideology, however I think, it was taught at a much more more mature age than Capitalist indoctrination in US -- I think, I was at least 15 years old then.
Over the whole course of study I have never seen a multiple-choice test -- students had to present the whole procedure they used to derive the answer, teachers graded the whole work and discussed errors made by students. On the other hand, long essays were limited to Russian and Belarusian (I lived in Belarus) language and literature classes.
In the last two years the school where I studied had expanded Physics course, and additional Electronics course, however that was uncommon. Computer courses were introduced when I was in the 9th year of study, however since then they were moved to something very early -- I don't know how the computer-related curriculum looks now.
When I was eight years old, I was in the second year of secondary school in Gomel, Belarus, then-USSR. Typical school day WAS five hours long. Almost every day included Math (arithmetics), Russian, Russian Literature, fundamentals of science (not differentiated into Physics, Chemistry and Biology until the sixth year) and some "less rigorous" class like shop, physical education, music or drawing. Later one year was added before this, so it's third year now, however I am sure, the content of curriculum is the same. None of those things are optional.
This model only works for people studying simple operations (assembly line workers, customer support monkeys) or memorizing endless lists of facts, rules and instructions (lawyers). It's a great way to make people even dumber than they are now.
Teaching of math, physics, chemistry, biology, and all kinds of engineering should give students deep understanding of the subject, something that can't be evaluated with multiple-choice questions, or adjusted with some Civilization-like scenarios.
...and force all public schools to teach it. Without a curriculum designed by people who understand psychology of learning, can analyze experience of other countries, have resources to keep that research running for years, and can mandate the use of textbooks based on their recommendations (surprisingly, everyone but US seems to have that) curriculum will be up to a bunch of stupid rednecks in school boards and huge number of overworked teachers.
"State rights" are for things like slavery, gay marriage bans, and who-gets-earlier-primary drama, you morans.
Of course, it's easier to proselytize when you can say that you weren't Christian originally, however it isn't any different from every pro-Windows troll (a more common occurrence on Slashdot than born-again Christians) claiming great familiarity with other systems.
Even if they use flash-based iPod as opposed to hard drive-based one, it's still a very bad choice. You can have two flash cards with any interface recording flight data in parallel (pretending to be RAID1), in a continuous stream treating both cards as ring buffers, so each bit is overwritten once per cycle. Even with $15-20 modern flash cards you can get 1G before old data will be overwritten, and 100T (assuming that the cards are never changed) before 1/2 probability of an error. If we assume that black box records voice at 8Kbps and another 8Kbps records the flight data, that's more than 135 hours. If the voice is recorded at 64Kbps, you get 30 hours from a pair of cards. That's at most $40 of storage -- it can be expanded if necessary by adding more cards.
What is the point of adding some expensive, unreliable device that contains completely useless audio playback and input circuitry?
When harmless applications trigger the same "security" popups as administrative applications, nothing can be safe. ZoneAlarm has to save the list of "safe" applications because its prompts are nearly useless in the first place, and would be completely useless if users didn't "accidentally" install programs that should be blacklisted.
If an application needs elevated access that can directly affect the system, it should be requested from the user before application starts, but the point is, there are very few applications that should be allowed to have it in the first place, so those prompts are supposed to be very, very rare.
As I have explained, the problem is, user can't make an informed decision when in the middle of running a program he gets this question. iTunes accesses the Internet because iTunes is supposed to -- so if someone really thinks that there should be a "firewall" that keeps some applications from accessing the Internet, LAN or whatever else, the decision should be made even before the application is installed. Worse yet, some applications are not supposed to be interactive in the first place, yet they have to talk to the network, so having this dialog waiting for weeks when some application is trying to download its first update is something opposite to good security policy.
With administrative applications, of course, things are even easier -- there are few of them and they are easy to identify.
Following the example of two of the most annoying programs ever, ZoneAlarm and Norton Firewall, Microsoft implements a feature that requests a permission to do something from the person least likely able to make an informed choice, and absolutely not interested in knowing about it -- current desktop user. However in ZoneAlarm the reason for this is psychological -- if ZoneAlarm didn't constantly remind user that something is threatening his precious computer, user wouldn't know if ZoneAlarm does anything useful at all. In Vista it's pointless because it's not like user has a choice of buying or not buying some feature with it.
There are few specific APPLICATIONS, explicitly called by the user, that may have to run with elevated privileges, and beyond them there is nothing that is supposed to access system settings, write configuration files or executables. If anything other than those few select applications try to do that, user shouldn't be asked -- the action should be denied, just like it always was in Unix and occasionally even in Windows. If someone has to edit any system files, he knows that he has to run editor as administrator -- and if he doesn't, he has no reason to manually edit them in the first place. If someone runs installer, installer always has to run as administrator.
The reason why Gnome and KDE desktops have password dialogs is not to ask user if he does or doesn't want to do something privileged -- of course, he does if he just started some administrative application. It's to ask him for a password that malicious application or user with no sudo access can't enter by themselves, and to give him the application's name so he can be sure that the application that will run is the same application that he just asked for. The dialog can just as well be a captcha for users that can't remember their own passwords -- the point is to confirm that a program is started by a real human user in front of the keyboard. A piece of malware can run gksudo, and user will see the dialog with a program that he didn't run -- it's assumed that he will cancel it if he doesn't recognize the name. But this is actually a suboptimal use of sudo, a limitation of typical sudoers file configuration. A much better idea will be to supply sudoers file with all possible applications and arguments that may be used in this manner -- then anything else will be simply denied without any user's interaction, or user will be just notified that something tried to run gksudo with invalid arguments.
While the decision that administrative application may still run at reduced privileges unless it does something that requires true administrative access is a good idea, switching between those modes is not something that should be asked from user -- it should be asked at the very beginning when application starts, and should be done only for administrative applications.
Because science does not operate with beliefs, and should not conform to them. It would be a shitty kind of science if it didn't occasionally discover that some widely accepted beliefs were wrong.
If Ballmer will identify the "IP" that he claims, Linux violates as particular patents, and will have to waste enough time to match them to particular code, then it should be possible to invalidate those particular patents, and demonstrate that Microsoft can't cause any damage. The only real reason for this is that we know that almost all Microsoft patents can be invalidated, it just takes too much resources to attack all of them. But if Microsoft is forced to waste time on finding code that "violates" their patents, or be accused of fraudulent threats (acting in bad faith is a pretty shitty thing to have when you sue someone for parent violation), then it becomes feasible to attack whatever Microsoft claims to support them, and spend the amount of time and money that companies that depend on Linux can afford.
So whatever Microsoft does, as long as it doesn't have "real" patents that can stand a challenge _and_ remain applicable to Linux code regardless of possible changes, Ballmer can be demonstrated to be either wrong or malicious, and in either case harmless as far as litigation is concerned.
It can be also useful for the next antitrust lawsuit.
Name a single thing essential to good web site design that is impossible, or significantly more difficult to do in Gimp, but can be done in Photoshop.
I am sure, you can't because you know absolutely nothing about Gimp in the first place, and just spouting second-hand bullshit to make excuses for using your pirated copy of Photoshop.
Slashdot Mirror
A combination of GPL'ed and your code is a derived work because (duh) it includes GPL'ed code. Since your code is unusable without GPL'ed one, even if you distribute it separately it's still a part of derived work. You can use it internally but you can't distribute it -- a part of derived work is still a derived work.
If you modify your code so it does not use anything GPL'ed (by means other than adding a mangled version of GPL'ed code, of course), it becomes usable on its own, therefore you can do with it whatever you please as long as you don't violate licenses of other pieces that you use in it. No one would care if in the process of development you made intermediate versions that would be illegal to distribute as long as you didn't actually distribute them.
IANAL, however this is really, really simple -- this is why people told you to just fix your product. GPL isn't designed to prevent all possible kinds of freeloading by commercial entities or to exterminate non-free code, it specifically prevents one thing -- converting free code into non-free one. If your code never was free to begin with, it does not hurt you.
Developing code that is supposed to be redistributed with a GPL-incompatible license around GPL'ed libraries in a hope that eventually they can be replaced by something else is stupid because such a replacement usually involves a switch to a different interface provided by a different product, or reimplementation by people who may have trouble inventing a code that is different from GPL'ed source that they have seen. However this is not something that one can do accidentally. It's much more likely that someone "accidentally" will include restrictively licensed sample code from a book, or perform a modification/redistribution of some product commercially distributed in a source form but with closed license -- in those situations license terms can be unknown or misunderstood by the developer.
Name three GPL'ed libraries that are not readline. Explain why at least one of them is in any way relevant to your work.
Yeah! Someone will buy a BSD-licensed piece of code just to be able to file nuisance lawsuits that will at most award him a fraction of his legal costs and your announcement that you credit him. He can't retroactively add stuff to the license, so all he can do is to pester people who managed to run afoul of the attribution clause, and only if it wasn't already revoked like it happened with most of BSD-licensed software anyway. If he indeed has nothing better to do, it would be easier to just buy a building with your office to evict your company.
Next Microsoft shill, please.
This is one of the "BUT GIMP DOES NOT WORK WITH CMYK!!!!" (actually it does) arguments that Windows fanboys repeat constantly. You are slightly better than the rest of them because you have bothered to find out completely useless for "business user" scenarios where OO is slightly slower or uses different (though better justified) syntax. Congratulations. Now shut up and open a bunch of .xls files someone emailed to you -- there is a possibility that they aren't full of viruses.
The only situation when Microsoft Office documents (or ODF documents if it matters) should ever leave the computer where they are created is when someone else is supposed to modify them. This should never happen when a document crosses boundaries of organization.
First.
Otherwise dollar is 500 times cheaper, because the volume of production is physically limited by two people hunting goats on Tobago, the only situation when this example can be possibly relevant.
This is very important. Satellite design and launch are costly -- design takes time and launch requires large amount of fuel and other materials. Building the satellite is nothing compared to those costs. So if the alternatives are to replace the satellite with its exact copy plus/minus some error fixed, or to launch a repair satellite, then assuming that repair satellite design is also a fixed cost (one repair satellite can fix various satellites), the cost to build can be seen as negligible, and only the difference in launch cost should be important.
On one end of the scale are small, single-function satellites -- communications relays, GPS/navigation, weather. To repair them you need a satellite much heavier than themselves. You just launch an updated satellite, and if it's in GSO, it's a good idea to deorbit the old one to avoid crowding.
On the other end are massive space stations -- even if there were completely automated ones, used as platforms to carry multiple devices, "park" other satellites, etc., it's obvious that failure of, say, a communications relay would warrant a replacement, especially if that heavy station is regularly refueled anyway.
If you get even larger, station can simply carry its own repair robots capable of fixing anything that is attached to it. At a GSO it probably will make more sense to keep a station with hundreds of "attachments" and a bunch of robots than to send people or a special repair satellite there. Launching people, life support and supplies for them every few months to GSO would require huge amount of fuel, and keeping living space for them would increase the amount of fuel for the initial launch, however refueling and occasionally replacing the robots would be much cheaper than maintaining hundreds of individual satellites that provide the same functionality.
Repair satellite that is sitting in orbit and maneuvers between other satellites still has to burn a large amount of fuel to change its orbit, so unless broken satellites are in a "lucky" location, it's still too much fuel compared to replacement.
It also bothers me that EVERY RECENT LARGE-SCALE DECISION IN US SPACE PROGRAM IS THE MOST OBVIOUS DECISION MADE BY A LAYMAN WHO DOESN'T UNDERSTAND THE PROBLEM. Send people to orbit? Build an airplane with boosters and a large fuel tank! Make space program more useful for military? Put a missile launcher there! So it will shoot missiles DOWN, man! Establish colonies on Moon and Mars? Ignore the details, but let politicians declare that it will be done somehow! You need a heavy launcher for that? Oh, it's too expensive, we thought, you can reach Mars on a Shuttle! Expensive satellites only live for 20 years? Make a repair-bot that is even more expensive to launch, and its overcomplicated mechanism will likely break in a year!
No, he would feather his friends' nests and concentrate on more important things -- in case of Bill Gates it's convincing himself that he is not intellectually inferior to people who developed better software, something that seems to be his elusive lifelong goal.
Solution: Make sure that housing is no longer overpriced, so it's worth $20-30K, that would be a fair price for construction of most modern houses, plus worthless land that they are built on. Of course, that will also mean that you won't be able to take hundreds of thousands in equity loans just because you own a dry wall shack in the middle of nowhere, what causes massive inflation for the rest of us.
Can they pass a bill to bribe themselves to do something that makes sense?
Long division IS a primitive set of instructions. This system still wouldn't work with it because at the age when kids study it, the problem is more likely inability to understand and follow clearly explained operations, boredom, distractions or sloppy writing -- something that would be very difficult to correctly recognize without a human teacher.
When a real knowledge is taught, the only way to make sure that student understands it is to make him write his own explanation, something that no program in a foreseeable future is going to understand, leave alone find holes in.
And yet in USSR, where state-sponsored education was among few properly implemented things, I have got a far better education than most of American students -- in both private and public schools.
We definitely didn't have social studies (aka mind-numbing propaganda of government's ideology in both countries) at that age, though it was an unavoidable part of "modern history" -- thankfully introduced much later, after students developed sufficient critical thinking skills.
Spelling was a part of Russian, however it was more a study of numerous rules of the language, with memorization of relatively few exceptions. Russian language is probably easier to study in a more systematic manner because words have many easy to derive forms, and usually proper spelling can be derived from a set of rules -- English seems to be more random, and it provides less possibilities to verify the spelling through comparison of different form of the word.
Some kind of "penmanship" was a part of the first-year study of letters, but it was expected that after that students are supposed to be able to write everything in cursive. Except foreign language classes (in my case English, introduced at the 4th year, now 5th) where cursive wasn't used at all.
Algebra and Physics started at 6th year (now 7th), Geometry, I think, two years before that. Chemistry and parts of Biology are tied together with Physics, so material taught in one course was used in subsequent study of other courses -- knowledge about atoms and molecules, physical properties of various substances from Physics were used in Chemistry, and knowledge of inorganic and some organic compounds used in later parts of Biology. In its turn Physics course relied on Math (Algebra/Geometry/Trigonometry/beginning of Calculus), but connection was mostly noticeable in 8th-10th (now 9th-11th) years when Physics went through a more math-heavy study of mechanics/heat/electricity mentioned earlier.
History started from 4th (now 5th) year as Ancient history, what was more or less free of ideological idiocy, and along with actual history mentioned ancient art and mythology. Later it switched to Middle Ages, Modern history, USSR history (as Russia and then USSR itself) and at the very end 20th century in general. Obviously anything about late 19th and the whole 20th century was heavily peppered with Communist ideology, however I think, it was taught at a much more more mature age than Capitalist indoctrination in US -- I think, I was at least 15 years old then.
Over the whole course of study I have never seen a multiple-choice test -- students had to present the whole procedure they used to derive the answer, teachers graded the whole work and discussed errors made by students. On the other hand, long essays were limited to Russian and Belarusian (I lived in Belarus) language and literature classes.
In the last two years the school where I studied had expanded Physics course, and additional Electronics course, however that was uncommon. Computer courses were introduced when I was in the 9th year of study, however since then they were moved to something very early -- I don't know how the computer-related curriculum looks now.
When I was eight years old, I was in the second year of secondary school in Gomel, Belarus, then-USSR. Typical school day WAS five hours long. Almost every day included Math (arithmetics), Russian, Russian Literature, fundamentals of science (not differentiated into Physics, Chemistry and Biology until the sixth year) and some "less rigorous" class like shop, physical education, music or drawing. Later one year was added before this, so it's third year now, however I am sure, the content of curriculum is the same. None of those things are optional.
This model only works for people studying simple operations (assembly line workers, customer support monkeys) or memorizing endless lists of facts, rules and instructions (lawyers). It's a great way to make people even dumber than they are now.
Teaching of math, physics, chemistry, biology, and all kinds of engineering should give students deep understanding of the subject, something that can't be evaluated with multiple-choice questions, or adjusted with some Civilization-like scenarios.
...and force all public schools to teach it. Without a curriculum designed by people who understand psychology of learning, can analyze experience of other countries, have resources to keep that research running for years, and can mandate the use of textbooks based on their recommendations (surprisingly, everyone but US seems to have that) curriculum will be up to a bunch of stupid rednecks in school boards and huge number of overworked teachers.
"State rights" are for things like slavery, gay marriage bans, and who-gets-earlier-primary drama, you morans.
I call bullshit on that.
Of course, it's easier to proselytize when you can say that you weren't Christian originally, however it isn't any different from every pro-Windows troll (a more common occurrence on Slashdot than born-again Christians) claiming great familiarity with other systems.
What the world is coming to?
Even if they use flash-based iPod as opposed to hard drive-based one, it's still a very bad choice. You can have two flash cards with any interface recording flight data in parallel (pretending to be RAID1), in a continuous stream treating both cards as ring buffers, so each bit is overwritten once per cycle. Even with $15-20 modern flash cards you can get 1G before old data will be overwritten, and 100T (assuming that the cards are never changed) before 1/2 probability of an error. If we assume that black box records voice at 8Kbps and another 8Kbps records the flight data, that's more than 135 hours. If the voice is recorded at 64Kbps, you get 30 hours from a pair of cards. That's at most $40 of storage -- it can be expanded if necessary by adding more cards.
What is the point of adding some expensive, unreliable device that contains completely useless audio playback and input circuitry?
When harmless applications trigger the same "security" popups as administrative applications, nothing can be safe. ZoneAlarm has to save the list of "safe" applications because its prompts are nearly useless in the first place, and would be completely useless if users didn't "accidentally" install programs that should be blacklisted.
If an application needs elevated access that can directly affect the system, it should be requested from the user before application starts, but the point is, there are very few applications that should be allowed to have it in the first place, so those prompts are supposed to be very, very rare.
As I have explained, the problem is, user can't make an informed decision when in the middle of running a program he gets this question. iTunes accesses the Internet because iTunes is supposed to -- so if someone really thinks that there should be a "firewall" that keeps some applications from accessing the Internet, LAN or whatever else, the decision should be made even before the application is installed. Worse yet, some applications are not supposed to be interactive in the first place, yet they have to talk to the network, so having this dialog waiting for weeks when some application is trying to download its first update is something opposite to good security policy.
With administrative applications, of course, things are even easier -- there are few of them and they are easy to identify.
Following the example of two of the most annoying programs ever, ZoneAlarm and Norton Firewall, Microsoft implements a feature that requests a permission to do something from the person least likely able to make an informed choice, and absolutely not interested in knowing about it -- current desktop user. However in ZoneAlarm the reason for this is psychological -- if ZoneAlarm didn't constantly remind user that something is threatening his precious computer, user wouldn't know if ZoneAlarm does anything useful at all. In Vista it's pointless because it's not like user has a choice of buying or not buying some feature with it.
There are few specific APPLICATIONS, explicitly called by the user, that may have to run with elevated privileges, and beyond them there is nothing that is supposed to access system settings, write configuration files or executables. If anything other than those few select applications try to do that, user shouldn't be asked -- the action should be denied, just like it always was in Unix and occasionally even in Windows. If someone has to edit any system files, he knows that he has to run editor as administrator -- and if he doesn't, he has no reason to manually edit them in the first place. If someone runs installer, installer always has to run as administrator.
The reason why Gnome and KDE desktops have password dialogs is not to ask user if he does or doesn't want to do something privileged -- of course, he does if he just started some administrative application. It's to ask him for a password that malicious application or user with no sudo access can't enter by themselves, and to give him the application's name so he can be sure that the application that will run is the same application that he just asked for. The dialog can just as well be a captcha for users that can't remember their own passwords -- the point is to confirm that a program is started by a real human user in front of the keyboard. A piece of malware can run gksudo, and user will see the dialog with a program that he didn't run -- it's assumed that he will cancel it if he doesn't recognize the name. But this is actually a suboptimal use of sudo, a limitation of typical sudoers file configuration. A much better idea will be to supply sudoers file with all possible applications and arguments that may be used in this manner -- then anything else will be simply denied without any user's interaction, or user will be just notified that something tried to run gksudo with invalid arguments.
While the decision that administrative application may still run at reduced privileges unless it does something that requires true administrative access is a good idea, switching between those modes is not something that should be asked from user -- it should be asked at the very beginning when application starts, and should be done only for administrative applications.
Because science does not operate with beliefs, and should not conform to them. It would be a shitty kind of science if it didn't occasionally discover that some widely accepted beliefs were wrong.
If Ballmer will identify the "IP" that he claims, Linux violates as particular patents, and will have to waste enough time to match them to particular code, then it should be possible to invalidate those particular patents, and demonstrate that Microsoft can't cause any damage. The only real reason for this is that we know that almost all Microsoft patents can be invalidated, it just takes too much resources to attack all of them. But if Microsoft is forced to waste time on finding code that "violates" their patents, or be accused of fraudulent threats (acting in bad faith is a pretty shitty thing to have when you sue someone for parent violation), then it becomes feasible to attack whatever Microsoft claims to support them, and spend the amount of time and money that companies that depend on Linux can afford.
So whatever Microsoft does, as long as it doesn't have "real" patents that can stand a challenge _and_ remain applicable to Linux code regardless of possible changes, Ballmer can be demonstrated to be either wrong or malicious, and in either case harmless as far as litigation is concerned.
It can be also useful for the next antitrust lawsuit.
Name a single thing essential to good web site design that is impossible, or significantly more difficult to do in Gimp, but can be done in Photoshop.
I am sure, you can't because you know absolutely nothing about Gimp in the first place, and just spouting second-hand bullshit to make excuses for using your pirated copy of Photoshop.