Which is exactly why I love the GPL. I know the code I will release might be reused by other people but they MUST release their modification of it.
So you like the GPL because you don't understand the GPL? 90% of all software developers are employed writing in-house software that is never distributed. If any of these use or modify your code, then they have no legal obligation to release their modifications.
They usually do. And they quite often then release it under a permissive (e.g. BSD) license and so other people help find bugs and improve it. Then the next company to come along has a choice of using the GPL or BSDL version. The GPL version may be more mature, but the BSDL one comes with less potential liability, so they use it and improve it. It's cheaper for them to send patches upstream than to maintain a private fork, so they do that. Then another company comes along, sees a fairly mature GPL'd project and a slightly less mature, but more rapidly improving, BSDL project. Guess which one they pick?
That's rather the point of TFA. Given the choice between using GPL'd code or writing it themselves, a lot of people will pick option 2.
Which BSD kernel? The XNU kernel that Apple uses? They release all of the sources. It's based on CMU Mach, 4BSD, and has more recently some code from the FreeBSD kernel.
Did Apple contribute HFS+ back to BSD?
You mean this code? It's APSL, and someone did port it to FreeBSD, but no one really cared so it bit rotted.
Offhand I can't think of any apple subsystem that made its way into BSD
How about the pthread_workqueue_*() family of functions? They are used by libdispatch to manage a pool of worker threads.
Mach-O loader? no
Why would any BSD want to switch from ELF to Mach-O? You probably could replace rtld with dyld, but what would be the point?
Only if there is a single owner. Most non-trivial projects have multiple contributors. I've released code under an MIT license and had some fairly significant improvements contributed back. I am free to use the resulting version of the project in any way I want. There are no restrictions on what I can do with it. If I'd GPL'd the original, then people would have given me GPL'd patches and I would only have been able to use the improved version in accordance with the GPL, although I could use the original version under any terms I wanted.
This was NeXT, and it is a case study in the failure of the GPL. Initially, NeXT released their Objective-C front end as a binary-only front end for GCC. The FSF threatened them with lawyers and forced them to release the source. There were just two problems:
They didn't release the Objective-C runtime, so the compiler was actually pretty useless by itself.
The code was shockingly bad.
Unfortunately, after trumpeting the 'success' of the GPL in forcing NeXT to release the code, they were forced to integrate it with GCC or lose face. They then hacked in some ugly changes to support their own runtime (changes which NeXT / Apple imported into their own branch). If they'd started from scratch and written their own Objective-C front end (which was only about 10KLoC back then), then they'd almost certainly have ended up with better code.
You spend a million hours creating a great program, call it Foo, and you release it under the BSD license. Someone like Microsoft comes along, takes Foo, improves it a little bit, calls it FooBar and distributes it as a binary without publishing any of the code
Your reasoning ignores the fact that off-the-shelf software is a tiny niche - it's responsible for employing about 10% of software developers. For most types of software, there is no Microsoft. There is no big company that will sell you a license to use an existing product. There are, however, lots of companies that will look at your business needs and provide you with some software that addresses them.
The GPL does nothing to protect you from this kind of competition. Let's say you release an improvement to the Linux kernel or a piece of GPL'd software that runs on Linux. When someone wants a Linux-based system, are they likely to go to you, or to a company like Red Hat or IBM? These companies will happily take your code, modify it, and charge their customers for the modifications (actually, charge the customer for the finished product, but typically less than they'd charge for writing it from scratch). The customer then receives the code under the GPL, but has no obligation to share the modifications with you, or with anyone else - indeed, if they give the provide a competitive advantage then they may persuade Red Hat or IBM to keep the modifications private.
All the GPL does for you here is stop you getting in the door at a lot of companies, who won't touch your code at all because of the license, but would happily use (and contribute back changes they make to) a BSD licensed project.
Nope, that only applies for customer-facing things. If Google takes, say, the Linux kernel and adds a filesystem driver and some performance optimisations, and runs it in their datacentres, then there is absolutely nothing in GPLv2 or v3 requiring them to share the code with anyone. As the author of the original code, you would get nothing: no money, no code.
When people say 'Linux' they usually mean Linux-and-all-of-the-associated-cruft. Typically this at least includes GNU libc, GNU binutils, and GNU coreutils (which, between them, are more code than the kernel), and typically the GNU shell (bash) and GNU libstdc++. All of these have no moved to GPLv3 (in some cases with the runtime exemption). Remove them, and even though you still have 'Linux' you don't have a system that can run any of your existing code.
Yes there would. Compile a program, give a less-computer-literate friend a copy of the binary. Don't give them the source, because they've no need of it (and can grab it from upstream if they really need it). Forget to include a written offer good for three years to provide the source. You've just violated the GPL, meaning that you have no license to distribute the software, and are therefore a committing copyright infringement (or 'piracy' in the vernacular).
In my experience, it's a little bit more complex than that. To most companies, the GPL is complicated. They can almost certainly use GPL'd code without violating the license, but their lawyers aren't 100% certain. Their lawyers are certain that they can use BSD licensed code without violating the license. Their lawyers are also certain that they can use proprietary code without violating the license, because they get a license that explicitly permits them to do what they want.
I've seen several cases of the GPL driving companies to buy proprietary solutions: given the choice between buying a proprietary license and using free GPL'd code, they'll pick the proprietary solution and limit their (perceived) liability. If there's a BSD licensed alternative, they'll use that and quite often contribute changes back (after all, it's usually cheaper than maintaining a private fork).
The problem with that claim is that it's not even remotely true. For example, consider Google. They have their own private fork of Linux (GPLv2) which includes things like their own filesystem. Some changes are contributed back to the community because maintaining them in a private fork costs more than the loss of competitive advantage from sharing them. Some are kept private, because the scales tip the other way.
In contrast, Yahoo uses a private fork of FreeBSD on a lot of their systems. They employ several FreeBSD developers and contribute a lot of changes back if doing so won't give away a serious competitive advantage, but they keep some things private.
One project has a permissive license, the other has a strong copyleft license, but the behaviour of downstream users is identical in both cases. The GPL doesn't stop you using, modifying, or profiting from the code without giving anything back, it only prevents you from refusing to share the source for your modifications with anyone who receives a binary.
So, you moved from their service because you did the wrong thing? If they ignore a DMCA takedown notice, then they are liable for any copyright infringement. Any US-based hosting company will do the same (which is a good reason not to do business with them, but I digress). If you send a DMCA counter notice, then they are absolved of any liability if they restore the content. The dispute is then between the person filing the notice and the counter notice (and, since the original notice is signed under penalty of perjury, you can probably sue the person who filed it for loss of earnings while the site was down if it was invalid). Sending them other documents, no matter how relevant, does not absolve them of liability.
In short, your complaint makes as much sense as saying 'they asked for my password, I sent them a copy of my passport and drivers license, and they didn't let me connect'.
Momentum is conserved, so if you're both thrusting towards each other and pulling apart you would remain stationary. Or you could just use a padded room and bounce off the walls periodically. The bigger problem with zero-g sex is that most people's reaction to being weightless is nausea, not horniness. Puking from being spun around while already feeling motion sick is a real turn off.
No, that comes under the first bullet point. If you don't know enough C to understand the above sentence, then you almost certainly aren't capable of judging what your employees are capable of accomplishing if your product depends on software. Management doesn't have to be skilled programmers, but they do need to understand the basics of what a computer can do and how long it takes to make a computer do things.
My school had a uniform, and I don't recall any teasing over it (at least, by other people wearing the same thing). Poorer students got theirs second hand, others got them new. I had a mixture because my parents thought the new price for some things (like the tie, for example) were stupidly high given that the second-hand ones looked almost the same. If anything, having a brand new uniform was looked down on as being a bit too poncy.
More importantly, I'm a little surprised that this doesn't open them up to more liability. I was under the impression that the DMCA's safe harbour status only applied to site that didn't actively police their own content. If they do, then they fall under the same rules as a publisher and are liable for wilful commercial infringement.
To retain their safe harbour status, which makes them immune to copyright infringement lawsuits on their hosted content, all that they need to do is take down any content on receipt of a DMCA take-down notice and restore it on receipt of a counter notice.
But efficiency, at least measured in battery life, was something they never seemed to quite get right
Really? It was one of the things that impressed me about my TouchPad. Watching a 45-minute (flash) video on iPlayer took the battery life down to 91% from a full charge. A day of using it as a dev / test machine via ssh used about 50% of the battery life, and it lasts on standby for... actually, I've no idea how long, I've never managed to leave it unplugged long enough to drain more than about 20% of the battery life in standby.
A transparent lie? Of course, that was my first thought too! No one could possibly accidentally type 4.3 when they meant 1.3 - it must be a deliberate attempt at misinformation!
The question is does that commercial backing undermine you (e.g. like OS/X undermined FreeBSD)
How do you think OS X undermined FreeBSD? Apple hired quite a few FreeBSD developers to work on Darwin, but they still contribute stuff back. And things like libdispatch and libc++ that are initially developed on Darwin typically get FreeBSD ports before they run on any other non-Darwin platforms because of the similarities in the libc and kernel interfaces (e.g. libdispatch on FreeBSD just required some #ifdefs around the Mach-port stuff, on Linux it requires you to use libkqueue to emulate some *BSD kernel APIs).
They don't believe them all. For example, the core tenet of Islam is that there is no god except Allah. If you believe in Islam, then you explicitly don't believe in any other religions. You may believe in the abrahamic God, but that doesn't mean that you believe in Islam (or Judaism or any flavour or Christianity).
Except that's not really how it works, because there is not a choice of religion on the one hand and not-religion on the other hand. Instead, there is a choice of either no religion or one from a large number of other religions. Or, as another wise man once said: there are no religious people, just two kinds of atheist. Some disbelieve all religions, some disbelieve all except one.
Management isn't about having the ability to offer an extended work knowledge base to your subordinates. Management is all about leadership and when to crack the whip
This kind of thinking has caused a number of companies to fail. That's part of management, but the most important part of senior management is providing direction (there's a hint in the job title for a lot of them...). This requires them to have a detailed understanding of the market. A successful manager needs to understand three things:
The first is what his subordinates are capable of. When an engineer says 'that's impossible' does he mean 'I can't be bothered?' Or 'we could do that, but only with five times the allotted budget?' How about 'we can't do that with current technology, but we probably will be able to in 2-3 years?' Or does he really mean that it's impossible.
The second is what his competitors are capable of. There is no point in producing a great product a year from now if your competitors can produce almost the same thing in six months and for 10% less. You'll spend a lot on development and then end up with no sales to recoup it.
The third, and often ignored, is what your customers will buy. You can produce the best buggy whip possible, but you won't get many car owners to buy it. This is often the hardest of the three. As His Fordship said 'if I asked my customers what they wanted, they'd have said a faster horse'. Working out what your customers actually want is very difficult, because often they don't know themselves. This also relates to the first point, because what they actually want may not have been possible a few years ago.
Slashdot Mirror
Which is exactly why I love the GPL. I know the code I will release might be reused by other people but they MUST release their modification of it.
So you like the GPL because you don't understand the GPL? 90% of all software developers are employed writing in-house software that is never distributed. If any of these use or modify your code, then they have no legal obligation to release their modifications.
Boo hoo, so write it yourself
They usually do. And they quite often then release it under a permissive (e.g. BSD) license and so other people help find bugs and improve it. Then the next company to come along has a choice of using the GPL or BSDL version. The GPL version may be more mature, but the BSDL one comes with less potential liability, so they use it and improve it. It's cheaper for them to send patches upstream than to maintain a private fork, so they do that. Then another company comes along, sees a fairly mature GPL'd project and a slightly less mature, but more rapidly improving, BSDL project. Guess which one they pick?
That's rather the point of TFA. Given the choice between using GPL'd code or writing it themselves, a lot of people will pick option 2.
I'm referring to the BSD kernel proper
Which BSD kernel? The XNU kernel that Apple uses? They release all of the sources. It's based on CMU Mach, 4BSD, and has more recently some code from the FreeBSD kernel.
Did Apple contribute HFS+ back to BSD?
You mean this code? It's APSL, and someone did port it to FreeBSD, but no one really cared so it bit rotted.
Offhand I can't think of any apple subsystem that made its way into BSD
How about the pthread_workqueue_*() family of functions? They are used by libdispatch to manage a pool of worker threads.
Mach-O loader? no
Why would any BSD want to switch from ELF to Mach-O? You probably could replace rtld with dyld, but what would be the point?
HFS+? no sure
Was ported, no one cared, it died.
Only if there is a single owner. Most non-trivial projects have multiple contributors. I've released code under an MIT license and had some fairly significant improvements contributed back. I am free to use the resulting version of the project in any way I want. There are no restrictions on what I can do with it. If I'd GPL'd the original, then people would have given me GPL'd patches and I would only have been able to use the improved version in accordance with the GPL, although I could use the original version under any terms I wanted.
their contributions to GCC to handle Objective-C
This was NeXT, and it is a case study in the failure of the GPL. Initially, NeXT released their Objective-C front end as a binary-only front end for GCC. The FSF threatened them with lawyers and forced them to release the source. There were just two problems:
Unfortunately, after trumpeting the 'success' of the GPL in forcing NeXT to release the code, they were forced to integrate it with GCC or lose face. They then hacked in some ugly changes to support their own runtime (changes which NeXT / Apple imported into their own branch). If they'd started from scratch and written their own Objective-C front end (which was only about 10KLoC back then), then they'd almost certainly have ended up with better code.
You spend a million hours creating a great program, call it Foo, and you release it under the BSD license. Someone like Microsoft comes along, takes Foo, improves it a little bit, calls it FooBar and distributes it as a binary without publishing any of the code
Your reasoning ignores the fact that off-the-shelf software is a tiny niche - it's responsible for employing about 10% of software developers. For most types of software, there is no Microsoft. There is no big company that will sell you a license to use an existing product. There are, however, lots of companies that will look at your business needs and provide you with some software that addresses them.
The GPL does nothing to protect you from this kind of competition. Let's say you release an improvement to the Linux kernel or a piece of GPL'd software that runs on Linux. When someone wants a Linux-based system, are they likely to go to you, or to a company like Red Hat or IBM? These companies will happily take your code, modify it, and charge their customers for the modifications (actually, charge the customer for the finished product, but typically less than they'd charge for writing it from scratch). The customer then receives the code under the GPL, but has no obligation to share the modifications with you, or with anyone else - indeed, if they give the provide a competitive advantage then they may persuade Red Hat or IBM to keep the modifications private.
All the GPL does for you here is stop you getting in the door at a lot of companies, who won't touch your code at all because of the license, but would happily use (and contribute back changes they make to) a BSD licensed project.
Well, there's the AGPL for that
Nope, that only applies for customer-facing things. If Google takes, say, the Linux kernel and adds a filesystem driver and some performance optimisations, and runs it in their datacentres, then there is absolutely nothing in GPLv2 or v3 requiring them to share the code with anyone. As the author of the original code, you would get nothing: no money, no code.
When people say 'Linux' they usually mean Linux-and-all-of-the-associated-cruft. Typically this at least includes GNU libc, GNU binutils, and GNU coreutils (which, between them, are more code than the kernel), and typically the GNU shell (bash) and GNU libstdc++. All of these have no moved to GPLv3 (in some cases with the runtime exemption). Remove them, and even though you still have 'Linux' you don't have a system that can run any of your existing code.
Yes there would. Compile a program, give a less-computer-literate friend a copy of the binary. Don't give them the source, because they've no need of it (and can grab it from upstream if they really need it). Forget to include a written offer good for three years to provide the source. You've just violated the GPL, meaning that you have no license to distribute the software, and are therefore a committing copyright infringement (or 'piracy' in the vernacular).
to most companies GPL == toxic
In my experience, it's a little bit more complex than that. To most companies, the GPL is complicated. They can almost certainly use GPL'd code without violating the license, but their lawyers aren't 100% certain. Their lawyers are certain that they can use BSD licensed code without violating the license. Their lawyers are also certain that they can use proprietary code without violating the license, because they get a license that explicitly permits them to do what they want.
I've seen several cases of the GPL driving companies to buy proprietary solutions: given the choice between buying a proprietary license and using free GPL'd code, they'll pick the proprietary solution and limit their (perceived) liability. If there's a BSD licensed alternative, they'll use that and quite often contribute changes back (after all, it's usually cheaper than maintaining a private fork).
The problem with that claim is that it's not even remotely true. For example, consider Google. They have their own private fork of Linux (GPLv2) which includes things like their own filesystem. Some changes are contributed back to the community because maintaining them in a private fork costs more than the loss of competitive advantage from sharing them. Some are kept private, because the scales tip the other way.
In contrast, Yahoo uses a private fork of FreeBSD on a lot of their systems. They employ several FreeBSD developers and contribute a lot of changes back if doing so won't give away a serious competitive advantage, but they keep some things private.
One project has a permissive license, the other has a strong copyleft license, but the behaviour of downstream users is identical in both cases. The GPL doesn't stop you using, modifying, or profiting from the code without giving anything back, it only prevents you from refusing to share the source for your modifications with anyone who receives a binary.
So, you moved from their service because you did the wrong thing? If they ignore a DMCA takedown notice, then they are liable for any copyright infringement. Any US-based hosting company will do the same (which is a good reason not to do business with them, but I digress). If you send a DMCA counter notice, then they are absolved of any liability if they restore the content. The dispute is then between the person filing the notice and the counter notice (and, since the original notice is signed under penalty of perjury, you can probably sue the person who filed it for loss of earnings while the site was down if it was invalid). Sending them other documents, no matter how relevant, does not absolve them of liability.
In short, your complaint makes as much sense as saying 'they asked for my password, I sent them a copy of my passport and drivers license, and they didn't let me connect'.
Momentum is conserved, so if you're both thrusting towards each other and pulling apart you would remain stationary. Or you could just use a padded room and bounce off the walls periodically. The bigger problem with zero-g sex is that most people's reaction to being weightless is nausea, not horniness. Puking from being spun around while already feeling motion sick is a real turn off.
No, that comes under the first bullet point. If you don't know enough C to understand the above sentence, then you almost certainly aren't capable of judging what your employees are capable of accomplishing if your product depends on software. Management doesn't have to be skilled programmers, but they do need to understand the basics of what a computer can do and how long it takes to make a computer do things.
My school had a uniform, and I don't recall any teasing over it (at least, by other people wearing the same thing). Poorer students got theirs second hand, others got them new. I had a mixture because my parents thought the new price for some things (like the tie, for example) were stupidly high given that the second-hand ones looked almost the same. If anything, having a brand new uniform was looked down on as being a bit too poncy.
More importantly, I'm a little surprised that this doesn't open them up to more liability. I was under the impression that the DMCA's safe harbour status only applied to site that didn't actively police their own content. If they do, then they fall under the same rules as a publisher and are liable for wilful commercial infringement.
To retain their safe harbour status, which makes them immune to copyright infringement lawsuits on their hosted content, all that they need to do is take down any content on receipt of a DMCA take-down notice and restore it on receipt of a counter notice.
But efficiency, at least measured in battery life, was something they never seemed to quite get right
Really? It was one of the things that impressed me about my TouchPad. Watching a 45-minute (flash) video on iPlayer took the battery life down to 91% from a full charge. A day of using it as a dev / test machine via ssh used about 50% of the battery life, and it lasts on standby for... actually, I've no idea how long, I've never managed to leave it unplugged long enough to drain more than about 20% of the battery life in standby.
A transparent lie? Of course, that was my first thought too! No one could possibly accidentally type 4.3 when they meant 1.3 - it must be a deliberate attempt at misinformation!
The question is does that commercial backing undermine you (e.g. like OS/X undermined FreeBSD)
How do you think OS X undermined FreeBSD? Apple hired quite a few FreeBSD developers to work on Darwin, but they still contribute stuff back. And things like libdispatch and libc++ that are initially developed on Darwin typically get FreeBSD ports before they run on any other non-Darwin platforms because of the similarities in the libc and kernel interfaces (e.g. libdispatch on FreeBSD just required some #ifdefs around the Mach-port stuff, on Linux it requires you to use libkqueue to emulate some *BSD kernel APIs).
They don't believe them all. For example, the core tenet of Islam is that there is no god except Allah. If you believe in Islam, then you explicitly don't believe in any other religions. You may believe in the abrahamic God, but that doesn't mean that you believe in Islam (or Judaism or any flavour or Christianity).
Except that's not really how it works, because there is not a choice of religion on the one hand and not-religion on the other hand. Instead, there is a choice of either no religion or one from a large number of other religions. Or, as another wise man once said: there are no religious people, just two kinds of atheist. Some disbelieve all religions, some disbelieve all except one.
You should sell your company to a big company in exchange for a seat on the board and enough money to slack off all of the time, obviously...
Management isn't about having the ability to offer an extended work knowledge base to your subordinates. Management is all about leadership and when to crack the whip
This kind of thinking has caused a number of companies to fail. That's part of management, but the most important part of senior management is providing direction (there's a hint in the job title for a lot of them...). This requires them to have a detailed understanding of the market. A successful manager needs to understand three things:
The first is what his subordinates are capable of. When an engineer says 'that's impossible' does he mean 'I can't be bothered?' Or 'we could do that, but only with five times the allotted budget?' How about 'we can't do that with current technology, but we probably will be able to in 2-3 years?' Or does he really mean that it's impossible.
The second is what his competitors are capable of. There is no point in producing a great product a year from now if your competitors can produce almost the same thing in six months and for 10% less. You'll spend a lot on development and then end up with no sales to recoup it.
The third, and often ignored, is what your customers will buy. You can produce the best buggy whip possible, but you won't get many car owners to buy it. This is often the hardest of the three. As His Fordship said 'if I asked my customers what they wanted, they'd have said a faster horse'. Working out what your customers actually want is very difficult, because often they don't know themselves. This also relates to the first point, because what they actually want may not have been possible a few years ago.
I'd believe most of what a scientist says to a journalist. I have more of a problem believing things that a journalist hears from a scientist...