The use of the BerkeleyDB do not put any restrictions on your software, as long as you do not statically link it.
The Sleepycat license doesn't trigger based on linking; it's triggered by compiling against it. See The Sneaky Sleepycat License and comments from Oracle's forums. The existing license was already very "viral" in terms of how aggressively it required either open source distribution or a commercial license.
phpBB is currently under the GPLv2. The person you replied to didn't say they are unwilling to share their source code, just that such work would be unproductive. Do you always rant about straw men like this?
The FSF's best practice for software licenses involves zero license fees, always. There are multiple practices the FSF follows that people accept only because they are the FSF. Copyright assignment is another thing the FSF can do, but when it's adopted by a commercial company it's presumed they are violating the spirit of free software by taking contributor work into a private commercial version. The exact motives and license recommendations of the FSF may change over time, but they are transparent and as consistent as they can be. When a commercial entity emulates part of their behavior, but they add a profit motive and license inconsistency, they are not acting like the FSF anymore at all.
There was a very clear meaning for "pattern" in a computer gaming context by 1980. In 1980 Pac-Man was first released in the US. By 1981 the word "pattern" to describe navigating the maze was so popular that the "Pac-Man Fever" album included patterns for each level.
It's (1), Sid Meier, super genius. I've spoken to the man twice, the tech side of Baltimore where we both live is pretty small. Sid is exactly the sort of guy who will stare at a game, note the patterns, and then figure out what algorithms must be driving them, all while a regular person is just playing. There is not a hint of boasting from the guy in person, he's just that good at what he does.
Ubuntu has been bootable on a Mac from a USB drive for a while. Before 12.04.2 it was harder to then install the system sometimes. See UEFI for comments on what changed. Everything should work at this point on versions after that one. Recent changes in Knoppix should allow this latest version to work too. There are a lot more people working on Mac support in Ubuntu than Knoppix though.
If you have a PC system available, it's helpful to test booting there, so you can be sure the drive is fine before moving on to fighting with whatever EFI issues pop up. It should be possible to make the drive bootable just by holding down the Option key during boot. But there are a few common EFI headaches that get easier if you just install rEFind on your boot drive, that's what I always do.
The Chief here says basically that if you don't let them have their way, you won't be able to use their services.
That is the only useful point to take away from this. DRM advocates regularly use extortion tactics to get what they want. The most important thing to know about paying protection money is that you'll keep paying forever once you start. Even worse, the earlier money collected will be backing later threats. Kipling got it right a long time ago.
Marketing will try to sell things to anyone that will buy them. The only companies who can easily demonstrate a real need for Oracle are larger installations. My job involves getting a lower total cost of ownership out of PostgreSQL for the sizes it can handle, rough practical limit is around 5TB right now, and that's easy to pull off in most cases. For a fraction of what an Oracle license costs, you can get support from a company that does source code level contributor to one of the open-source databases. And since you're not dependent on the company itself for help, you can take on as much of the administrator work as you'd like. There are so many horror stories about companies getting giant monstrosities built because they got Oracle consultants involved in a project. They love using the fancy features because it gives job security and your license costs go up.
Oracle's marketing is starting to dip into FUD mode now about just how viable the open-source alternatives are to them. It's pretty funny to watch that coming out of one end, while they continue buying those very companies with the other. The main thing Oracle has going for it is that while good people are expensive, there are a lot of them out there. It is pretty hard to hire a good PostgreSQL DBA right now.
Recovery is also awesome. "ALTER DATABASE RECOVER UNTIL [timestamp]", "ALTER DATABASE RECOVERY UNTIL CANCEL", "ALTER DATABASE UNTIL CHANGE [transaction number]" and so on.
PostgreSQL has had this sort of feature for a few years now, see recovery target. The main thing Oracle is better at is giving a simpler UI to this work. This sort of ALTER DATABASE interface is easier to use than the config file approach Postgres uses. Similarly, Oracle Flashback is a slicker interface to look at old data than you get out of the box with Postgres, but the same under the hood look is possible in both databases.
The differences in the backup/recovery area are pretty small now though, especially now that UI clones such as pgbarman are available. The last time I wanted a better UI to part of this that wasn't in any of the available free tools yet, I paid a few thousand dollars to the pgbarman developers and they added it for me. Companies who are willing to spend some money on licensed software really should consider what spending the same amount on improving PostgreSQL would get them.
Yes, people can copy a program, modify it, and only release the source to users who get the binaries. That's usually a meaningless edge case now though. In your average case, where the source and binaries are being distributed via the web to anyone who wants it, the GPL terms will still compel release of the modified source code.
And in the case where people make large changes such that they can't be merged upstream, who cares? Let them hack up a private version and maintain that junk forever. I don't want crappy code like that anyway. Why get stressed about "oh noes, people will fork my code and ruin it and not share"...just ignore those forkers and move along. If your changes won't merge upstream, there's some bad news: you won't be able to pull from upstream either. People who haven't figured that out are going to fail at open source work anyway, better to ignore them altogether.
Joomla is acting like the library component here, because the extensions are using its APIs. If Joomla added an exception that said "you can use the Joomla API under conditions X in non-GPL programs", yes that would be similar to things like the OpenSSL exception. But if the Joomla is GPL and is serious about advocating that license for their own ecosystem, they shouldn't do that without a seriously good reason. One-off licenses with exceptions create legal headaches that make them less useful than ones using a standard license.
This sort of thing is a hack to cope with projects with bad license, not something you should be relying on for code where you can influence them to pick the "right" license--where rightness means aligning with upstream code license choices. All of the GPL-with-exception licenses are pragmatic choices, where crappy license terms are worked around because they can't be easily removed. That's not a model anyone should use if it's possible to avoid it, and it seems Joomla could avoid it here if they tried. That's why the spirit of the license is relevant for the right thing to do in this situation.
I won't deal with GPL projects because people will just fork it so they don't have to give their changes back.
Forking a project doesn't change its license that way. You can complain about people who outright ignore the code's license, and those who bypass it via "reverse engineering" that's mostly copying code between projects. Those both happen, and that's a problem sometimes. But neither of those can be called forking.
These trademark issues aren't even specific to RedHat or CentOS. Debian even rebrands their version of Firefox and Thunderbird as IceWeasel and IceDove, also due to trademark restrictions.
If it's possible to rebrand a program easily in this way, I take that as proof the artwork involved is not really an essential part of the program. The changes of the CentOS and Debian artwork forks do not diminish any feature of the software. That can't be said about most CSS and Javascript, and even some images are mandatory for a program to be useful.
Of course the GPL has a spirit that goes beyond its legal wording. The clearest example of that in action was when the GPL v3 was introduced to block things like Tivoization. That sort of loophole was against the spirit of the license, but not explicitly blocked by it. Once the problem was clearly identified, the legal language of the license was updated. You can think of that as turning more of the spirit of the license into explicit legalese. People who think the wording of these licenses are some sort of challenge, where victory is finding a loophole, can expect the licenses to evolve against them too.
Saying that the FSF wasn't complaining suggests they were happy with the license. Looking over the longer timeline, I see it more as that they just stopped complaining so much in public, probably because they were worn out by then. They had already been bitching for a long time by 2005, they had gotten significant license concessions, and the end goal they really wanted (code under one free license) was very clear by then.
Much easier for who though? It's certainly easier for manufacturers to move all their firmware issues so the OS has to deal with them. But the cost of doing that work is being pushed toward kernel developers and packagers. Whether the end result is better or worse is complicated, but that's not why people like Stallman complain. What you can't argue with is that it's frustrating for a Linux kernel developer to spend time chasing down a bug that's actually inside of the firmware blob, or in the part of the code that is uploading the thing to the card. They often end up working blind in situations where source code would make things easier. And for many of those developers, wanting to have source code to everything relevant is why they work on open source software.
The point of things like Debian's social contract is that if you follow the guidelines and fit into the development community well, that community will both work on supporting your hardware and recommend its purchase. When a company doesn't do that, the people who have to handle the job don't like it. Now, if you like a Linux where binary blobs are a first class citizen, good for you. That sort of thing is how Ubuntu became more popular than Debian. But when a company doesn't do that, they can't expect to be a preferred vendor by the strictly licensed projects. Debian and others with similar goals are not going to ignore their rules and make a special exception for anyone.
There are a few types of overhead involved in firmware distribution, and making that part of your system software pushes that work toward open source communities in a way they resent. If you look at things like Debian's policy, none of these blobs fit their guidelines. That means those firmware blobs go into their non-free repository. That wart is annoying enough that people regularly try to eliminate it altogether. All of that means some of the overhead manufacturers are saving by not having flash on the board is being passed on to free software packagers instead.
Similarly, there's also a very real cost involved with building, maintaining, and distributing the firmware updates and flashing code. Why should the Linux community pay for that? They'll do it, sure, but don't be surprised when people prefer options that avoid it.
This is not just history, it's ongoing. MariaDB requires copyright assignment for contributing code today. They claim "This is needed so that Monty Program Ab can give your code forwards to other projects (like MySQL)." But that's slight of hand to distract from the real issue here. If MySQL won't accept straight GPL code, that's their problem to fix, not mine to work around as a contributor. When I support a new fork of something I don't worry about pandering to the needs of the other forks anymore. That is the whole point of forking code--you shouldn't even go there unless you have given up on working within the constraints of the original project.
Copyright assignment isn't always associated with bad behavior; the FSF requires it too. But when you're assigning your copyright to someone who has taken advantage of dual-licensing to their personal benefit before, it's really hard to see it as anything other than self-serving the next time they try.
Start by making intellectual property rights vest only in the creator, and make it non transferrable. This will force commercial entities to grant a fair share of the profits to the real innovators instead of the giving an unearned bonus to the patent troll who own a large number of the patents today.
Dream on. What would actually happen is a combination of two things. One, employees doing research for large companies will get paid less, on the theory that now they get directly rewarded for the things they invent. It will work just like tipping does for food service people: their base income will drop because this expected (but not guaranteed) revenue would get lumped in as part of their compensation. Of course researchers will actually make nothing that way in the average case, due to the large cost of prosecuting a patent infringement, and the result will be they are paid less overall.
Second, really vital patents will end up assigned to the company management instead of the researchers, and then those people will license them out. Don't like it? You're fired, and we'll take your name off the patent application too the minute your ass is out the door. (This has happened to me) Instead of buying the patent, in this new arrangement you'll buy a share in the guy who owns it, and that won't be the real people at the bottom doing the work anymore. Currently patent rights when you work at a large company turns into a resume builder, while the company takes the profits from that work. Restructure the profits, and you'll find the credit moves away from the real inventor too. "That's just business" says the sort of executive who will happily execute the new style of employee agreements.
Don Lancaster wrote an interesting book twenty-ish years ago called "The Incredible Secret Money Machine", and there's a chapter in there about patent protection. He makes the point that if you really are the sort of genius who can invent something brand new as a researcher, the odds you are also a savvy business man who knows how to make a buck from that idea are almost zero. The only thing I would add to that is that you can't even just make the buck--you have to keep it away from assholes with money who'd prefer to litigate it out of you too.
That's generous. Ideas without a working implementation are in almost every case worth $0. If it's innovative enough to be considered worthy of a patent, it will be complicated enough that you can't prove it will work without building at least a prototype.
And many of the cases where something can be built, but couldn't until now, are simply based on underlying technology being available. Every year I churn out a few ideas that, while new at the time, are obvious next steps based on the current state of the art. That is the category Personal Audio's junk patents fall into. I "invented" a web site for online auctions in the summer of 1995, just before ebay started as AuctionWeb. Was that worth 10% of ebay? Of course not; it was worth nothing.
What these patent trolls are doing is a form of curve fitting. Companies throw out a bunch of garbage, hand waving over the details, filing patents for what the next generation of products might look like. Then, after the world sees which of them turned out to be right--which some of them will be just by random chance--the owners of the lucky patents sue. If many people re-invent your idea without taking any advice on how to do it from your patent, that alone should torpedo any patent as being an obvious next step.
I asked the EFF question, surprised that made it though. Describing the EFF as a sort of digital oriented ACLU is a great analogy for explaining its role to people unfamiliar with it, thanks for that.
Slashdot Mirror
The use of the BerkeleyDB do not put any restrictions on your software, as long as you do not statically link it.
The Sleepycat license doesn't trigger based on linking; it's triggered by compiling against it. See The Sneaky Sleepycat License and comments from Oracle's forums. The existing license was already very "viral" in terms of how aggressively it required either open source distribution or a commercial license.
phpBB is currently under the GPLv2. The person you replied to didn't say they are unwilling to share their source code, just that such work would be unproductive. Do you always rant about straw men like this?
The FSF's best practice for software licenses involves zero license fees, always. There are multiple practices the FSF follows that people accept only because they are the FSF. Copyright assignment is another thing the FSF can do, but when it's adopted by a commercial company it's presumed they are violating the spirit of free software by taking contributor work into a private commercial version. The exact motives and license recommendations of the FSF may change over time, but they are transparent and as consistent as they can be. When a commercial entity emulates part of their behavior, but they add a profit motive and license inconsistency, they are not acting like the FSF anymore at all.
There was a very clear meaning for "pattern" in a computer gaming context by 1980. In 1980 Pac-Man was first released in the US. By 1981 the word "pattern" to describe navigating the maze was so popular that the "Pac-Man Fever" album included patterns for each level.
It's (1), Sid Meier, super genius. I've spoken to the man twice, the tech side of Baltimore where we both live is pretty small. Sid is exactly the sort of guy who will stare at a game, note the patterns, and then figure out what algorithms must be driving them, all while a regular person is just playing. There is not a hint of boasting from the guy in person, he's just that good at what he does.
Ubuntu has been bootable on a Mac from a USB drive for a while. Before 12.04.2 it was harder to then install the system sometimes. See UEFI for comments on what changed. Everything should work at this point on versions after that one. Recent changes in Knoppix should allow this latest version to work too. There are a lot more people working on Mac support in Ubuntu than Knoppix though.
If you have a PC system available, it's helpful to test booting there, so you can be sure the drive is fine before moving on to fighting with whatever EFI issues pop up. It should be possible to make the drive bootable just by holding down the Option key during boot. But there are a few common EFI headaches that get easier if you just install rEFind on your boot drive, that's what I always do.
Hack into a computer at a US airport like this, and in addition to free WiFi they'll include a free trip to Cuba!
The Chief here says basically that if you don't let them have their way, you won't be able to use their services.
That is the only useful point to take away from this. DRM advocates regularly use extortion tactics to get what they want. The most important thing to know about paying protection money is that you'll keep paying forever once you start. Even worse, the earlier money collected will be backing later threats. Kipling got it right a long time ago.
You can never donate too much to politicians. That spending has some of the best return on investment around!
Marketing will try to sell things to anyone that will buy them. The only companies who can easily demonstrate a real need for Oracle are larger installations. My job involves getting a lower total cost of ownership out of PostgreSQL for the sizes it can handle, rough practical limit is around 5TB right now, and that's easy to pull off in most cases. For a fraction of what an Oracle license costs, you can get support from a company that does source code level contributor to one of the open-source databases. And since you're not dependent on the company itself for help, you can take on as much of the administrator work as you'd like. There are so many horror stories about companies getting giant monstrosities built because they got Oracle consultants involved in a project. They love using the fancy features because it gives job security and your license costs go up.
Oracle's marketing is starting to dip into FUD mode now about just how viable the open-source alternatives are to them. It's pretty funny to watch that coming out of one end, while they continue buying those very companies with the other. The main thing Oracle has going for it is that while good people are expensive, there are a lot of them out there. It is pretty hard to hire a good PostgreSQL DBA right now.
Recovery is also awesome. "ALTER DATABASE RECOVER UNTIL [timestamp]", "ALTER DATABASE RECOVERY UNTIL CANCEL", "ALTER DATABASE UNTIL CHANGE [transaction number]" and so on.
PostgreSQL has had this sort of feature for a few years now, see recovery target. The main thing Oracle is better at is giving a simpler UI to this work. This sort of ALTER DATABASE interface is easier to use than the config file approach Postgres uses. Similarly, Oracle Flashback is a slicker interface to look at old data than you get out of the box with Postgres, but the same under the hood look is possible in both databases.
The differences in the backup/recovery area are pretty small now though, especially now that UI clones such as pgbarman are available. The last time I wanted a better UI to part of this that wasn't in any of the available free tools yet, I paid a few thousand dollars to the pgbarman developers and they added it for me. Companies who are willing to spend some money on licensed software really should consider what spending the same amount on improving PostgreSQL would get them.
Yes, people can copy a program, modify it, and only release the source to users who get the binaries. That's usually a meaningless edge case now though. In your average case, where the source and binaries are being distributed via the web to anyone who wants it, the GPL terms will still compel release of the modified source code.
And in the case where people make large changes such that they can't be merged upstream, who cares? Let them hack up a private version and maintain that junk forever. I don't want crappy code like that anyway. Why get stressed about "oh noes, people will fork my code and ruin it and not share"...just ignore those forkers and move along. If your changes won't merge upstream, there's some bad news: you won't be able to pull from upstream either. People who haven't figured that out are going to fail at open source work anyway, better to ignore them altogether.
Joomla is acting like the library component here, because the extensions are using its APIs. If Joomla added an exception that said "you can use the Joomla API under conditions X in non-GPL programs", yes that would be similar to things like the OpenSSL exception. But if the Joomla is GPL and is serious about advocating that license for their own ecosystem, they shouldn't do that without a seriously good reason. One-off licenses with exceptions create legal headaches that make them less useful than ones using a standard license.
This sort of thing is a hack to cope with projects with bad license, not something you should be relying on for code where you can influence them to pick the "right" license--where rightness means aligning with upstream code license choices. All of the GPL-with-exception licenses are pragmatic choices, where crappy license terms are worked around because they can't be easily removed. That's not a model anyone should use if it's possible to avoid it, and it seems Joomla could avoid it here if they tried. That's why the spirit of the license is relevant for the right thing to do in this situation.
I won't deal with GPL projects because people will just fork it so they don't have to give their changes back.
Forking a project doesn't change its license that way. You can complain about people who outright ignore the code's license, and those who bypass it via "reverse engineering" that's mostly copying code between projects. Those both happen, and that's a problem sometimes. But neither of those can be called forking.
These trademark issues aren't even specific to RedHat or CentOS. Debian even rebrands their version of Firefox and Thunderbird as IceWeasel and IceDove, also due to trademark restrictions.
If it's possible to rebrand a program easily in this way, I take that as proof the artwork involved is not really an essential part of the program. The changes of the CentOS and Debian artwork forks do not diminish any feature of the software. That can't be said about most CSS and Javascript, and even some images are mandatory for a program to be useful.
Of course the GPL has a spirit that goes beyond its legal wording. The clearest example of that in action was when the GPL v3 was introduced to block things like Tivoization. That sort of loophole was against the spirit of the license, but not explicitly blocked by it. Once the problem was clearly identified, the legal language of the license was updated. You can think of that as turning more of the spirit of the license into explicit legalese. People who think the wording of these licenses are some sort of challenge, where victory is finding a loophole, can expect the licenses to evolve against them too.
Saying that the FSF wasn't complaining suggests they were happy with the license. Looking over the longer timeline, I see it more as that they just stopped complaining so much in public, probably because they were worn out by then. They had already been bitching for a long time by 2005, they had gotten significant license concessions, and the end goal they really wanted (code under one free license) was very clear by then.
Much easier for who though? It's certainly easier for manufacturers to move all their firmware issues so the OS has to deal with them. But the cost of doing that work is being pushed toward kernel developers and packagers. Whether the end result is better or worse is complicated, but that's not why people like Stallman complain. What you can't argue with is that it's frustrating for a Linux kernel developer to spend time chasing down a bug that's actually inside of the firmware blob, or in the part of the code that is uploading the thing to the card. They often end up working blind in situations where source code would make things easier. And for many of those developers, wanting to have source code to everything relevant is why they work on open source software.
The point of things like Debian's social contract is that if you follow the guidelines and fit into the development community well, that community will both work on supporting your hardware and recommend its purchase. When a company doesn't do that, the people who have to handle the job don't like it. Now, if you like a Linux where binary blobs are a first class citizen, good for you. That sort of thing is how Ubuntu became more popular than Debian. But when a company doesn't do that, they can't expect to be a preferred vendor by the strictly licensed projects. Debian and others with similar goals are not going to ignore their rules and make a special exception for anyone.
There are a few types of overhead involved in firmware distribution, and making that part of your system software pushes that work toward open source communities in a way they resent. If you look at things like Debian's policy, none of these blobs fit their guidelines. That means those firmware blobs go into their non-free repository. That wart is annoying enough that people regularly try to eliminate it altogether. All of that means some of the overhead manufacturers are saving by not having flash on the board is being passed on to free software packagers instead.
Similarly, there's also a very real cost involved with building, maintaining, and distributing the firmware updates and flashing code. Why should the Linux community pay for that? They'll do it, sure, but don't be surprised when people prefer options that avoid it.
The only reason GNOME exists is exactly because the FSF objected to the Qt license. See Stallman on Qt for some history there.
This is not just history, it's ongoing. MariaDB requires copyright assignment for contributing code today. They claim "This is needed so that Monty Program Ab can give your code forwards to other projects (like MySQL)." But that's slight of hand to distract from the real issue here. If MySQL won't accept straight GPL code, that's their problem to fix, not mine to work around as a contributor. When I support a new fork of something I don't worry about pandering to the needs of the other forks anymore. That is the whole point of forking code--you shouldn't even go there unless you have given up on working within the constraints of the original project.
Copyright assignment isn't always associated with bad behavior; the FSF requires it too. But when you're assigning your copyright to someone who has taken advantage of dual-licensing to their personal benefit before, it's really hard to see it as anything other than self-serving the next time they try.
Start by making intellectual property rights vest only in the creator, and make it non transferrable. This will force commercial entities to grant a fair share of the profits to the real innovators instead of the giving an unearned bonus to the patent troll who own a large number of the patents today.
Dream on. What would actually happen is a combination of two things. One, employees doing research for large companies will get paid less, on the theory that now they get directly rewarded for the things they invent. It will work just like tipping does for food service people: their base income will drop because this expected (but not guaranteed) revenue would get lumped in as part of their compensation. Of course researchers will actually make nothing that way in the average case, due to the large cost of prosecuting a patent infringement, and the result will be they are paid less overall.
Second, really vital patents will end up assigned to the company management instead of the researchers, and then those people will license them out. Don't like it? You're fired, and we'll take your name off the patent application too the minute your ass is out the door. (This has happened to me) Instead of buying the patent, in this new arrangement you'll buy a share in the guy who owns it, and that won't be the real people at the bottom doing the work anymore. Currently patent rights when you work at a large company turns into a resume builder, while the company takes the profits from that work. Restructure the profits, and you'll find the credit moves away from the real inventor too. "That's just business" says the sort of executive who will happily execute the new style of employee agreements.
Don Lancaster wrote an interesting book twenty-ish years ago called "The Incredible Secret Money Machine", and there's a chapter in there about patent protection. He makes the point that if you really are the sort of genius who can invent something brand new as a researcher, the odds you are also a savvy business man who knows how to make a buck from that idea are almost zero. The only thing I would add to that is that you can't even just make the buck--you have to keep it away from assholes with money who'd prefer to litigate it out of you too.
Coming up with an idea is 10% of the effort.
That's generous. Ideas without a working implementation are in almost every case worth $0. If it's innovative enough to be considered worthy of a patent, it will be complicated enough that you can't prove it will work without building at least a prototype.
And many of the cases where something can be built, but couldn't until now, are simply based on underlying technology being available. Every year I churn out a few ideas that, while new at the time, are obvious next steps based on the current state of the art. That is the category Personal Audio's junk patents fall into. I "invented" a web site for online auctions in the summer of 1995, just before ebay started as AuctionWeb. Was that worth 10% of ebay? Of course not; it was worth nothing.
What these patent trolls are doing is a form of curve fitting. Companies throw out a bunch of garbage, hand waving over the details, filing patents for what the next generation of products might look like. Then, after the world sees which of them turned out to be right--which some of them will be just by random chance--the owners of the lucky patents sue. If many people re-invent your idea without taking any advice on how to do it from your patent, that alone should torpedo any patent as being an obvious next step.
I asked the EFF question, surprised that made it though. Describing the EFF as a sort of digital oriented ACLU is a great analogy for explaining its role to people unfamiliar with it, thanks for that.
Only a Miracle can save them now!