So, what if there are no other ways of doing something because the patent has been wisely written not to stray too far into implementation? It could preclude even the existence of your FOSS program. Period.
Patents are on specific functionality, and patents are published and well-known. For every major piece of FOSS, people have a good idea of what the patent-related risks are.
For most pieces of FOSS, any patents that could possibly read on it apply to small, optional features.
There are a few FOSS packages where the primary functionality might infringe on some patent (e.g., panorama software); those tend to be highly specialized and both the authors and users are aware of the risk and have taken it into account. In fact, exactly the same case occurs frequently for commercial software.
Statements about things that "might" and "could" happen are just FUD. Patents don't strike out of the blue, and the kind of generalities in which people like you talk are designed to hurt FOSS.
I'm sure there's a ton of patents (valid or not) winding their way through the patent process for the Office 2007 and Windows Vista interfaces.
Well, both tabbed toolbars ("ribbons") and galleries have plenty of prior art, so if those patents actually matter, they will be invalidated. Except for the specific visual style, OpenOffice, in fact, already has the functionality in several places.
Getting rid of the main menu bar is such an incredibly stupid idea that it's hard to understand what Microsoft could possibly be thinking. The menu bar gives you an enormous amount of bang-for-the-buck in terms of user interface.
Microsoft has been investing Billions in research and design for years, the new Office UI is simply an extension of that.
If the Office 2007 the result of spending "billions in research and design", Microsoft's R&D isn't paying off. The Office UI is not innovative; those ideas have been around for a while, and many applications already use them.
Microsoft will always spend Billions on creating slicker and easier to use interfaces.
I think it's highly debatable whether the new Office UI is "easier to use", and Microsoft has provided zero evidence that it is.
Office 2007 is a great opportunity for FOSS to make real advances in UI design, after Microsoft dropped the ball so badly with their next generation products.
I doubt Java would meet anybody's desktop quality standards.
KDE has the problem that commercial software development for it is far more expensive than for Gnome, Windows, or Macintosh. People may like KDE, but when push comes to shove for big companies like Sun, IBM, or Google to support a desktop platform, KDE is a non-starter.
So, I think if Google were to pick a desktop, it would be Gnome or a modified version of it.
In support of the nocturnal mammal evolution theory, our vision system has only three colors whereas most non-mammals have four.
Actually, current thinking is that somewhere along the line, our ancestors lost all color vision (or were down to two colors), and humans managed to re-evolve color vision.
However, if your software uses some patented algorithm (yes, I know, that's a hideous turn of phrase) then odds are you'd never risk releasing the source, so you could probably infringe forever and nobody would ever know.
If the algorithm is important and difficult to work around, then the patent holder will know even if you ship just binaries, and they can and will compel you to produce source code. In addition, your behavior will likely be interpreted as willful infringement, exposing you to extra damages.
That is really the greatest threat to open source software from software patents: the fact that it is substantially easier to determine if an open source package is infringing. In a litigious environment, it's easy to say, "why take the risk?"
That's pure FUD (do you work for Microsoft?). Open source has been around for several decades, and I'm not aware of any serious consequences for end users from patent infringement by FOSS. First of all, for the very reasons you mention--people know they are being scrutinized--patent infringement by FOSS is rare, and when it does, people simply remove the offending code.
You're far more at risk with closed source software--infringement seems to be far more frequent, lawsuits happen often and with serious consequences, and whether you as the customer are directly liable for infringement or not, you will often still face substantial costs if your vendor is found guilty.
So far as DOJ scrutiny is concerned, does anyone know if the DOJ has ever charged a large corporation with an antitrust violation for using a patent portfolio to suppress competition, given that that is the intended function of patents?
There is ample precedent for the government interfering in how companies license patents. But what's at issue here is not the exclusionary nature of patents in general, it's the inequitable way in which it is being used: companies who cross-license the entire portfolio have no costs, while newcomers to the market may not be able to enter at all.
Patent portfilios for the Microsoft/IBM/Oracle/Sun/HP crowd (or Intel/AMD/nVidia/ATI for that matter) have become exactly the same kind of "Mutually Assured Desctruction" scenario.
Actually, that's not all they are: they are also barriers to entry, because small, commercial, closed-source competitors find it hard to enter a market in this situation. That's not what the patent system was supposed to do. And, sooner or later, it may lead to some serious scrutiny by the DOJ.
Nevertheless, it may work to the advantage of open source, since it means that new software companies may find it advantageous to figure out open source models for software that they would otherwise have released under a proprietary license.
MySQL is absolutely trivial to set up; on most Linux distributions, setting it up is a single command to install the package, create the databases, install the service, and start it up.
There are standard GUIs for administering MySQL, probably for people like you, but most people I know prefer the command line. The command line is easier even for non-DBAs, because it's far easier to write down and document what needs to be done.
An experienced DBA can set up a new installation [of Microsoft SQL Server] in a couple of minutes
Probably true. But with MySQL, you don't even have to be a DBA at all, and it's probably even quicker.
I think it's good to equip police with cameras. It deters "offensive behavior" not only against the police, but by the police. However, what needs to be in place is a mandatory requirement that these cameras be used, that citizens have the same rights as the government to have them used in court, and that they be tamper-proof and cryptographically secured.
Hopefully, even if the government that's putting in the cameras isn't thinking of doing this, courts will sooner or later force the government to comply.
um , they do not encourage people to release stuf under LGPL, quite the opposite.
Quite right: they do not "encourage" people to release stuff under the LGPL, but they do use the LGPL or GPL w/linking exception when it serves the purposes of open source software.
The GNU C library, as well as Gtk continue to be released under licenses that permit commercial developers to deliver applications for free on Linux. The FSF and lots of other people believe that that's important.
Troll Tech deliberately makes it costly for commercial developers to develop for desktop and embedded Linux because that's where they make their profit, but in doing so, they make Linux uncompetitive for commercial developers. Fortunately, in the case of desktop Linux, we have a good alternative. For embedded Linux, Troll Tech's licensing structure has probably doomed the platform for now until the open source world can establish a good, free alternative standard.
You and Troll Tech are trying to confuse the issue by talking about generalities and association: yes, the GPL is a FSF license and the FSF generally discourages non-GPL licenses. But each license needs to be considered on its own merits, and Troll Tech's licensing structure is harmful to the goals of free software.
If Microsoft had anything substantial and usable, they'd have sued by now, but they probably have figured out that that's pointless. Many of their patents are probably invalid or unenforceable, or even have prior art in open source software.
Furthermore, FOSS developers try hard to avoid infringing on people's patents, and Microsoft's patents are scrutinized, so the number of infringing software packages is likely small. In the few cases where Microsoft might have a valid patent claim against a piece of FOSS and could actually identify someone to sue, it would be hard for them to be able to claim willful infringement or get any real damages, and the infringing code would be removed instantaneously, making the case fall apart.
If Microsoft actually believes they have IP that's being violated, they should stop bluffing and start asserting it in court. That way, they can get what they deserve, and they create certainty for everybody else. Of course, certainty is the last thing they want.
I've never heard of Christians beheading people- though this response does seem mild in comparison to what other Islamic sects do.
I think the traditional Christian methods are burning at the stake, drowning, and stoning. They were practiced widely until the rise of secular society in the West finally put a lid on the most egregious religious excesses (fairly recently, historically speaking).
The stated goal of Islam is to convert the entire world, by the sword if necessary.
So is, unfortunately, the stated goal of Christianity. Judaism doesn't try to convert much, it simply preaches separateness and superiority. No matter which way you look at it, the Abrahamic religions are pretty messed up.
The issue isn't whether you have a problem with paying for it, the issue is that people should understand the consequences of their choice and make an informed decision. The KDE/QPL license problem itself is a good example that people don't. I'm not convinced you do, but that's your problem.
Good, you found the right quote. Now, you seem to be implying that people think that Qt offers "a significant unique capability" and should therefore be released GPL, not LGPL.
Well, I disagree; toolkits for C++ are a dime a dozen. But, more importantly, the FSF disagrees, since they have put their own toolkit under the LGPL.
Qt's licensing costs are way out of line with what people pay for Mac and Windows development. If Qt were the only toolkit for Linux, Linux development would cost a premium. That doesn't bother the kind of premium customers Troll Tech likes, but it would have killed desktop Linux if it had been the only choice. Fortunately, on the desktop, there are alternatives, in part precisely because the FOSS community did not want to hobble the Linux desktop with that sort of burden. However, for phones, there is no alternative, and Qt/Embedded's licensing structure is probably in part responsible for the failure of Linux for phones and PDAs.
When dealing with the OSS community, TrollTech could easily have kept with their original QPL license, or even gone fully closed-source
Good! I think that would have been the right thing to do.
So you blame Troll Tech for wanting to run a viable company that makes money -- you think they should be writing top-quality code for the good of the Linux community alone
I don't "blame" them, I'm saying people shouldn't use their software. I'm saying that because people probably don't understand the costs and implications involved, and because it hurts the open source community.
The above makes me question your understanding of the GPL... or maybe you just mistyped something above.
I meant exactly what I said: the creators of the GPL chose, for many libraries, the LGPL and the GPL-with-linking-exception in order to enable commercial developers to develop freely for open source platforms. Troll Tech deliberately chose the GPL precisely because they didn't have that goal.
Open source is not a religion, it's just a tool like anything else.
Quite right. And I'm saying: Qt is a bad deal for both commercial developers and for open source, for simple, rational business and policy reasons. It's a bad deal for commercial developers because of hidden costs. It's a bad deal for open source because it discourages commercial developer from moving to open source platforms because Qt is so expensive compared to Windows and Macintosh.
Seems to me they are getting a top-quality GUI library
People permit commercial usage of Linux and many FOSS libraries in the hope that others create useful commercial applications and thereby actually help open source platforms and APIs succeed.
But in the phone market, Troll Tech has managed to insert itself into Linux in such a way that they are the only ones reaping financial benefit. And Troll Tech designed their system so that other toolkits would be difficult to port to their platform, and they made license agreements with manufacturers that ensured that users wouldn't have a choice. And they screwed up so badly with their platform that they probably put off lots of manufacturers from it.
Of course, they can legally do all that--the license permits it.
But people like me can and should also point out how harmful their behavior is for open source and users in the long run, just like people can and should point out how Microsoft's licensing model is a bad deal for users.
You are welcome to accept their terms, or reject them -- nobody is forcing you to develop with Qt if you don't like their license.
Indeed, I don't develop with Qt. And I am trying to convince others not to develop with Qt either because I think it's a lousy deal for the open source community. I'm saying that if more people behaved like Troll Tech did, open source would die. In fact, I think Troll Tech is largely responsible for the failure of Linux on phones and PDAs.
Now, what's your problem with me stating my opinion and analysis? In your world, should everybody other than corporate marketing just shut up?
Troll Tech isn't unfairly taking advantage of anything.
Oh, yes, they are. I mean, look at what happened with the KDE project: originally, the started using Qt despite the fact that their license was incompatible with the QPL. People don't understand licenses well, and people don't understand the consequences of choosing software under particular licenses, and Troll Tech marketing is taking full advantage of that. That's why one needs to talk about this, over and over again until people get it. Qt is bad for open source.
I imagine a lot more software would be GPL'd, since a lot more developers wouldn't want to pay those fees and would therefore avoid them by GPL'ing their software.
Well, apparently the people who created the GPL don't believe so, since they have chosen a license that permits commercial developers to link to many of their libraries for free. They have done so because they believe such licenses to be in the best interest of open source.
In contrast, Troll Tech has chosen the GPL as their second license because it's in their best corporate interest, not the interest of open source. Troll Tech chose a nuisance library that effectively prohibits forking to keep control of Qt while getting the open source community to do the marketing for them. I think open source is getting a bad deal.
For an OSS product. We seem to get a lot of those around here
Qt is not an "OSS product"; it's a commercial product that happens also to be released under an open source license. That's a big difference.
For example, it's not possible to fork Qt and compete with the original. That means, among other things, that Troll Tech has full control over where Qt is going and there's not a damned thing anybody can do about that. Furthermore, unlike OSS projects, licensees of Qt end up having to pay for contributions that third parties make to Qt.
Trolltech's modell is an excellent example of how you can make money on free software.
Imagine every piece of software on Linux were to adopt Troll Tech's model: $2k/developer for the commercial exception for the C library, another $2k/developer for the math library, another $1k/developer for XML libraries, another $4k/developer for the kernel. Where do you think Linux would be today?
Both GNU and Linux have deliberately adopted licensing policies that permit commercial usage. Troll Tech is unfairly taking advantage of the situation. If Troll Tech's model caught on, Linux would be in deep trouble. And, in fact, on PDAs and cell phones, Troll Tech is probably one of the major reasons why Linux hasn't caught on that much there.
Yeah, so you're right: Troll Tech is an excellent example of how you can make money on free software--by unfairly taking advantage of the free commercial usage exceptions of other open source contributors.
(A further problem with Qt is that it's not an open source project, it's a proprietary project that happens to be released under an open source license.)
People like to say how wonderful Qt is, but you should think carefully about what you're getting into.
Almost all of our projects are open source, but occasionally will do some custom commercial stuff. Yet, because the commercial version of Qt has a per developer license, we'd end up paying as much for it as if we did all closed-source development, since it would be impractical to divide our developers that way. And Qt isn't cheap: a couple of thousand dollars per developer. Think carefully about what it would cost you if you introduce Qt and start using it.
An additional problem with it is that it (gratuitously) uses non-standard C++ extensions. That causes additional development headaches. And the Qt/Embedded version is not even fully compatible with the desktop version, and it's an all-or-nothing proposition (forget about using other toolkits on Qt/Embedded devices).
Overall, I fail to see the point of Qt for most people. For cross-platform development needs, between Java, J2ME, and wxWidgets, I think all the bases are covered at lower licensing costs and (in the case of Java and J2ME) lower development and maintenance costs and better platform coverage.
Slashdot Mirror
So, what if there are no other ways of doing something because the patent has been wisely written not to stray too far into implementation? It could preclude even the existence of your FOSS program. Period.
Patents are on specific functionality, and patents are published and well-known. For every major piece of FOSS, people have a good idea of what the patent-related risks are.
For most pieces of FOSS, any patents that could possibly read on it apply to small, optional features.
There are a few FOSS packages where the primary functionality might infringe on some patent (e.g., panorama software); those tend to be highly specialized and both the authors and users are aware of the risk and have taken it into account. In fact, exactly the same case occurs frequently for commercial software.
Statements about things that "might" and "could" happen are just FUD. Patents don't strike out of the blue, and the kind of generalities in which people like you talk are designed to hurt FOSS.
That would be "so far, the EU has imposed fines of 497 MILLION and 280 MILLION onto Microsoft". Of course, it's still spare change for Microsoft.
Shipping and supporting an OS is a lot of work and requires a lot of people; why should Google bother?
Ubuntu is a nearly ideal OS for them, and anything they want to accomplish with a desktop OS, they can by contributing to, and supporting, Ubuntu.
I'm sure there's a ton of patents (valid or not) winding their way through the patent process for the Office 2007 and Windows Vista interfaces.
Well, both tabbed toolbars ("ribbons") and galleries have plenty of prior art, so if those patents actually matter, they will be invalidated. Except for the specific visual style, OpenOffice, in fact, already has the functionality in several places.
Getting rid of the main menu bar is such an incredibly stupid idea that it's hard to understand what Microsoft could possibly be thinking. The menu bar gives you an enormous amount of bang-for-the-buck in terms of user interface.
Microsoft has been investing Billions in research and design for years, the new Office UI is simply an extension of that.
If the Office 2007 the result of spending "billions in research and design", Microsoft's R&D isn't paying off. The Office UI is not innovative; those ideas have been around for a while, and many applications already use them.
Microsoft will always spend Billions on creating slicker and easier to use interfaces.
I think it's highly debatable whether the new Office UI is "easier to use", and Microsoft has provided zero evidence that it is.
Office 2007 is a great opportunity for FOSS to make real advances in UI design, after Microsoft dropped the ball so badly with their next generation products.
I doubt Java would meet anybody's desktop quality standards.
KDE has the problem that commercial software development for it is far more expensive than for Gnome, Windows, or Macintosh. People may like KDE, but when push comes to shove for big companies like Sun, IBM, or Google to support a desktop platform, KDE is a non-starter.
So, I think if Google were to pick a desktop, it would be Gnome or a modified version of it.
In support of the nocturnal mammal evolution theory, our vision system has only three colors whereas most non-mammals have four.
Actually, current thinking is that somewhere along the line, our ancestors lost all color vision (or were down to two colors), and humans managed to re-evolve color vision.
... and the intellectual property law that backs this demand is what exactly?
Other way around, I'd say.
That's the common view, and it's wrong
However, if your software uses some patented algorithm (yes, I know, that's a hideous turn of phrase) then odds are you'd never risk releasing the source, so you could probably infringe forever and nobody would ever know.
If the algorithm is important and difficult to work around, then the patent holder will know even if you ship just binaries, and they can and will compel you to produce source code. In addition, your behavior will likely be interpreted as willful infringement, exposing you to extra damages.
That is really the greatest threat to open source software from software patents: the fact that it is substantially easier to determine if an open source package is infringing. In a litigious environment, it's easy to say, "why take the risk?"
That's pure FUD (do you work for Microsoft?). Open source has been around for several decades, and I'm not aware of any serious consequences for end users from patent infringement by FOSS. First of all, for the very reasons you mention--people know they are being scrutinized--patent infringement by FOSS is rare, and when it does, people simply remove the offending code.
You're far more at risk with closed source software--infringement seems to be far more frequent, lawsuits happen often and with serious consequences, and whether you as the customer are directly liable for infringement or not, you will often still face substantial costs if your vendor is found guilty.
So far as DOJ scrutiny is concerned, does anyone know if the DOJ has ever charged a large corporation with an antitrust violation for using a patent portfolio to suppress competition, given that that is the intended function of patents?
There is ample precedent for the government interfering in how companies license patents. But what's at issue here is not the exclusionary nature of patents in general, it's the inequitable way in which it is being used: companies who cross-license the entire portfolio have no costs, while newcomers to the market may not be able to enter at all.
Patent portfilios for the Microsoft/IBM/Oracle/Sun/HP crowd (or Intel/AMD/nVidia/ATI for that matter) have become exactly the same kind of "Mutually Assured Desctruction" scenario.
Actually, that's not all they are: they are also barriers to entry, because small, commercial, closed-source competitors find it hard to enter a market in this situation. That's not what the patent system was supposed to do. And, sooner or later, it may lead to some serious scrutiny by the DOJ.
Nevertheless, it may work to the advantage of open source, since it means that new software companies may find it advantageous to figure out open source models for software that they would otherwise have released under a proprietary license.
MySQL is absolutely trivial to set up; on most Linux distributions, setting it up is a single command to install the package, create the databases, install the service, and start it up.
There are standard GUIs for administering MySQL, probably for people like you, but most people I know prefer the command line. The command line is easier even for non-DBAs, because it's far easier to write down and document what needs to be done.
An experienced DBA can set up a new installation [of Microsoft SQL Server] in a couple of minutes
Probably true. But with MySQL, you don't even have to be a DBA at all, and it's probably even quicker.
I think it's good to equip police with cameras. It deters "offensive behavior" not only against the police, but by the police. However, what needs to be in place is a mandatory requirement that these cameras be used, that citizens have the same rights as the government to have them used in court, and that they be tamper-proof and cryptographically secured.
Hopefully, even if the government that's putting in the cameras isn't thinking of doing this, courts will sooner or later force the government to comply.
um , they do not encourage people to release stuf under LGPL, quite the opposite.
Quite right: they do not "encourage" people to release stuff under the LGPL, but they do use the LGPL or GPL w/linking exception when it serves the purposes of open source software.
The GNU C library, as well as Gtk continue to be released under licenses that permit commercial developers to deliver applications for free on Linux. The FSF and lots of other people believe that that's important.
Troll Tech deliberately makes it costly for commercial developers to develop for desktop and embedded Linux because that's where they make their profit, but in doing so, they make Linux uncompetitive for commercial developers. Fortunately, in the case of desktop Linux, we have a good alternative. For embedded Linux, Troll Tech's licensing structure has probably doomed the platform for now until the open source world can establish a good, free alternative standard.
You and Troll Tech are trying to confuse the issue by talking about generalities and association: yes, the GPL is a FSF license and the FSF generally discourages non-GPL licenses. But each license needs to be considered on its own merits, and Troll Tech's licensing structure is harmful to the goals of free software.
If Microsoft had anything substantial and usable, they'd have sued by now, but they probably have figured out that that's pointless. Many of their patents are probably invalid or unenforceable, or even have prior art in open source software.
Furthermore, FOSS developers try hard to avoid infringing on people's patents, and Microsoft's patents are scrutinized, so the number of infringing software packages is likely small. In the few cases where Microsoft might have a valid patent claim against a piece of FOSS and could actually identify someone to sue, it would be hard for them to be able to claim willful infringement or get any real damages, and the infringing code would be removed instantaneously, making the case fall apart.
If Microsoft actually believes they have IP that's being violated, they should stop bluffing and start asserting it in court. That way, they can get what they deserve, and they create certainty for everybody else. Of course, certainty is the last thing they want.
I've never heard of Christians beheading people- though this response does seem mild in comparison to what other Islamic sects do.
I think the traditional Christian methods are burning at the stake, drowning, and stoning. They were practiced widely until the rise of secular society in the West finally put a lid on the most egregious religious excesses (fairly recently, historically speaking).
The stated goal of Islam is to convert the entire world, by the sword if necessary.
So is, unfortunately, the stated goal of Christianity. Judaism doesn't try to convert much, it simply preaches separateness and superiority. No matter which way you look at it, the Abrahamic religions are pretty messed up.
The issue isn't whether you have a problem with paying for it, the issue is that people should understand the consequences of their choice and make an informed decision. The KDE/QPL license problem itself is a good example that people don't. I'm not convinced you do, but that's your problem.
Funny, I seem to remember fusion researchers from Livermore in the 70s say that commercial power was 20 years away...
It was, for the longest time. This century, it will be 35 years away for the rest of the century.
Good, you found the right quote. Now, you seem to be implying that people think that Qt offers "a significant unique capability" and should therefore be released GPL, not LGPL.
Well, I disagree; toolkits for C++ are a dime a dozen. But, more importantly, the FSF disagrees, since they have put their own toolkit under the LGPL.
Qt's licensing costs are way out of line with what people pay for Mac and Windows development. If Qt were the only toolkit for Linux, Linux development would cost a premium. That doesn't bother the kind of premium customers Troll Tech likes, but it would have killed desktop Linux if it had been the only choice. Fortunately, on the desktop, there are alternatives, in part precisely because the FOSS community did not want to hobble the Linux desktop with that sort of burden. However, for phones, there is no alternative, and Qt/Embedded's licensing structure is probably in part responsible for the failure of Linux for phones and PDAs.
When dealing with the OSS community, TrollTech could easily have kept with their original QPL license, or even gone fully closed-source
Good! I think that would have been the right thing to do.
So you blame Troll Tech for wanting to run a viable company that makes money -- you think they should be writing top-quality code for the good of the Linux community alone
I don't "blame" them, I'm saying people shouldn't use their software. I'm saying that because people probably don't understand the costs and implications involved, and because it hurts the open source community.
The above makes me question your understanding of the GPL... or maybe you just mistyped something above.
I meant exactly what I said: the creators of the GPL chose, for many libraries, the LGPL and the GPL-with-linking-exception in order to enable commercial developers to develop freely for open source platforms. Troll Tech deliberately chose the GPL precisely because they didn't have that goal.
Open source is not a religion, it's just a tool like anything else.
Quite right. And I'm saying: Qt is a bad deal for both commercial developers and for open source, for simple, rational business and policy reasons. It's a bad deal for commercial developers because of hidden costs. It's a bad deal for open source because it discourages commercial developer from moving to open source platforms because Qt is so expensive compared to Windows and Macintosh.
Seems to me they are getting a top-quality GUI library
It only seems that; but that's a separate debate.
People permit commercial usage of Linux and many FOSS libraries in the hope that others create useful commercial applications and thereby actually help open source platforms and APIs succeed.
But in the phone market, Troll Tech has managed to insert itself into Linux in such a way that they are the only ones reaping financial benefit. And Troll Tech designed their system so that other toolkits would be difficult to port to their platform, and they made license agreements with manufacturers that ensured that users wouldn't have a choice. And they screwed up so badly with their platform that they probably put off lots of manufacturers from it.
Of course, they can legally do all that--the license permits it.
But people like me can and should also point out how harmful their behavior is for open source and users in the long run, just like people can and should point out how Microsoft's licensing model is a bad deal for users.
You are welcome to accept their terms, or reject them -- nobody is forcing you to develop with Qt if you don't like their license.
Indeed, I don't develop with Qt. And I am trying to convince others not to develop with Qt either because I think it's a lousy deal for the open source community. I'm saying that if more people behaved like Troll Tech did, open source would die. In fact, I think Troll Tech is largely responsible for the failure of Linux on phones and PDAs.
Now, what's your problem with me stating my opinion and analysis? In your world, should everybody other than corporate marketing just shut up?
Troll Tech isn't unfairly taking advantage of anything.
Oh, yes, they are. I mean, look at what happened with the KDE project: originally, the started using Qt despite the fact that their license was incompatible with the QPL. People don't understand licenses well, and people don't understand the consequences of choosing software under particular licenses, and Troll Tech marketing is taking full advantage of that. That's why one needs to talk about this, over and over again until people get it. Qt is bad for open source.
I imagine a lot more software would be GPL'd, since a lot more developers wouldn't want to pay those fees and would therefore avoid them by GPL'ing their software.
Well, apparently the people who created the GPL don't believe so, since they have chosen a license that permits commercial developers to link to many of their libraries for free. They have done so because they believe such licenses to be in the best interest of open source.
In contrast, Troll Tech has chosen the GPL as their second license because it's in their best corporate interest, not the interest of open source. Troll Tech chose a nuisance library that effectively prohibits forking to keep control of Qt while getting the open source community to do the marketing for them. I think open source is getting a bad deal.
For an OSS product. We seem to get a lot of those around here
Qt is not an "OSS product"; it's a commercial product that happens also to be released under an open source license. That's a big difference.
For example, it's not possible to fork Qt and compete with the original. That means, among other things, that Troll Tech has full control over where Qt is going and there's not a damned thing anybody can do about that. Furthermore, unlike OSS projects, licensees of Qt end up having to pay for contributions that third parties make to Qt.
Trolltech's modell is an excellent example of how you can make money on free software.
Imagine every piece of software on Linux were to adopt Troll Tech's model: $2k/developer for the commercial exception for the C library, another $2k/developer for the math library, another $1k/developer for XML libraries, another $4k/developer for the kernel. Where do you think Linux would be today?
Both GNU and Linux have deliberately adopted licensing policies that permit commercial usage. Troll Tech is unfairly taking advantage of the situation. If Troll Tech's model caught on, Linux would be in deep trouble. And, in fact, on PDAs and cell phones, Troll Tech is probably one of the major reasons why Linux hasn't caught on that much there.
Yeah, so you're right: Troll Tech is an excellent example of how you can make money on free software--by unfairly taking advantage of the free commercial usage exceptions of other open source contributors.
(A further problem with Qt is that it's not an open source project, it's a proprietary project that happens to be released under an open source license.)
People like to say how wonderful Qt is, but you should think carefully about what you're getting into.
Almost all of our projects are open source, but occasionally will do some custom commercial stuff. Yet, because the commercial version of Qt has a per developer license, we'd end up paying as much for it as if we did all closed-source development, since it would be impractical to divide our developers that way. And Qt isn't cheap: a couple of thousand dollars per developer. Think carefully about what it would cost you if you introduce Qt and start using it.
An additional problem with it is that it (gratuitously) uses non-standard C++ extensions. That causes additional development headaches. And the Qt/Embedded version is not even fully compatible with the desktop version, and it's an all-or-nothing proposition (forget about using other toolkits on Qt/Embedded devices).
Overall, I fail to see the point of Qt for most people. For cross-platform development needs, between Java, J2ME, and wxWidgets, I think all the bases are covered at lower licensing costs and (in the case of Java and J2ME) lower development and maintenance costs and better platform coverage.