Uhh, no. If you were a bit more aware of the subject, you would know that there are "GPL-compatible" licenses Uh, no. The GPL explicitly requires the entire program to be covered by the GPL. So there is simply no license "compatible with GPL" in this regard (apart perhaps from the LGPL, which I find almost ironical).
where you can link code between them and the clauses of each license permit so without trouble. Yes, all open source licenses allow you to do that, except the GPL.
I'm actually quite suprised how little most people know on the subject. You're right, and, no offence, but you are a good example of that.
If you want a license that frees your code while letting others link proprietary code to it, you have the LGPL, which allows just that, and is IMO the best license for core libraries such as GTK a libc.
Sounds to me like you people should take a look at the LGPL. The FSF and Stallman realized that some of the freedoms of the GPL might translate into unreasonable restrictions to others, and that's why they created that license. Nobody has talked about LGPL. We (including the grand parent poster) have talked about the GPL. So please try to stay on-topic. The topic was GPL vs. BSD.
Licenses only do things that they actually say, not things that someone hopes they say.
Not always. Judges should and often do attempt to interpret and infer the original intents of the author of the license. Even laws are sometimes vague, broad, or even ambiguous, and in such cases they cannot be taken literally. They need to be sensibly interpreted.
Let's try a different approach, so that everyone understands:
Simply, if we remove from the GPL the requirement that the whole program must be under the GPL, and if we add a new requirement that other licenses covering other parts of the program must require complete source code to be open, will we have more freedom? Yes, we definitely and obviously will.
The "if-you-are-not-with-us-you-are-against-us" attitude reminds me of the communists. The GPL insists that it should be the only license in the world...
Need an example? If I create software under the GPL I cannot take any code under other open source licenses (MPL, BSD, Apache, etc.) and integrate it as part of my GPLed program. Why? Because the GPL requires that all parts of my software are released under the GPL! See?
The other licenses allow me to combine open source code covered by different licenses in one single product. That's why they give us more freedom than the GPL.
At first I thought this was a scam or a joke. But then I noticed the source at the bottom of the article. And it is one of the most credible news sources there is: BBC.
It would have been better if the article linked to that, rather than to some, at least to me, rather obscure blogger.
The two camps that seem to have the most concern about too many licenses are the FUD-spinners trying to damage OSS or the Free-bies that are trying to steer everyone towards GPL 3 and FSF hegemony. (Yes, I'm a bit biased.)
Good grief your right, I looked to the left and the right and found I was in the exact center of the Hubble Volume [wikipedia.org]!
What a misleading comment and link. The Wikipedia article redirects to "Observable universe". I've never talked about "Observable universe", but about the entire matter that exists in this sapce. And that (not "Observable universe") is currently believed to be finite (space in which the matter exists may be infinite though).
"Observable universe" is obviously finite and only an idiot would comment on that fact.
It could be finite, but unbounded. Think of the surface of a sphere with 2D beings living on it.
Space like that (a la the PacMan game) would violate quite a lot of laws in physics. Two objects going in opposite directions (i.e. not on a collision course) would eventually collide. That clearly doesn't seem to be the case in this space, does it?
I see that old bullshit repeated over and over. If the matter in space is finite, and AFAIK it is believed by scientists to be finite, then it has to have some overall shape. Therefore, there must be a center of the matter.
Uh, you do realize that if you run an OS with the hard required MINIMUM of RAM, entire free RAM is just about a few MBs? And that all memory for apps is swapped to a disk?
Vista's MINIMUM memory requirement is 512 MB. That means it will run, but slow like hell due to swapping. Minimum is minimum. If you want something good, go for the RECOMENDED figure.
You know what? Some people never learn. So I suggest you continue playing with a legacy-design OS from the 60's, which makes it impossible to develop third-party drivers that don't stop working after each minor update of the kernel.
I will continue using binaries of excellent open-source and closed source drivers written for Windows NT4... on Windows XP and Vista.
This was my last post to you. If you post anything in response I won't read it, so don't bother. You're a stupid moron and I dislike communicating with such people.
I'm glad you acknowledged that a stable API is needed and that there are problems with the Linux model. What you call "out-of-tree" we in the real world call just regular "software". You know the world where anyone can develop and release any software under any license they wish (not just under the GPL as part of a bloated monolithic kernel that includes every driver ever written).
Hitachi et al want a stable driver API so they can write binary drivers. Linux doesn't want or need binary drivers.
It's not only closed source drivers. But also open source THIRD PARTY EXTERNAL drivers, such as TrueCrypt. Every MINOR update of the Linux kernel caused the TrueCrypt driver to stop working and it required that the developers REWROTE parts of the driver! Do you get it now? Why a stable API for drivers is needed?
Now go play with your toy designed in the 1960's for mainframes. And don't bother replying -- I don't read messages from religious zealots.
It's not necessary given Linux's development model
Really? Then ask yourself why Hitachi et al asked for it, and why the grand parent poster wrote "Linux is slowly becoming a microkernel" (certainly because monolithic kernel is great, eh?)
Ok, you're yet another ignorant fool and a blind zealot who believes that a legacy system designed in the 60's is suitable for desktop in the year 2007.
Funny example, that - Slashdot probably has one of the highest ratios of users capable of ad-blocking of any site on the entire web, yet manages to pay the bills
As their bandwidth must cost millions, if everyone blocked their ads, they (Slashdot, Sourceforge, etc.) would have the following options:
The "device driver" written "for" Linux is supposed to be integrated _into_ the main Linux source tree. If they didn't do that, it's their own fault and they deserve breakage.
There is no stable (i.e. standardized) external driver API in the Linux kernel, so any further discussion in that direction is pointless and makes you look like a fool.
A device driver written for Windows NT4 works on Windows Vista without any modifications, whereas a device driver written for Linux kernel 2.6.18 doesn't work on (nor compile against) kernel 2.6.19.
You missed the point completely. I also wonder why you posted as AC... anyway back to the point.
Where as, when I combine this piece of public domained code "Activation Crack.exe" with Microsoft Windows XP, the XP license does not "infect" the combined result, and I can distribute it as I feel like?
A Microsoft Windows license and the GPL don't allow you to combine your work with their work and distribute the result and two licenses (theirs and yours.) However, (and that's the main point) most open source licenses DO allow you to do exactly that. For example, BSD, Apache, Mozilla, and other licenses.
Wrong. Just like the GPL doesn't disappear when you make a derived work, neither does most other licenses. This is what you call "viral".
Wrong. What I call viral is the fact the GPL requires that ALL portions of a collective work must be put under a single license. Most open source licenses don't require that that (for example, BSD, Apache, Mozilla, and other licenses).
Any code that gets close to a GPLed code (combined with it in a single product) must be put under GPL. That's how the GPL attaches to non-GPL code. See? It does spread like a virus. And if you don't want or cannot put your code portions under GPL, what can you do? The thing you do when... Avoid the GPLed code like,... like a... contagious illness (e.g. flu).
There's no doubt. It is a viral license and it is intentionally viral. Are you saying that the world should try to cover up or make that FACT* less obvious?
* If anyone of you believes that GPL is not viral, you should perhaps finally find some time to actually read it.
Slashdot Mirror
Sounds to me like you people should take a look at the LGPL. The FSF and Stallman realized that some of the freedoms of the GPL might translate into unreasonable restrictions to others, and that's why they created that license. Nobody has talked about LGPL. We (including the grand parent poster) have talked about the GPL. So please try to stay on-topic. The topic was GPL vs. BSD.
Licenses only do things that they actually say, not things that someone hopes they say.
Not always. Judges should and often do attempt to interpret and infer the original intents of the author of the license. Even laws are sometimes vague, broad, or even ambiguous, and in such cases they cannot be taken literally. They need to be sensibly interpreted.
Let's try a different approach, so that everyone understands:
Simply, if we remove from the GPL the requirement that the whole program must be under the GPL, and if we add a new requirement that other licenses covering other parts of the program must require complete source code to be open, will we have more freedom? Yes, we definitely and obviously will.
The "if-you-are-not-with-us-you-are-against-us" attitude reminds me of the communists. The GPL insists that it should be the only license in the world...
it is no less free than the BSD
GPL certainly is less free than BSD.
Need an example? If I create software under the GPL I cannot take any code under other open source licenses (MPL, BSD, Apache, etc.) and integrate it as part of my GPLed program. Why? Because the GPL requires that all parts of my software are released under the GPL! See?
The other licenses allow me to combine open source code covered by different licenses in one single product. That's why they give us more freedom than the GPL.
At first I thought this was a scam or a joke. But then I noticed the source at the bottom of the article. And it is one of the most credible news sources there is: BBC.
It would have been better if the article linked to that, rather than to some, at least to me, rather obscure blogger.
http://news.bbc.co.uk/2/hi/africa/6959209.stm
NASA's Open Source license is based on the GPL, for example
Sorry, but that is an absolutely false statement. They are two completely different licenses and it's certainly not based on the GPL.
The two camps that seem to have the most concern about too many licenses are the FUD-spinners trying to damage OSS or the Free-bies that are trying to steer everyone towards GPL 3 and FSF hegemony. (Yes, I'm a bit biased.)
If I had mod points, I'd mod you up.
Good grief your right, I looked to the left and the right and found I was in the exact center of the Hubble Volume [wikipedia.org]!
What a misleading comment and link. The Wikipedia article redirects to "Observable universe". I've never talked about "Observable universe", but about the entire matter that exists in this sapce. And that (not "Observable universe") is currently believed to be finite (space in which the matter exists may be infinite though).
"Observable universe" is obviously finite and only an idiot would comment on that fact.
something traveling at the speed of light couldn't have made it all the way "around" even once.
Sounds like an easy way to prevent anyone from disproving the "hypothesis".
It could be finite, but unbounded. Think of the surface of a sphere with 2D beings living on it.
Space like that (a la the PacMan game) would violate quite a lot of laws in physics. Two objects going in opposite directions (i.e. not on a collision course) would eventually collide. That clearly doesn't seem to be the case in this space, does it?
all locations in the universe are at the centre
I see that old bullshit repeated over and over. If the matter in space is finite, and AFAIK it is believed by scientists to be finite, then it has to have some overall shape. Therefore, there must be a center of the matter.
BTW, don't confuse matter in space with space.
512MB of RAM
Uh, you do realize that if you run an OS with the hard required MINIMUM of RAM, entire free RAM is just about a few MBs? And that all memory for apps is swapped to a disk?
Vista's MINIMUM memory requirement is 512 MB. That means it will run, but slow like hell due to swapping. Minimum is minimum. If you want something good, go for the RECOMENDED figure.
You know what? Some people never learn. So I suggest you continue playing with a legacy-design OS from the 60's, which makes it impossible to develop third-party drivers that don't stop working after each minor update of the kernel.
I will continue using binaries of excellent open-source and closed source drivers written for Windows NT4... on Windows XP and Vista.
This was my last post to you. If you post anything in response I won't read it, so don't bother. You're a stupid moron and I dislike communicating with such people.
Good bye.
Yup, out-of-tree development tends to do that.
I'm glad you acknowledged that a stable API is needed and that there are problems with the Linux model. What you call "out-of-tree" we in the real world call just regular "software". You know the world where anyone can develop and release any software under any license they wish (not just under the GPL as part of a bloated monolithic kernel that includes every driver ever written).
Hitachi et al want a stable driver API so they can write binary drivers. Linux doesn't want or need binary drivers.
It's not only closed source drivers. But also open source THIRD PARTY EXTERNAL drivers, such as TrueCrypt. Every MINOR update of the Linux kernel caused the TrueCrypt driver to stop working and it required that the developers REWROTE parts of the driver! Do you get it now? Why a stable API for drivers is needed?
Now go play with your toy designed in the 1960's for mainframes. And don't bother replying -- I don't read messages from religious zealots.
It's not necessary given Linux's development model
Really? Then ask yourself why Hitachi et al asked for it, and why the grand parent poster wrote "Linux is slowly becoming a microkernel" (certainly because monolithic kernel is great, eh?)
Ok, you're yet another ignorant fool and a blind zealot who believes that a legacy system designed in the 60's is suitable for desktop in the year 2007.
Funny example, that - Slashdot probably has one of the highest ratios of users capable of ad-blocking of any site on the entire web, yet manages to pay the bills
As their bandwidth must cost millions, if everyone blocked their ads, they (Slashdot, Sourceforge, etc.) would have the following options:
1) Become a pay site (get ready your credit card)
2) End
Google text ads are the answer to this.
That would explain why AdBlock blocks Google ads by default...
I've never seen a single ODF document on any website. But I've seen a lot of .docs and zillions of .pdfs.
Hello Anonymous Coward,
The "device driver" written "for" Linux is supposed to be integrated _into_ the main Linux source tree. If they didn't do that, it's their own fault and they deserve breakage.
There is no stable (i.e. standardized) external driver API in the Linux kernel, so any further discussion in that direction is pointless and makes you look like a fool.
What, like how Windows
A device driver written for Windows NT4 works on Windows Vista without any modifications, whereas a device driver written for Linux kernel 2.6.18 doesn't work on (nor compile against) kernel 2.6.19.
But, come on and look at Linux: it's slowly becoming a microkernel:
In other words, it's slowly getting towards the superior desktop OS design -- the one that Windows uses. And it's really no wonder.
You missed the point completely. I also wonder why you posted as AC... anyway back to the point.
Where as, when I combine this piece of public domained code "Activation Crack.exe" with Microsoft Windows XP, the XP license does not "infect" the combined result, and I can distribute it as I feel like?
A Microsoft Windows license and the GPL don't allow you to combine your work with their work and distribute the result and two licenses (theirs and yours.) However, (and that's the main point) most open source licenses DO allow you to do exactly that. For example, BSD, Apache, Mozilla, and other licenses.
Wrong. Just like the GPL doesn't disappear when you make a derived work, neither does most other licenses. This is what you call "viral".
Wrong. What I call viral is the fact the GPL requires that ALL portions of a collective work must be put under a single license. Most open source licenses don't require that that (for example, BSD, Apache, Mozilla, and other licenses).
So, as you see, the GPL is clearly not viral.
... like a ... contagious illness (e.g. flu).
Let me explain to you, why GPL is viral:
Any code that gets close to a GPLed code (combined with it in a single product) must be put under GPL. That's how the GPL attaches to non-GPL code. See? It does spread like a virus. And if you don't want or cannot put your code portions under GPL, what can you do? The thing you do when... Avoid the GPLed code like,
perception of the GPL as a "viral" license
There's no doubt. It is a viral license and it is intentionally viral. Are you saying that the world should try to cover up or make that FACT* less obvious?
* If anyone of you believes that GPL is not viral, you should perhaps finally find some time to actually read it.