"Respects Your Freedom" certification?

[_wb_]

My point is that the contribution of GNU/FSF is more than just the collection of code they wrote. They also designed the GPL, pushed other projects like BSD and X to adopt FOSS licenses, and started the whole FOSS movement in the first place. While the Linux kernel is not part of the GNU project, I think it is still reasonable to say that GNU is an important part of the reason for Linux' success - if GNU software like gcc and bash would not have been available in 1991, and if the GPL would not have existed, the history of Linux would have been very different.

Maybe it was a bit confusing of me to say "we should still credit GNU". By that phrase I did not mean to say that we have to call it GNU/something (because obviously that would be an incorrect name if there is no GNU software included). I just think there are other reasons than only code contributions to credit something or someone. And of course there other ways to credit something than to include it in the name.

Edit: The point is probably moot anyway, since we probably want to use a GNU/Linux-based OS anyway, including stuff like glibc, bash and gcc. All of this discussion was about a very weird scenario in which we for some strange reason want to avoid GNU software and use other Free Software instead (e.g. from *BSD).

 

[sebt3]

So you're rewriting history now...

Opensource existed before the FSF. Neither the FSF nor RMS did start showing their source. How would RMS had the idea to start this whole mess if lexmark never provided the sources of their drivers ?

Many copyleft licenses existed before the GPL.

Telling that Linux would not have existed without the FSF is also rewriting history. At first, Linux was using Minix userland, not the GNU one (which was n ot ready at that time anyway). The reason why Linux is opensource is that it was a fun/research projet where Linus wanted to get more hands in.

While I'm at it, let me tell you that the reason why Linux kickstarted and Hurd not is that Linux was fun to hack while hacking hurd meant having to respect a tons of dogma. It's the dogma that killed hurd... and these same kind of dogma are going to kill the FSF. Believe it or not, but that's is. In a sentence : Too much dogma kill the tenet...

If having Exophase and Notaz telling you that it is way over mark not enough for you, I cant help you much.

And if Notaz is serious about kicking the FSF projects out of the potential P2 I'm in. I'll work on this OS. 'cause I'm becoming more and more pissed by these zealots that take credit from others' work. And I realy like the idea to request a RYF sticker with not a single GNU lines in the projet. Sound like a good way to reground the zealots...

 

[WizardStan]

Could we even produce a BSD based P2 that is backwards compatible with the PNDs already produced for the P1? Don't a lot of them rely on the GNU clibs? What about SDL and all those additional libraries we use?

 

[Exophase]

My point is that the contribution of GNU/FSF is more than just the collection of code they wrote. They also designed the GPL, pushed other projects like BSD and X to adopt FOSS licenses, and started the whole FOSS movement in the first place. While the Linux kernel is not part of the GNU project, I think it is still reasonable to say that GNU is an important part of the reason for Linux' success - if GNU software like gcc and bash would not have been available in 1991, and if the GPL would not have existed, the history of Linux would have been very different.

So what you're saying is, if not for the collection of code GNU wrote Linux wouldn't have taken off the ground. But it's about crediting the philosophy, not the software. It sounds like two contradictory viewpoints. Face it, GNU is that software (what they call an operating system, although IMO only a fairly small part of GNU really qualifies as operating system components), putting it in means nothing else than describing the operating system. The only reason FSF is stringent on the argument that this isn't about crediting software is because it opens up the can of worms for the other significant things no one's crediting, but that doesn't mean that these viewpoints are logically consistent.

I don't know what this bit about Linux starting out under a more restrictive license is about, Linus said in the thread where he introduced it "It will be free though (probably under gnu-license or similar)." At this point he hadn't even released anything. Given his goals I wouldn't be surprised at all if Linux was open source in some form if not exactly like GPL or really necessarily "Free" by the FSF's requirements. I'm sure he had no aspirations at pretty much any point of doing the thing entirely by himself.

Could we even produce a BSD based P2 that is backwards compatible with the PNDs already produced for the P1? Don't a lot of them rely on the GNU clibs? What about SDL and all those additional libraries we use?

I doubt there's a lot of relying on functionality that's specific to GNU's standard C library, and besides that it's an open standard so it's not like other C runtimes can't implement the same functions if they really want to (and I don't think that alone will force them to become GPL). SDL isn't part of GNU.

 

[WizardStan]

SDL isn't part of GNU.

I'm aware, I was thinking more about library requirements, going back to needing standard C libraries. That was an instinctive reaction based on the fact that I've not consciously been aware of any non-GNU standard C libraries, I build pretty much everything with some form of gcc. Obviously that's not correct and swapping the toolchain shouldn't pose any sort of problem, I would hope.

 

[Exophase]

I'm aware, I was thinking more about library requirements, going back to needing standard C libraries. That was an instinctive reaction based on the fact that I've not consciously been aware of any non-GNU standard C libraries, I build pretty much everything with some form of gcc. Obviously that's not correct and swapping the toolchain shouldn't pose any sort of problem, I would hope.

The compiler doesn't dictate the standard C library. You can build with GCC and link with whatever. Or build with something that isn't GCC and link with glibc. There are a lot of non-GNU free libcs, like BSD libc, newlib, musl, uclibc..

Note that glibc itself is actually very slightly newer than Linux.

This goes without saying, but there are also a lot of non-GNU shells, and I actually don't think that Linux would have been that different if not for bash. Linus and other early contributors would have probably written their own shell and core utilities, it's not at all hard to get enough functionality to allow minimal development and usability, and the ball would have rolled from there.

 

[WizardStan]

There are a lot of non-GNU free libcs, like BSD libc, newlib, musl, uclibc..

Exactly, but the standard gcc toolchains I've used all use the same GNU libraries, or some variation thereof. I just wasn't thinking when I posted my worry the first time.

 

[_wb_]

Many copyleft licenses existed before the GPL.

Such as...?

My point is that the contribution of GNU/FSF is more than just the collection of code they wrote. They also designed the GPL, pushed other projects like BSD and X to adopt FOSS licenses, and started the whole FOSS movement in the first place. While the Linux kernel is not part of the GNU project, I think it is still reasonable to say that GNU is an important part of the reason for Linux' success - if GNU software like gcc and bash would not have been available in 1991, and if the GPL would not have existed, the history of Linux would have been very different.

So what you're saying is, if not for the collection of code GNU wrote Linux wouldn't have taken off the ground. But it's about crediting the philosophy, not the software. It sounds like two contradictory viewpoints. Face it, GNU is that software (what they call an operating system, although IMO only a fairly small part of GNU really qualifies as operating system components), putting it in means nothing else than describing the operating system. The only reason FSF is stringent on the argument that this isn't about crediting software is because it opens up the can of worms for the other significant things no one's crediting, but that doesn't mean that these viewpoints are logically consistent.

Those are not contradictory viewpoints, just two different reasons to give credit to GNU.

I don't know what this bit about Linux starting out under a more restrictive license is about, Linus said in the thread where he introduced it "It will be free though (probably under gnu-license or similar)." At this point he hadn't even released anything. Given his goals I wouldn't be surprised at all if Linux was open source in some form if not exactly like GPL or really necessarily "Free" by the FSF's requirements. I'm sure he had no aspirations at pretty much any point of doing the thing entirely by himself.

The original license said


You may not distibute this for a fee, not even "handling" costs.

(source: https://www.kernel.org/pub/linux/kernel/Historic/old-versions/RELNOTES-0.01)

This would for example not allow selling distributions on CD-ROM.

 

[Exophase]

Okay, I didn't see he originally had that in there. But Linux moving to GPL isn't strong evidence that that restriction wouldn't have been dropped. He couldn't have felt that strongly about it, or was taking into consideration the wishes of the other people who were contributing to it.

Personally I'd love to have seen Linux take a stand on the non-commercial thing

 

[comradekingu]

What was left of any feature parity was smashed once big business started contributing more code than indie devs. And in this alternate reality maybe the devs that are hired to work on linux would have worked on it anyhow, or that the total number of clever people are funneled to that cause, but alas, it is what it is.

So the FSF actually has a hardware license that applies to BSD and others. Which are for the case of argument systems they have little control over. Whats so evil about this? If you dont want to agree with their software claims somehow a point is being made out of applying to their freedom guidelines for hardware?

Thats the very level of support we are discussing here, at a hardware level, you need support for free drivers, otherwise no BSD, haiku etc. And for whatever reason linux is big, FSF, GPL, big business, first to the table, it doesn't take away from the point in question, free drivers are good, and RYF is a good license, since its a hardware license and its good for all software.

Not everything FSF does is good, and i think i made a list of bad things at page 5, their marketing department is beyond bad for example. However, they arent going anywhere, so lets support their good initiatives, like this. We use enough GPL stuff to atleast show some gratitude for how that is a working license that produces good software. That is nice regardless of whos merit it is.

How about we just not write linux anywhere on the package and instead print the logo? If thats all it takes for someone to be able to run their system of choice, good. Its semantics, the common goal here is that we all want good free drivers for our linux kernel systems, regardless of what we call it.

 

[WizardStan]

How about we just not write linux anywhere on the package and instead print the logo?

I love how this discussion has evolved from arguing about what consequences each requirement may or may not have into discussions about how to use loopholes in the wording to completely circumvent the requirements and still be compliant.

 

[HelenF]

I also hate all that enforcing of some ideals of what's "good" and what's "bad", like why I can't call something like MAME free, it's not just gratis, it's open, hell, it's even under copyleft license. But no, for GNU, no GPL, not free software.

For the FSF, something can still be "free software" without being under the GPL, or even compatible with the GPL.

MAME doesn't offer "The freedom to run the program, for any purpose."

http://www.gnu.org/philosophy/free-sw.html

I think MAME's non-commercial license is unfortunate. Many of the consequences are listed on the legal FAQ. And since MAME was written by hundreds of developers who each hold the copyright to their own code, the license can never be changed even if almost everyone involved later regrets it.

http://www.mamedev.org/legal.html

 

[Exophase]

I've always found the claim that "the freedom to run the program, for any purpose" implies "the freedom to charge any amount of money for the program" completely off the wall. I can't even begin to understand the logical rationalization of this. Stallman wanted to make a nice neat set of 4 axioms so he could make it sound really universal and self-evident but botched it on the commercialization part. Should have just added a fifth thing.

The people who contributed to MAME knew the license when they got involved, if they all suddenly regret it later that's on them. It's no different than if it were a bunch of people working on a GPL program who later decided that they regret that.

 

[comradekingu]

So how is that not covered? The GPL is up to version 3, it really needs to be correct, legal, and universial. Commercial is tied to a monetary system. What if in the future we start offering to the gods, software, trying to rid the world of technology, in return for cows. GPL still valid. MAME license wins in this scenario though, but its restrictive!

How about we just not write linux anywhere on the package and instead print the logo?

I love how this discussion has evolved from arguing about what consequences each requirement may or may not have into discussions about how to use loopholes in the wording to completely circumvent the requirements and still be compliant.

Hack the hackers!

 

[Exophase]

I didn't say it wasn't covered in the GPL, it very clearly is covered there. I said that I don't see at all how it's implied by any of the "four freedoms" despite the many arguments I've seen that claim otherwise. As far as I understand it's supposed to be the four freedoms that the FSF uses to determine whether or not a license is "free."

Of course MAME license is restrictive, GPL is restrictive too, isn't that the underlying point? Earlier in this thread it was said that GPL restricts the rights of developers so it can respect the rights of users. I find it hard to really get behind that, using any kind of definition for user that makes sense in the real world.. I'd rather say that it retricts the rights of the original developers so it can respect the rights of future developers, something that the commercialization requirements definitely fall under. Unless we're talking about programs which somehow make the users money, which is something very different.

 

[HelenF]

I've always found the claim that "the freedom to run the program, for any purpose" implies "the freedom to charge any amount of money for the program" completely off the wall. I can't even begin to understand the logical rationalization of this. Stallman wanted to make a nice neat set of 4 axioms so he could make it sound really universal and self-evident but botched it on the commercialization part. Should have just added a fifth thing.

The GPL mostly protects against unfair commercial exploitation, while still allowing people to make money from working on free software (including building on other people's work that came before).
In an alternate universe where the main free software ecosystem was non-commercial, all development which builds on other people's work would have to be done by hobbyists in their spare time.

And if commercial hardware products and hosting services couldn't use a free software OS, we would all be stuck with products and services with a proprietary OS instead. We wouldn't even have the Pandora.

The people who contributed to MAME knew the license when they got involved, if they all suddenly regret it later that's on them. It's no different than if it were a bunch of people working on a GPL program who later decided that they regret that.

That's true. Either way, the most likely reason for large numbers of people to regret it later is the changes in licensing-related cultural norms over the course of many years; and who knows what's yet to happen in the future.

 

[Exophase]

The GPL mostly protects against unfair commercial exploitation, while still allowing people to make money from working on free software (including building on other people's work that came before).

From what I can tell the vast majority of money made by GPLed software is done on ports to platforms where the users don't really understand that the software is a port of something GPL and where there's a big associated value with the branding and public reception of being first - such that releasing another redistribution of the same software under a different name for a different price doesn't actually stop the first person from making money. These people tend to contribute little if anything back to the pre-existing versions of the program.

There are a few cases where GPLed programs are sold by the original developers but they're rare.

In an alternate universe where the main free software ecosystem was non-commercial, all development which builds on other people's work would have to be done by hobbyists in their spare time.

I disagree - today much of the work on open source software is done by those employed by corporations, but those corporations are not directly selling the software but indirectly benefiting from it in some way. I understand that this probably isn't enabled by the MAME license, but there's a continuum of non-commercial licenses that are not accepted by FSF, but would allow for distributing software freely as part of some other commercial product. Even MAME development could work this way - while you can't include MAME on a product you sell improving MAME could still improve your product's viability. But it's hard to draw a lot specifically with MAME, because it's a niche program with a niche license catered to it.

I think I have to make something clearer here. I don't have a problem with licenses like GPL that require allowing commercial distribution. So I'm really not arguing against people using the license. What I have a problem with is the notion that disallowing commercial distribution is universally unethical, and I don't think it's mincing words to say that that's the FSF's position. I believe that the software developer has the prerogative to license the work based on his or her personal requirements and circumstances, regardless of how much or how little this provides to the betterment of society. And it's userbase's prerogative to ignore software that they don't like for whatever reason, including license.

And if commercial hardware products and hosting services couldn't use a free software OS, we would all be stuck with products and services with a proprietary OS instead. We wouldn't even have the Pandora.

Software serves different roles and attracts different audiences. A license that makes sense for the developers of MAME might not make sense to the developers of an operating system or a library. Sometimes the very same developers will work on GPL, LGPL, BSD, etc projects at the same time. As far as the developers of MAME were concerned, they likely reasoned that the biggest commercial incentive for the emulator was to be included in game packs by game companies. This has been done with other emulators in the past, and I guess they don't like it. With something like Pandora you don't need to load it with MAME, users can easily put MAME on it themselves.

 

[HelenF]

You were saying "Personally I'd love to have seen Linux take a stand on the non-commercial thing". If that happened, it would have less work done on it, and it couldn't be used for some of the things we enjoy now. Maybe we'd be using a proprietary kernel for those things; or maybe Hurd would have gone further if Linux hadn't displaced it.

Most of the money in GPL'd software is made indirectly rather than by selling the software. Usually a non-commercial license means not allowing the product to be used to make money in any way. Though there could be licenses which only disallow particular commercial uses e.g. selling the software.

What I have a problem with is the notion that disallowing commercial distribution is universally unethical, and I don't think it's mincing words to say that that's the FSF's position. I believe that the software developer has the prerogative to license the work based on his or her personal requirements and circumstances, regardless of how much or how little this provides to the betterment of society. And it's userbase's prerogative to ignore software that they don't like for whatever reason, including license.

Fair enough.
The userbase can still campaign for a particular license in future projects. And can debate about which license most benefits society (which might be a factor in some people's choice).

 

[Exophase]

You were saying "Personally I'd love to have seen Linux take a stand on the non-commercial thing". If that happened, it would have less work done on it, and it couldn't be used for some of the things we enjoy now. Maybe we'd be using a proprietary kernel for those things; or maybe Hurd would have gone further if Linux hadn't displaced it.

Most of the money in GPL'd software is made indirectly rather than by selling the software. Usually a non-commercial license means not allowing the product to be used to make money in any way. Though there could be licenses which only disallow particular commercial uses e.g. selling the software.

Okay, I feel like I already touched on this but let's be clear here - we're talking about the clause Linus put in originally which is common in a lot of non-commercial licenses. That's the clause which says that you can't distribute it for a fee. My interpretation of this is that you can't ask for money solely in exchange for receiving the software. That doesn't mean you couldn't bundle Linux as a "freebie" as MAME puts it, or an included part of a broader system that you charge money for. I think that the vast majority of commercial involvement in Linux wouldn't have been impacted by this. There isn't a one-size fits all description for what "non-commercial" means as a license stipulation.

You can look at it this way - the spirit of a lot of non-commercial licenses is that they don't want you to take their work and sell it while contributing zero or close to zero work on your own. That's a little different from taking your work and selling it with their work. In most people's minds there's a pretty big division between where someone's software ends and where your hardware begins - or where your other software begins. Having a non-commercial license interact with commercial products sold by you would be analogous to distributing a software package that contains both GPLed software and proprietary software. While FSF may frown on this it doesn't actually violate the GPL, much less does distributing hardware that contains GPLed software but a proprietary hardware design.

While being non-commercial in this form would have prevented Linux distros from showing up in stores I honestly don't think this ever did a lot to spur its popularity. Maybe if the "handling fee" part would have been dropped and some standard caps were set in place for exactly how much you can charge with handling. I have a really good feeling that's what would have happened if the Linux developers were actually opposed to commercial gain from selling the kernel directly.

Still, my comment was really pretty tongue in cheek. I don't have a personal problem with Linux being GPL and I don't think that got in its way, I just think that the controversy between Linux and GPL would have resulted in some very interesting and very public arguments. Far beyond what we've seen, despite Linus and rms already butting heads on several occasions.

 

[comradekingu]

How about we turn the argument and say that rebranding/forking things for profit isnt a big problem?

Its the users that decide, and they seem to do just fine. No need for complex governing structures when its all about who sends money to what.

Only money i ever regret funding into free software was cragix software developer fund...

Linus said gplv2 for the kernel is his best decision, and there is even an added clause that makes it gplv2 only. But we are fast approaching theoretical equilibrium.

As for RYF its very good prospects for that on ARM atm, minus imagination technologies PowerVR stuff. On Mips i dont really know, the vivante combo seems to work for gwx-zero. I have only seen Imagination get behind mips lately, they seem very eager to push it, along with their IP GPU driver/hardware.

So before we start the x86 silliness, things are looking good on ARM, except for powerVR.

PowerVR is a high priority FSF project, but it seems there is nothing happening or to throw money at.