"Respects Your Freedom" certification?

[Neelix]

Most people are fine with the person who started a project getting to name it and don't insist on renaming it once a bunch of other people have contributed to it, since that's petty and pointless.

 So who started the overall project (fully free OS, kernel+userspace+development tools) that I now prefer to call GNU/Linux, but which most people insist on calling just "Linux"? Linus Torvalds or Richard Stallman? Maybe we should call it Stallmanix after all, no?

That's cute, but I think you missed the point. The person who started it gets to pick a name, not the person who started it gets it named after them. There's a big difference there.

However, my point is that we should not call it GNU/Linux because of the weight of the GNU code contributions, but because "GNU" is a clear reference to the goals of Free Software. If you just call it "Linux", it could as well be a tivo'ized locked-down Android; if you call it "GNU/Linux", you immediately know that it will be open, dev-friendly, malware-free.

That strikes me as being dangerously naive. Someone could make a distro containing all manner of malware... They could include untainted source files to make it look legit... and you would trust it because they happen to call it GNU/Linux?

- Neelix

 

[comradekingu]

Im not seeing that being a huge security risk, besides, the FSF is very on the ball with violations. Fact of the matter is there is need for a word to differentiate a tivo linux distro from a real one.  "Foss linux", doesnt really work, because its foss to begin with.

 

[_wb_]

Most people are fine with the person who started a project getting to name it and don't insist on renaming it once a bunch of other people have contributed to it, since that's petty and pointless.

 
So who started the overall project (fully free OS, kernel+userspace+development tools) that I now prefer to call GNU/Linux, but which most people insist on calling just "Linux"? Linus Torvalds or Richard Stallman? Maybe we should call it Stallmanix after all, no?

That's cute, but I think you missed the point. The person who started it gets to pick a name, not the person who started it gets it named after them. There's a big difference there.

Right, so that would mean Stallman gets to pick the name, and he chose the name "GNU/Linux".

However, my point is that we should not call it GNU/Linux because of the weight of the GNU code contributions, but because "GNU" is a clear reference to the goals of Free Software. If you just call it "Linux", it could as well be a tivo'ized locked-down Android; if you call it "GNU/Linux", you immediately know that it will be open, dev-friendly, malware-free.

That strikes me as being dangerously naive. Someone could make a distro containing all manner of malware... They could include untainted source files to make it look legit... and you would trust it because they happen to call it GNU/Linux?

Obviously I wouldn't trust anything just because of its name. I'm talking about the connotations I associate with that name. "Linux" is good, but it's a bit vague on the actual-freedom-to-modify front. "GNU/Linux" is a bit better in that regard. The RYF label would take away even more doubts.

 

[milkshake]

omg this thread is absolutely pointless because there will never be a complete agreement between the far-left and the mainstream, this is the same in the politics world as well as the software world.

 

[Exophase]

So who started the overall project (fully free OS, kernel+userspace+development tools) that I now prefer to call GNU/Linux, but which most people insist on calling just "Linux"? Linus Torvalds or Richard Stallman? Maybe we should call it Stallmanix after all, no?

Nobody did, because GNU/Linux is not a project. That's like asking "who started the LAMP project" (Linux Apache MySQL PHP if that wasn't clear). Just because GNU calls Linux "the final piece" doesn't mean that it was ever part of GNU or was ever made for the purpose of fulfilling anything with GNU.

It is the GNU project that started the whole movement to make a fully free OS. This was necessary because by then (early 80s), all popular OS'es had become proprietary (and quite expensive too). Of course the GNU project did not invent Free software - there has been public domain software from the very beginning of computing. But the lesson was learned that "public domain" was too weak as a license to avoid things to become proprietary in the long run. It is the GNU project that pointed this out and mostly solved the issue by inventing the GPL license.

Of course the GNU project inspired others (e.g. BSD, X) to become free as well.

To claim anything else is to rewrite history.

GNU is not synonymous with GPL, but as you know, it is called the GNU GPL because the GNU project designed the GPL license. Of course that does not mean that anything that uses the GPL license is directly inspired by the GNU project, but at least some of it is.

GNU's goal was to create a free collection of software that they called an operating system. This was more or less reimplementation of common *nix stuff at the time. The idea of what's important in an operating system has changed since then, and there has hardly been a consensus on what collection of software an operating system should mean (most people don't think this is something that merits some kind of strong definition to begin with). GNU simply does not deserve credit for all free software in an a package being called an operating system today, just because they started making free software in a package they called an operating system 30 years ago.

You keep coming back to the same thing, GNU invented GPL so GPLed "OS" stuff deserves to be credited as GNU. It doesn't make any sense. Either credit all GPL software by having GNU in the name or credit all GPL software as having GPL in the name. Both are crazy but they're more consistent than what you want. GPL being invented for GNU doesn't matter.

As to numbers on contribution sizes: I guess a lot depends on how you measure these things. I think it's fair to assign more weight to "core" components like the kernel, the C library, compilers, coreutils, binutils, the X server, the shell, standard libraries, etc. than to applications like The Gimp or LibreOffice. By any metric, I think it is clear that GNU is still the biggest single contributor, but of course the number of contributors has grown a lot.

That idea of "core" weighting is extremely subjective. And over time many of those things have become much less central. GCC isn't the only thing compiling user code, and on a lot of systems it doesn't even come as part of the base distro. glibc isn't the only significant language runtime. A lot of people aren't using shells (and a lot of them aren't GNU). For the average user the web browser is going to be very important, and for many office and productivity software is going to be very important.

I think it's precisely GNU's modus operandi of making free versions of proprietary software that has caused them to gradually stagnate. They innovate sometimes but not really enough.

However, my point is that we should not call it GNU/Linux because of the weight of the GNU code contributions, but because "GNU" is a clear reference to the goals of Free Software. If you just call it "Linux", it could as well be a tivo'ized locked-down Android; if you call it "GNU/Linux", you immediately know that it will be open, dev-friendly, malware-free.

No, it means it has GNU software on it, that's it. It could have all sorts of closed, dev-unfriendly, malware-infested software in addition to that. Maybe the terms have those connotation to you but I think that's heavily influenced by your own biases.

 

[Grench]

recommends for use in conjunction with the product

It also says

Encumbered formats are those whose implementation is impeded by patents, interface copyrights or other restrictions. If the device supports encumbered formats, it must also support free formats that serve the same general purpose.

It even goes to talk specifically about MP3 playback being allowed despite the fact that MP3s are a restrictive format by patent and copyright. It does not take any stretch to see that ROMs can be considered an encumbered format. Maybe they wouldn't be, maybe the FSF would agree with you, but there is enough room for interpretation that what you are saying is not a logically necessary conclusion.

I believe this to be a false argument.  The FSF uses MP3, essentially a linear-read data file as an example of encumbered format.  I could see this exception being extended to DVD playback, but not game ROMs.

This distinction from the FSF creates two classes of 'stuff' - programs and data files.

You're making the assumption that a game ROM is the same thing as a linear-read data file in an encumbered format that requires a license to read.  It isn't.  A game ROM is a compiled and licensed program with interactive decisions based on user input.  The game ROMs are programs, not simple data files.  They are programs in much the same way as an OS or any stand-alone game is.  They are software, not an encumbered format data file or stream.

The emulation video in the center of the Pandora website is a statement and claim of device purpose.  That statement and claim shows non-free licensed compiled programs (ROMs) operating on the device.  It advertises programs that are not only non-free, but that are not even available for general retail sale (anymore).  

I am quite confident that, based on the arguments above, the game ROMs running in the video on the Pandora's web site would be seen as a violation of the principles of the FSF's RYF certification program.

 

[Hồng Thất Công]

OK you guys work this mess out.  Either way, I will get a P2 for sure if there is one after the smokes clear

 

[sswam]

But it doesn't run a GNU userland, it includes several non-GNU user programs and I don't see an argument that GNU somehow defines or represents them.

The C library, the most used C compiler, the most used shell, most of the toolkit... they're from GNU, and that's the foundation of the userland.

We could use a different libc, a different C compiler, a different shell, a different toolkit, but the fact is we are using the GNU variety of all these.

As for the GUI, some of the most important components are X.org, SDL, GL, xfce, gtk+, qt.

 

I reckon we should mention or credit each of these loud and clear, regardless of RYF certification.

Somewhere on the packaging, in the manual, on the homepage.

Apart from giving credit where credit is due, this is good marketing - developers and hackers will like to know they can use these systems with the Pandora.

The GNU contribution is more visible and more important to developers and hackers, less so to regular users, so crediting GNU tends to be forgotten.

It's a major selling point that we have a "full Linux" system with GNU and X, not just Android or busybox.

edit: I forgot to mention Angstrom

 

[sswam]

To be clear:



- We don't need to call the OS GNU/Linux.

- Open Pandora company should not call the OS plain "Linux"; that's the kernel+modules.

- The OS can be called HyperZaxxon++ , Pandora OS , or whatever.

- We should loudly credit both GNU and Linux among others as major contributors to our OS.

I think that would go a long way towards satisfying the RYF requirements without any unreasonable drama or controversy.

 

[hedwards]

Attracting FSF people is much to the effect of attracting me, im an FSF member. (dun dun) Im not sure thats the only upside (if you will) but i think a certification is also going to educate users, which i guess the certification is good for.

I cant really put myself into "100% free you say?, - nosale.jpg i wanted to not be able to run the kernel i like or the OS i like, this is too much freedom, i cant handle that much"

Again, its not about a certification, its not about a sticker or not. FSF, BSD, Solaris etc- people WILL read up on things to avoid, and they are knowledgeable. So long as its 100% its not like they are going to be shaking their fists. They get agressive if you try to spy on them though.  But its not for them, its for us. A 100% free as in freedom hardware device makes sense and its good for _us_.

And I'm pretty sure that ED can do without the 5 or 6 sales that this will represent. Even most FSF supporters can recognize that spitting on the guys that are advancing things towards 100% is counterproductive in the long run. It's not always possible to do it 100% in one go. Now, when it's normal to be 90-95% free software, then this might be a reasonable stance.

95% is still massively better than most of the other options. But, to do the last 5% at the expense of choosing inferior hardware makes no sense whatsoever. Obviously, if there are two options that are technically similar and would both work, they should choose the more open one, but choosing open to placate zealots makes no sense whatsoever.

This isn't much different than _Wb_'s previous suggestion to remove all the proprietary software from the default install and make people seek it out. Most people are just not that enamored with GPL code. The difference between Pulic Domain, GPLv2, v3, BSD, MIT, APL and such is just not something that they really care about.

And really, if GPL is your sole deciding factor, you can be assured of getting lower quality code. Yes, there is tons of good GPL stuff, but GPL is also the dumping ground of amateur programmers that haven't spent the time to research other license options and think about which one best represents their objectives. If you don't know about licensing, the GPL is the one that you've probably heard about before any of the other licenses. Ever heard of the WTFPL?  Most software projects don't fail because they're stolen by a corporation and run out of the market, they fail because they're too obscure to attract the critical mass necessary to ensure continued development after the original programmer loses interest.

It all depends upon what ones goals are, and with the small community that is the OP community, they should be far more concerned with making these devices as good and as accessible to the mainstream as they can, without selling out on the essentials. You know the controls, the keyboard, the battery life and such.

 

[_wb_]

The RYF certification does not require GPL-only software. There is a big range of licenses that qualify as Free for the FSF.

 

[hedwards]

The RYF certification does not require GPL-only software. There is a big range of licenses that qualify as Free for the FSF.

Honestly, I tend to get confused by them, because the rules are a bit like trying to keep kosher. I'm sure they make sense to somebody, but I see regularly disagreement about which licenses are permitted to be called free software under the FSF's views.

I've got no problem with them trying for the RYF certification, just as long as they're able to do it without compromising the integrity of the project.

 

[_wb_]

Here's an annotated list of free software licenses: http://www.gnu.org/licenses/license-list.html

Here's an overview of free software licenses and their FSF / OSI / Debian etc approval status:

http://en.wikipedia.org/wiki/Comparison_of_free_software_licenses#Approvals

 

[comradekingu]

Being confused and lashing out against what you seem to be against don't go together that well. Five or six extra sales sounds nice. You produced zero evidence to the contrary.  Btw, there are many more fsf members than that already on the boards.

 

[_wb_]

I think the current Pandora already satisfies the requirements for RYF certification. All software that ships with the device is FOSS, except for the GPU and wifi blobs. On the requirements page, it says:
 

However, there is an exception for secondary embedded processors. The exception applies to software delivered inside auxiliary and low-level processors and FPGAs, within which software installation is not intended after the user obtains the product. This can include, for instance, microcode inside a processor, firmware built into an I/O device, or the gate pattern of an FPGA. The software in such secondary processors does not count as product software.

We want users to be able to upgrade and control the software at as many levels as possible. If and when free software becomes available for use on a certain secondary processor, we will expect certified products to adopt it within a reasonable period of time. This can be done in the next model of the product, if there is a new model within a reasonable period of time. If this is not done, we will eventually withdraw the certification.

Of course the GPU and wifi chip (and DSP for that matter) are in the category of secondary embedded processors. So if at some point there would be FOSS drivers for those, we would have to include them within a reasonable period of time in some future firmware update, or else we would lose the certification. But that's what we would most likely do anyway (provided the FOSS drivers have most or all features and they are stable etc; I assume that is implicit in "becomes available for use").

So I don't think there's anything stopping us from getting RYF certification for the current Pandora, and maybe we should do that!

Also even if the Pandora successor would get something like OMAP5 with a PowerVR with proprietary GPU drivers, it would still be possible to get the certification.

 

[comradekingu]

Ask them about it. Sounds interesting.

 

[WizardStan]

So I don't think there's anything stopping us from getting RYF certification for the current Pandora, and maybe we should do that!

Except that, as was pointed out, they've also got a requirement that all communication mention "GNU/Linux" and not just "Linux". That's something that most definitely isn't in place at the moment, if Linux is mentioned it's pretty much always "Angstrom Linux" or just "Linux". If this is a hard requirement it seems like a lot of work for something that doesn't benefit the consumer at all and I don't think it should be pursued.Which is to say you should ask them and find out. I honestly don't think it does qualify though: the exception is for blobs which run on coprocessors: the nubs with built in firmware and the wifi which has its firmware pushed to it when started qualify, but the GPU driver binary isn't uploaded, it is run on the CPU.

 

[_wb_]

So I don't think there's anything stopping us from getting RYF certification for the current Pandora, and maybe we should do that!

Except that, as was pointed out, they've also got a requirement that all communication mention "GNU/Linux" and not just "Linux". That's something that most definitely isn't in place at the moment, if Linux is mentioned it's pretty much always "Angstrom Linux" or just "Linux". If this is a hard requirement it seems like a lot of work for something that doesn't benefit the consumer at all and I don't think it should be pursued.
Which is to say you should ask them and find out. I honestly don't think it does qualify though: the exception is for blobs which run on coprocessors: the nubs with built in firmware and the wifi which has its firmware pushed to it when started qualify, but the GPU driver binary isn't uploaded, it is run on the CPU.

Oh, you're right, the GPU driver runs on the CPU too so it's problematic. I wonder if this would be a deal-breaker for the FSF though, if we say that we want to use a FOSS driver but there is none, and we don't exactly have the leverage to convince TI or Imagination Technology to open up their drivers.

I don't think it's "a lot of work" to do a search-and-replace s/Linux/GNU Linux/ in a few places on the website. It's true that it does not benefit the consumer (except by potentially triggering his/her interest in GNU, which arguably is a benefit of some kind), but it doesn't hurt either.

 

[Rockthesmurf]

I think we should make sure that when we advertise the Pandora we make it very clear that closed source, non GPL software, is a big part of the Pandora, and a very important part of the Pandora. Maybe something like:

"Without closed source software, the Pandora wouldn't be what it is today, consider keeping your source closed as in some cases the best software comes from closed source"

It covers things like DraStic which is a very successful piece of closed source software (there are of course others, but DraStic is a good example that extends out of the Pandora and onto Android).

Actually, I am only joking about the above, I personally find it a bit crazy to even consider signing up for on actively advertising anything about licenses/closed source/open source/paid/free software, just about the only thing I see as being appealing on any level worth counting is the fact that the vast majority of software for the Pandora costs nothing to download, as I can see this being interesting to a good number of people. But actively 'taking sides' with anything to do with software licenses, or bodies who have a strong stance on these matters, just seems to be segregating for no reason. This thread in itself underlines the problem to me.

Finally, it has been stated before, but I get a bit sleepy listening to all the talk about open source drivers being important, when the fact of the matter is so much of the Pandora is open source, yet other than notaz there seems to be pretty much no one looking at all this source, sure it is 'nice' to have the option, but if no one actually uses this option it feels somewhat redundant. There was threads talking about Wifi going back to pretty much release day, and although the issue has never been resolved, notaz (relatively) recently made some modifications to try and improve synchronisation between the Pandora and routers (I don't fully remember the details), it is like others could have done that, what he changed was in source we all had access to, it was clearly a problem, why didn't it get looked into? Probably because it is relatively complicated, relatively specialist, and probably only understand by a crazily small number of people. I believe notaz has also said that GPUs that do have open source drivers typically are very complicated code, and having it is only a tiny step towards actually fixing anything. So I still feel the same, if open source is available, great, if not it doesn't matter that much to me, and probably to a lot of people. So in the grand scheme of things, weigh it into the decision, but do not let it dominate the decision.

 

[_wb_]

The main advantage of FOSS is that everyone can improve stuff, fix stuff, replace stuff with other stuff, and fork stuff. Of course when something does the job fine, there's no need to change things. But having the option to do so is very important in the long run.

Take for instance my latest project, Pandora Image Viewer. It is based on vp, a minimalistic SDL-based GPL'ed image viewer. I took that source code to begin with, which helped a lot to get things started. It uses SDL, which is FOSS, which allowed notaz to tweak it for the Pandora hardware. It uses libjpeg and libpng, which are FOSS so other people could tweak those libraries with NEON optimizations. All those tweaks and optimizations were available to me because those optimized forks are also FOSS.

Other example: freamon wanted another GUI (even though we already had two: MiniMenu and XFCE). It is because the firmware is FOSS (and also because OpenBox is FOSS) that it was relatively easy for him to add another choice to the list.

Even though the vast majority of users will not touch or look at any of the code, the fact that they can is important. If something bothers you, you can always change it if it's FOSS. That is the point.

FOSS matters to many of us. If the Pandora would have a proprietary OS (e.g. Windows), I would not be interested in it at all.