"Respects Your Freedom" certification?


Outlandish Scenario:

Pandora successor is built with all free drivers and OS - and then Valve ports Steam and several associated non-free (paid) games to ARM. (substitute any paid software company or similar)

Would FSF certification of a Pandora successor preclude ED from being able to advertise and link to Steam from the Pandora successor's web site and OS linked application store (with ED receiving the appropriate compensation for driving business)?

Would ED and the P2 site be able to recommend/advertise/link to a paid software company and site?
 
This is a good question and I'm not entirely sure what the answer is. I guess the main question is: what counts as a recommendation? I'm assuming that it would be fine to advertise on the P2 site that the hardware is compatible with certain non-free software, but it would be a problem to advertise the hardware as being mainly intended to be used with non-free software, e.g. by calling it a "Steam handheld" or something like that.

Note that non-free and paid is not the same thing though: it's fine to ask money for free software, and non-free software can be gratis.
 
The fact that it has to even be a 'good question' as to how a private company can or cannot advertise, promote or partnership around their device without a 3rd party committee (FSF) making moral judgments around whether or not it's 'appropriate' in their world view makes it not worth pursuing.

It is granting a 3rd party committee a certain degree of power in the ongoing business.  After all, the publicity of getting the FSF certification would be nice.  The publicity of getting it revoked would be near catastrophic.  This would give the FSF certain leverage over business decisions that, quite frankly, aren't any of the FSF's business.

The most ironic, and somewhat satisfying, thing to me in this would be if the Pandora successor qualified in every way for FSF certification but decline to apply for certification due to it requiring that the project sacrifice certain freedoms if it were to apply.

Yes, efforts should help ensure that we have open and free developer access to the software and drivers on the Pandora successor.  No, ED shouldn't sacrifice -any- rights or freedoms within the product, advertising, partnerships, community or future capabilities to get any company or organizations seal of approval - with any form of restrictive strings attached.
 
The restrictive strings attached are just that if you don't satisfy the requirements for the certification any more, you lose the certification. That seems rather normal to me, and necessary if you want the certification to be meaningful. Otherwise you could ship a 100% Free OS with the first few units to get the certificate, and switch to Windows 8 from then on and get to keep the certificate. You can call that "giving the FSF certain leverage over business decisions that aren't any of their business" if you want, but to me it is common sense. How else are people supposed to be able to associate the "RYF" label with anything meaningful, if there is no link enforced between having the label and meeting its requirements?

If we agree that the requirements for the RYF certification are things we as a community want anyway, then I don't see why applying for the certification would be a bad thing: to the contrary, it would help to convey our intention that we have a long-term commitment to these goals. Because indeed, like you say, if at some point later on, the OP2T decides to, say, make a deal with Microsoft (maybe Microsoft approaches them and offers very cheap or gratis licenses) and to start shipping Windows 8 with the P2 from then on, it would result in the certification getting revoked and a lot of bad press. So getting the certification would be a good way of saying: we will not go that way, and you don't have to take our word for it, because we got endorsed by an independent third party that you can trust on such matters of software freedom.
 
Steam isnt on ARM, or MIPS. Thats the kind of power i dont want to give a 3rd party. They make it so that they can break it. So thats that. And speaking of  architectures, the X86 people are the worst, even get to me sometimes.

How is non non-free hardware going to ever going sour on anyone? Its not like it changes, so that certificate cant be revoked. Lets say we make a version 2.1 and there is a global shortage on the parts in the original version. We could at that point _not_ put the fsf approved on the box. We are that capable.

I dont really know what to call it, closed firmware encumbered hardware maybe? It isnt open hardware just because its non-nonfree, because open hardware means the schematics/gerber files etc are open so you can reproduce and modify the hardware. That is patent law, which isnt copyright. And lets not get into software patents.

As for shipping a FSF approved distro, I dont know, id say ship debian and take it from there. Any freedom is good freedom, but dont throw the baby out with the bathwater.

Debian is huge, its rock solid, and its what most custom distros are built atop. Since the base is stable that usually means newer software and GUI things.

How much non-free do we have? Its those games the gwc-guys put on there, then the drastic, and kami retro. The latter two are community centric, and paid for, so if its about money lets buy them and have them be pay as much as you like or whatever the creators see fit. Its not like you cant steal proprietary software either, so just because you can compile it from a sourcecode people will still buy it, just as much as they wont steal the closed binary. Im going to go on a limb here and guess that not every rom-dump for those emulators is within realms of legality.

I say lets set sight at shipping stock debian, it makes sense to tailor make the packages to the hardware, not a generic install, which isnt a problem unless you would add on external wifi, but if the internal one is flawless, it isnt a prevalent usercase scenario.  Besides, there are free options etc etc.  So once the hardware is free, the software is free, which the argument of saving space on nand takes care of before you get into morality.

I see the effect of having a certification possibly revoked, so maybe that part should be unofficial.  In my head it plays out like a 50s propaganda scenario^^. Every time you give users freedom, you are working for the FSF! Must read up on what that requires and how it works out though. Good thing we have wb for research. Maybe we could discuss it with them come to some sort of agreement. We dont exactly have a device that qualifies to put a certification on right now, so we could discuss and research for future endavours.

The FSF doesnt troll people who do good, they recently did a fundraiser for replicant i noticed, so they can be very nice. Replicant is a free android distro. They make sure the telephony stack is free and that the modem doesnt control the SoC and that they dont share memory. Plus stripping the google mandatory behinds the scenes usage survey and the google apps. Then there is GPS, camera, which needs firmware. I only lurk, but ive understood qualcomm is bad once you get LTE modem. So thats why i dont like the snapdragon 800, its fast and the adreno 330(?) is splendor, is has the second best drivers with lima being the best.

Etna_viv is the one for the gwc-zero graphics, but that isnt so widespread, and its not the fastest, and its not the most working freedriver. A free driver for the SGX powerVR (like what is in the pandora now) is a high priority FSF project.

Edit: Proprietary anything obviously isnt the way to go, but the question is how far to go to avoid it. There are a limitless number of players that ship android gaming centric devices, and most of them shit on freedom for the sake of convenience. Maybe going with freedom for freedoms case can be a benefit rather than going the way of convenience. I think the benefit of having a involved community that makes it our device is most important. In that respect we need as close as we can get.

TL;DR Since not having defective by design hardware was convenient and everyone got along, lets see where the software question is in practice. What closed source things are there, and what are the alternatives? Maybe we can switch the board software to phpBB or something that works well. Maybe do openstreetmap for the member map. Stop documenting good-old-games on the wiki? Im having trouble finding _real_ issues in the sense that they are insurmountable or that i care for the sake of freedom. The forum is a real issue for me, and it takes a bit of effort to change, is that the "problem"?
 
Last edited by a moderator:
@comradekingu

nice posts

why don't we aim at free and open-soure and just don't care about RYF certification? its not like those guys need that certification to be sadified to use a P2. Beeing FOSS should be totally awesome enough! ; ) And it wouldn't limit us i.a. discriminate against closed source and/or comerciall software.

There are simply too much restrictions for me to support it. The codec pack for example is pretty neat. NTFS is still the way to go for external drives because it can be read by Linux and Windows. Im not living in a bubble where everyone I know is a geek using Linux. In fact I'm the only one in my family and circle of friends. Not beeing able to use NTFS is very annoying. And no, using FAT32 is not an option because I use a lot of containers beeing larger than 4GB

and thats only one restriktion - reminds me a little of Apple, creating a "closed system" with free software xD

So, unless you have a good insight on how to explain the lack of these people, I don't see why an additional certification will make them come either.
a high price for old hardware + there are lots of other ineresting platforms: Arduino, cheap computers (Raspberry Pi, BBB,..) smartphones, tablets, Peable, and, and, and...

but I belief this will change with the next Pandora generation ; )
 
Last edited by a moderator:
The restrictive strings attached are just that if you don't satisfy the requirements for the certification any more, you lose the certification. That seems rather normal to me, and necessary if you want the certification to be meaningful. Otherwise you could ship a 100% Free OS with the first few units to get the certificate, and switch to Windows 8 from then on and get to keep the certificate. You can call that "giving the FSF certain leverage over business decisions that aren't any of their business" if you want, but to me it is common sense. How else are people supposed to be able to associate the "RYF" label with anything meaningful, if there is no link enforced between having the label and meeting its requirements?

If we agree that the requirements for the RYF certification are things we as a community want anyway, then I don't see why applying for the certification would be a bad thing: to the contrary, it would help to convey our intention that we have a long-term commitment to these goals. Because indeed, like you say, if at some point later on, the OP2T decides to, say, make a deal with Microsoft (maybe Microsoft approaches them and offers very cheap or gratis licenses) and to start shipping Windows 8 with the P2 from then on, it would result in the certification getting revoked and a lot of bad press. So getting the certification would be a good way of saying: we will not go that way, and you don't have to take our word for it, because we got endorsed by an independent third party that you can trust on such matters of software freedom.
Where I take issue with it is that by making that a goal the project would be -giving up freedom- to use Ubuntu or Debian as the default OS.  It would be -giving up freedom- to use a graphics, wireless or USB controller 10X faster (hypothetical) than what we can get free drivers for - all to bow to some 3rd party's twisted idea that the fruits of all software development labors should be, 'free', and that the project should surrender real freedom for their ideal.

Real freedom is independence.  If FSF 'approval' would require -limiting- the choice of OS, -limiting- the software that can be advertised and linked to the product and -limiting- the hardware choices available, then it is not promoting 'freedom' but some twisted and limited idealism bordering on communist philosophy that simply uses 'Free' as a marketing slogan.
 
Would FSF certification of a Pandora successor preclude ED from being able to advertise and link to Steam from the Pandora successor's web site and OS linked application store (with ED receiving the appropriate compensation for driving business)?
Certification is not forever, once you've got it you are not legally bound to keep it. In the event that such a thing happened I would say certification be damned, advertise the P2 as a walking Steam machine and let the FSF decide whether it still counts or not.

Where I take issue with it is that by making that a goal the project would be -giving up freedom-
If we agree that the requirements for the RYF certification are things we as a community want anyway
Very big if there. The general requirements of the RYF certification are definitely a good thing, free drivers are strictly better than their equivalent non-free, but not if such equivalence doesn't exist. If equivalence doesn't exist then a choice will need to be made: better hardware or free drivers, and that choice cannot possibly be made until we know how much better the hardware would be and how much trouble we might expect from the binary blob of a driver. Goals are allowed to change, there is no reason we can't say "RYF certification is a good thing" now, and then discard it later when it turns out the only compatible hardware sucks.
 
Might it work to sell the same hardware with two different names, branding, and recommended software? So people could choose whether to buy the Freedom Thing or the Walking Steam Machine ;)
 
Just to clarify something: the codec pack would not be a problem at all, as long as for every non-free file format there is also a free alternative offered, and that would be the case. So it's OK to have mp3 support if you also have ogg support, it's fine to have NTFS support if you also have ext2/3/4 support, etc.

If we agree that the requirements for the RYF certification are things we as a community want anyway
Very big if there. The general requirements of the RYF certification are definitely a good thing, free drivers are strictly better than their equivalent non-free, but not if such equivalence doesn't exist. If equivalence doesn't exist then a choice will need to be made: better hardware or free drivers, and that choice cannot possibly be made until we know how much better the hardware would be and how much trouble we might expect from the binary blob of a driver. Goals are allowed to change, there is no reason we can't say "RYF certification is a good thing" now, and then discard it later when it turns out the only compatible hardware sucks.
Of course - I would say it is a design consideration, and as far as I'm concerned an important one, but it's obviously not the only consideration. But as far as I can tell, there is no reason to assume that non-free hardware is that much better than free(d) hardware: e.g., it is not the case that the PowerVR GPUs (or the SoCs that have them) are better than the Mali ones. And even if we do want to have hardware components that are currently non-free, it's still a theoretical possibility that somehow the component vendor can be convinced/pressured to open up the drivers (not very likely for a small-scale project like this, but you don't know until you try) or that we can find someone who can RE the component and write a FOSS driver for it.
 
But on the repo, which does have the status of "officially recommended", there is a mix of all kinds of software. To satisfy the FSF criteria, it would be necessary to only officially recommend Free Software. Of course "third parties" are free to offer what they want, but the "Official" project should not offer non-free software, or even advertise or link to the "third parties" that do. They don't have to try to prevent people from using that non-free third party software, but they shouldn't make it very easy either.
I'd guess that 95% of Pandora owners / users like to play commercial games in addition to "free" games,

so when you say "it would attract the kind of people we want to have in our community"

you are not in fact talking about our community but some imaginary Pandora community that you might like better.

Even Debian doesn't qualify as an "RMS-approved free distro" and conversely GNU (documentation) does not qualify as Debian free software.  They are excessively concerned with their licenses and legalities.

It might be nice to flag apps whether they are fully "free software" or not on the repo, but I would not want to remove all the non-free software.

We have a strong free software OS and we can also run, share and enjoy free software and proprietary software.  All's well.
 
Last edited by a moderator:
I'd guess that 95% of Pandora owners / users like to play commercial games in addition to "free" games,

so when you say "it would attract the kind of people we want to have in our community"

you are not in fact talking about our community but some imaginary Pandora community that you might like better.

Even Debian doesn't qualify as an "RMS-approved free distro" and conversely GNU (documentation) does not qualify as Debian free software.  They are excessively concerned with their licenses and legalities.

It might be nice to flag apps whether they are fully "free software" or not on the repo, but I would not want to remove all the non-free software.

We have a strong free software OS and we can also run, share and enjoy free software and proprietary software.  All's well.
I am not talking about converting the repo to an FSF-approved Free GNU/Linux distro, or about removing all the non-free software. The "Respects Your Freedom" certification is about the hardware, and it only requires that it is possible to run a 100% Free OS on it, not that it is impossible to run anything non-free on it (in fact, it would be impossible by definition to have RYF-certified hardware that makes it impossible to run non-free software).

The current firmware that is shipped with the current Pandora already is a 99% Free OS: the only problem are two binary blobs (for the PowerVR GPU driver and for the wifi). So I'm arguing that it would not be that hard (or at least not impossible if we make it a design goal) to get to 100% in a Pandora successor and meet the criteria for the RYF certification.

Also, I'm not saying that the majority of the current Pandora community doesn't want to use non-free software or anything like that. All I'm saying is it would benefit our entire community - even those who want to play or develop non-free games - to attract more people who like the notion of Free software and who may themselves contribute new Free software or improve existing Free software.

We could also go the opposite way, make the hardware less free, add more DRM features to the hardware and the OS (e.g. make sure that only encrypted binaries with proper authentication can be executed), make a "walled garden" appstore, let the warranty be voided if you try to root your device and all that, and try to attract more non-free software developers to our community, who want to make sure that their copy-prevention mechanisms don't get circumvented. In my opinion that is not the direction we should go, and those are not the kind of people we want to attract.
 
stardroid

NTFS-3G is GPL software though, so there is no problem with that.

Samba too for that matter. Its not your friends choice you are making, you are making yours, and certain people have made that choice convenient. Pandora is not going to use fat32 or NTFS on what it ships, being a liability, because when you do that, microsoft sues you. That is to say untill 2015(?) when their short filename patent expires. (not used in exfat) on the precaution that you are not a big company with a portifolio of software patents (can be totally unrelating) to strike back with.

IBM has given rights to their patents to everyone who uses free software, and there are many others in that pool.

Tomtom got sued for using fat32, and since they use linux on their things, they went to the FSF lawyers and asked them to defend tomtom. In tomtom vs microsoft They did, but tomtom gives _nothing_ back to the community. And they settled for paying Microsoft. Thank you very much tomtom, hope your convenience is upheld the next time the big boys come a knocking.

And then they joined the pool...

You use Android, you get sued. Think it costs 5-15 dollar per device, to microsoft. Who wont even disclose what patents its trolling with. They make more on android than on windows phone. The kicker here is that they cant sue google, they go after the companies, so i guess that is why we cant use android (unless its replicant) Different rules apply when you are from china, because then MS cant touch you. Just as FSF cant touch them when they violate the GPL. They play by different rules, and they can include roms with emulators out of the box if they want to. If ED wants to sell them in Europe he has to reflash first, which has happened.

Edit: I think panda packs was a good idea for spreading software. Those would be much easier to assemble if you could determine the license on the repo easily. And its fri or non-fri that matters in that respect. As it is now you have to scourge deep to find it, as its often just left blank.
 
Last edited by a moderator:
he current firmware that is shipped with the current Pandora already is a 99% Free OS: the only problem are two binary blobs (for the PowerVR GPU driver and for the wifi).
The wifi driver is free, the binary blob is the firmware that runs on the wifi processor. I wonder if that falls under their "software run on a secondary processor" exception, or if it still needs to be free because it isn't actually built right into the IO device. The nubs are definitely described by the exception and the GPU driver blob runs on the CPU so it is definitely not covered.
Might it work to sell the same hardware with two different names, branding, and recommended software? So people could choose whether to buy the Freedom Thing or the Walking Steam Machine ;)
Unfortunately not.
To prevent confusion among customers about exactly what product has been certified, any other products offered by the seller, which are not certified by the FSF, must be easily distinguishable from certified products: their names must not be similar and their packaging must also not be similar.
 
he current firmware that is shipped with the current Pandora already is a 99% Free OS: the only problem are two binary blobs (for the PowerVR GPU driver and for the wifi).
The wifi driver is free, the binary blob is the firmware that runs on the wifi processor. I wonder if that falls under their "software run on a secondary processor" exception, or if it still needs to be free because it isn't actually built right into the IO device. The nubs are definitely described by the exception and the GPU driver blob runs on the CPU so it is definitely not covered.
Right. The wifi firmware blob would probably permitted, but if a Free version of the blob would become available, we would need to "adopt it within a reasonable amount of time".

We may still want to choose a wifi module which already has a fully Free driver and firmware (because it kind of sucks to only have the binary blob firmware, as anyone trying to debug Pandora's wifi trouble will have found out), but only the Free driver would be needed to get the RYF certification.

Might it work to sell the same hardware with two different names, branding, and recommended software? So people could choose whether to buy the Freedom Thing or the Walking Steam Machine ;)
Unfortunately not.
To prevent confusion among customers about exactly what product has been certified, any other products offered by the seller, which are not certified by the FSF, must be easily distinguishable from certified products: their names must not be similar and their packaging must also not be similar.

So if it has a sufficiently different name and sufficiently different packaging, this would be a possibility indeed: sell the same product in two different ways, one with the RYF label, the other without. This would make sense if one version of the product would be shipped with a 100% Free OS while the other would ship with, say, Windows 8 pre-installed and some DRM-enabled games with it (of course that second version would also be more expensive, unless maybe Microsoft wants to subsidize such a thing just to piss off the Free software people ;) ).
 
I suspect by "packaging" they mean the product as well, plastic housing and everything so that there is no confusion. If the only difference between the free and non-free Steam machine is a sticker under the battery then they will probably not take very kindly to it :p
 
Oh, I didn't consider that interpretation - I interpreted "packaging" as the cardboard box, not the case. If that's really what is meant, then it seems a bit overly harsh - after all, the current Pandora could be confused with a Nintendo DS from a suitable distance ;)

I'm assuming that what they want to avoid is having something like the hypothetical "Dragonfly DX0346" and the "Dragonfly DX0347", which are sold by the same vendor and look the same in every way except some of the fine print on the box, but they are completely different in the sense that the *6 has the RYF certification, while the *7 has DRM-enabling hardware and software. In such a case, it would be easy to confuse things, and have someone think "the Dragonfly is FSF-endorsed" and buy the *7 to find out later that is not at all.
 
Nope debian host its non-free repository. While not enabled by default it is there. I bet that simple existance is enough to not be certified
 
Back
Top