Building a free future in embedded devices [LWN.net]

Building a free future in embedded devices

Posted Oct 19, 2006 13:26 UTC (Thu) by cventers (subscriber, #31465) [Link]

Was I not clear about something? I believe I addressed that question in my original post as well as my response to the inquiry. I also addressed that in a past post on this subject pretty clearly. But since I have obviously not been clear enough, I will recap here once more.

Rockbox is only useful on embedded devices. If all embedded devices come with hardware DRM, Rockbox development will cease to be possible, because there will be no platform on which users and developers can build, try, and enjoy modified versions. At that point, the developers might as well have gone for the BSD license, because due to their nature of being software specifically for embedded devices, GPLv2 sadly cannot defend them in the ways truly necessary to prevent proprietary forks.

Now, you might point out that manufacturers have a choice of whether to use free software or proprietary software, and if the free software comes with difficult terms they'll just use proprietary software instead and everyone will lose. If you want to do that, go ahead, but I'm going to point out that you're using precisely the same argument that GPL dissenters / BSD license advocates have been using for years, and reality has been anything but. The _actual_ advantage of the GPL to the corporate user is that they can contribute without having their contribution eaten by a proprietary fork, and I believe that the Linux kernel has enjoyed years of impressive contributions from large companies as a result.

Now I'll come full circle to what I was saying before about the player market. Sure, you might see the majority of hardware manufacturers choosing to reject free software and use hardware DRM. If they do that with un-free software, it's going to cost them more to build a player that will be less good. This leaves room for an enterprising competitor to enter the market with cheaper hardware (lacking the DRM), cheaper software (software gratis and libre), and on the whole, a device with more features. So even if most everyone else goes crazy, there is a sensible business reason for someone to invest in a free device, by virtue of the fact that Rockbox exists.

Now you might claim that it doesn't matter, that's a weak argument because Rockbox isn't much of a reason to convince the production of a free device. If you go that way, I'll say that I hope you are wrong but the reason I will call you out for it is because you'll be effectively saying "Rockbox shouldn't have the opportunity to try."

The anti-anti-Tivoization argument often floated by the kernel community is something along the lines of "So what if they lock /their/ hardware up, you can just download the source and run it elsewhere!" Rockbox users don't have that choice. That's exactly the point - they can't. And it's damn foolish to look only at the current very prosporous nature of Linux and assume that such a thing would have always been possible under any hostile conditions. Linux might be safe without the anti-Tivoization clause (I'm not even sure that is true, but hey...), but what about all the other free software projects that collectively dwarf Linux? Are they not important?

Proprietary forks have a history of killing free software. GPLv2 totally prevented proprietary forks in covered software for many years until Tivo came along, and now GPLv3 aims to close the loop-hole letting embedded devices get away with proprietary forking now too.

Of course, I may be wrong about what I'm saying, but you ask me whether GPLv3 is even relevant in this discussion, and I'd be curious to hear why you think it is not. I'm not a Rockbox developer -- I don't speak for those guys, but I think it's clear that GPLv3 is by default better than GPLv2, most especially in their case. It may not be capable of doing enough to keep the project alive, but it does more than GPLv2.

Building a free future in embedded devices

Posted Oct 19, 2006 14:03 UTC (Thu) by vonbrand (subscriber, #4458) [Link]

Again, how does a GPLv3(draft) licensed third party product (capable of) running on <whatever embedded device> change anything if the manufacturer doesn't allow it (by whatever means, DRM or just keeping all specificacions to themselves and running only propietary stuff)?

Building a free future in embedded devices

Posted Oct 19, 2006 18:28 UTC (Thu) by cventers (subscriber, #31465) [Link]

Did you not read what I said? I addressed the point about most of the
manufacturers going off into hardware DRM land. Rockbox's very existence
holds the possibility of convincing another manufacturer to make a free
device, because if they do, they can take advantage of Rockbox and save on
engineering and product development costs in order to give them a
competitive edge. I've raised this point multiple times now -- I'm rather
amazed you missed it.

Building a free future in embedded devices

Posted Oct 19, 2006 19:06 UTC (Thu) by bronson (subscriber, #4806) [Link]

Encouraging manufacturers to go open because it saves on development cost, eh? That's true of just about every open source license ever drafted. As Ingo asked, how does this have anything to do with the GPLv3 specifically?

(other than the GPLv3 containing poorly-understood language that will likely encourage companies to NOT adopt it, of course!)

Building a free future in embedded devices

Posted Oct 19, 2006 19:15 UTC (Thu) by cventers (subscriber, #31465) [Link]

It's relevant because of why GPL is a better license for coprorate
contribution than BSD - it prevents proprietary forks and creates a level
playing field. GPLv2 has a loop-hole that seems to allow tivoization
(proprietary forks). I already pointed this out by the way.

I'll expand on this a little bit, though... if Rockbox could suffer under
hardware DRM, then the manufacturers I spoke of that might be encouraged
by saving development costs could just save those development costs and
make a proprietary Rockbox fork anyway.

I want to request that anyone who is going to reply to me read what I have
to say. I don't expect you to agree to it, but please address my actual
post rather than acting as if you didn't see something. I believe it was
fairly comprehensive.

Building a free future in embedded devices

Posted Oct 19, 2006 19:34 UTC (Thu) by sepreece (subscriber, #19270) [Link]

"a loop-hole that seems to allow tivoization (proprietary forks)"

I don't think a tivoized device is a "proprietary fork" in any real sense. Normally, a proprietary fork would mean that the code, and future changes to it, disappeared into a proprietary version that was no longer open to the community. From the corporate point of view, the "proprietary fork" issue would be about visibility of competitors' (or others') changes to their code. Tivoization, however, doesn't change the GPL requirements that the source code be made available, so that concern does not apply - the enhanced code is still visible/available.

Building a free future in embedded devices

Posted Oct 19, 2006 20:04 UTC (Thu) by cventers (subscriber, #31465) [Link]

Well, I suppose you're right about it not being a proprietary fork in the
sense that would offend commercial contribution -- perhaps I step too far
there. But I still think GPLv3 is important and relevant here because in a
licensing structure that permits Tivoization, no counter-incentive exists
for manufacturers wishing to make use of free software.

And again - I stress that GPLv3 can't crack this DRM issue all by itself,
but to the extent that it might possibly do better than GPLv2, I think it
is quite important.

Building a free future in embedded devices

Posted Oct 19, 2006 21:10 UTC (Thu) by bronson (subscriber, #4806) [Link]

in a licensing structure that permits Tivoization, no counter-incentive exists for manufacturers wishing to make use of free software.

What?! The most effective counter-incentive is the free market. Always has been, always will be. If a device doesn't make its customers happy, customers won't buy it. Anyone who thinks that the GPLv3 is the customer's last defense against the evil DRM manufacturers is simply deluding himself.

Think about DRM-fettered devices... TiVO has become mostly irrelevant in the PVR market. It's being eaten on the high end by Myth and Sage boxes and on the low-end by Comcast's junk. Customers have voted with their feet. If TiVO had been more open with their hardware, I think they would still be selling a ton of boxes. But, no, they were greedy and are now getting what they deserve.

Let's say Apple wants to produce an iPod that only plays iTunes music. Even with all the DRM in the world, could they? Not successfully. Would the GPLv3 affect Apple's decision? Not in the slightest.

Adding controversial, divisive, and poorly-understood wording to a widely-used license to try to protect against a problem that may or may not even exist... That sounds like a clear-cut case of overengineering to me.

Building a free future in embedded devices

Posted Oct 19, 2006 21:32 UTC (Thu) by cventers (subscriber, #31465) [Link]

I think you are making the mistake of confusing hardware DRM with, say,
content DRM. Apple couldn't get away with making an iPod that only played
iTunes music, but they could certainly get away with making an iPod that
only ran Apple-signed firmware.

That's the important distinction. I am like you in this way - I believe
the free market is _vastly_ powerful and will disrupt content DRM
completely. I'm eagerly awaiting the mass purchase of Zune players, for
instance, because I know lots of iTunes customers that are about to lose
their lunch when they realize that their paid-for "MP3s" won't play on
their new "MP3 player" because what they were sold is actually proprietary
DRM-laden crap.

Let's be clear, here...

Posted Oct 23, 2006 2:05 UTC (Mon) by Baylink (subscriber, #755) [Link]

The *issue* with GPL'd code, the *reason* why we don't want people to sequester the code, and why GPL prevents it, is so that purchasers have control over the things they buy.

If the manufacturer is permitted to do things to the hardware which make it impossible for the purchaser to exercise those GPL rights, then they might as well not have them -- and the manufacturer shouldn't be entitled to take advantage of the GPLd code.

I didn't know which side of this argument I was on until I had to compose this reply, but I do now.

Let's be clear, here...

Posted Oct 23, 2006 14:46 UTC (Mon) by vonbrand (subscriber, #4458) [Link]

Strange... GPL doesn't say anywhing about "consumer's control over the device they buy", it talks about the programmer's freedom to tinker with and reuse the source code if they somehow get a copy of the program to run.

You see, this is exactly the problem: GPLv3 advocates are trying to change the fundamental orientation of the license. Sure, this might be 100% in line with the FSF's (current) intentions, but it is not what GPLv2 says, and so is a breach of the solemn promise of "new versions similar in spirit" given with the license. Some people did select GPL because of what it says, not what the post of the week on some random website implies.

What is the most hilarious part of all this is that (as a post above says) the whole TiVo matter is getting resolved by TiVo going under... replaced by closed source alternatives.

Let's be clear, here...

Posted Oct 23, 2006 16:07 UTC (Mon) by cventers (subscriber, #31465) [Link]

I understand that your objection is one of the primary gripes about the anti-Tivoization clause. It is one that many people share. It is also one that I find patently absurd.

Quoting the GNU manifesto:

Complete system sources will be available to everyone. As a result, a user who needs changes in the system will always be free to make them himself, or hire any available programmer or company to make them for him. Users will no longer be at the mercy of one programmer or company which owns the sources and is in sole position to make changes. (emphasis mine)

(Keep in mind it is the GNU General Public License we are talking about here)

And here, quoting from the foreward to a GCC book, written by RMS in February 2004:

Good software must also be ethically good: it has to respect the users' freedom.

[snip]

By the early 90s, the nearly-finished GNU operating system was completed by the addition of a kernel, Linux, that became free software in 1992. The combined GNU/Linux operating system has achieved the goal of making it possible to use a computer in freedom. But freedom is never automatically secure, and we need to work to defend it. The Free Software Movement needs your support.

And here, quoting from the Free Software Definition:

The freedom to run the program means the freedom for any kind of person or organization to use it on any kind of computer system, for any kind of overall job and purpose, without being required to communicate about it with the developer or any other specific entity. In this freedom, it is the user's purpose that matters, not the developer's purpose; you as a user are free to run a program for your purposes, and if you distribute it to someone else, she is then free to run it for her purposes, but you are not entitled to impose your purposes on her.

At least for GNU, RMS, FSF and "free software", the goal has always been to be able to use a computer in freedom. When a manufacturer uses GPL-licensed free software and bolts on a crypto chip that prevents the end-user from running any kind of modified code the manufacturer themselves did not approve of, that does not maintain the right to use the computer in freedom, and in fact, that's not even upholding freedom 1 (the right to adapt the software to your needs).

It's at times like this that I really do sympathize with Stallman and his dislike of the term 'open source'. There has been for a long while people who don't align well with Stallman or his message and 'open source' was a deliberate effort to distance from that. In the process, I think a lot of people started forgetting / stopped caring about the goals of GNU or the FSF, or the 'freedom' in 'free software'. That sucks.

But what really sucks is the fact that there are people who now feel like using the computer in freedom or having the freedom to adapt was apparently never part of the equation in the first place. If you're not happy with that idea, let's talk about that, but this 'freedom' stuff isn't 'what the post of the week on some random website implies' - it's the fundamental philosophy.

And perhaps you never cared to listen to or consider that free software philosophy. That's fine. I read up on it extensively, which was why I'm totally unsurprised and unoffended by the GPLv3's anti-Tivoization clause. It makes perfect logical sense, and it's precisely in the spirit of GNU, FSF, RMS and GPL. But if you never realized this was about using a computer in freedom, and instead adopted the 'open source' idea entirely with no consideration for 'free software', perhaps you should have chosen another license than the GNU General Public License, written by Richard Stallman of the Free Software Foundation?

I agree with Linus. GPLv2 is a great license. But I don't see how GPLv3's anti-Tivoization clause counts for anything more than a clarification of legal terminology to stop something many people (even many kernel developers!) consider abusive.

Building a free future in embedded devices

Posted Oct 19, 2006 14:16 UTC (Thu) by tetromino (subscriber, #33846) [Link]

Your point is moot: no manufacturer in the world. no commercial entity, actually ships Rockbox on its audio players. IMHO, they are not going to start using Rockbox in the near future either (software patents are the most obvious reason). In other words, it doesn't matter whether Rockbox is GPLv2 or GPLv3: the hardware manufacturers don't care either way, because they don't care about Rockbox in the first place.

So by telling Rockbox to adopt GPLv3, you are telling Rockbox users and devs to sacrifice some of their freedom in order to solve a problem that does not exist.

Building a free future in embedded devices

Posted Oct 19, 2006 18:30 UTC (Thu) by cventers (subscriber, #31465) [Link]

You must not have read everything I said.

It's possible my point is entirely moot, because Rockbox might never
succeed. But there were people who said Linux would never go anywhere
either.

Do tell me - how am I telling Rockbox users and developers to sacrifice
freedom by going with GPLv3? What freedom do the users or developers of
Rockbox lose?

Building a free future in embedded devices

Posted Oct 26, 2006 13:31 UTC (Thu) by jond (subscriber, #37669) [Link]

I read on LWN that SanDisk were at least entertaining this notion: http://lwn.net/Articles/186091/

Building a free future in embedded devices

Posted Oct 19, 2006 22:56 UTC (Thu) by bojan (subscriber, #14302) [Link]

> This leaves room for an enterprising competitor to enter the market with cheaper hardware (lacking the DRM), cheaper software (software gratis and libre), and on the whole, a device with more features. So even if most everyone else goes crazy, there is a sensible business reason for someone to invest in a free device, by virtue of the fact that Rockbox exists.

Well, sort of. A device without content won't be successful no matter what. The market for such things just doesn't exist (or is really, really small).

And now we get back to exactly where Tivo is - it's not the manufacturers that want hardware DRM - it is the content providers. So, if they start demanding en masse "have hardware DRM or we won't give you our content", Rockbox under GPLv3 isn't going to play (sorry, bad pun). Some reports related to this Creative story suggest that they removed recording because RIAA insisted on that.

So, when the manufacturers "discover" hardware DRM through their content providing friends, the game in this embedded market will shift to the next level. Now, whether having GPLv2 software available for such devices is a good thing (i.e. because it enables you and me to start a business making such things based on free software for less money) or not - I don't know.

PS. I wouldn't be betting on DRM enabled hardware (i.e. chips) being more expensive forever. If the demand for it goes up, it's going to become "standard feature" and the price will drop to the level of other hardware, or below, depending on volume.