Linus Speaks Out On GPLv3

Slagged writes to mention the word that Linus Torvalds isn't a fan of the new GPL draft. News.com has the story, and someone purporting to be Linus is causing a ruckus in the Groklaw thread on the subject. From the News.com article: "Say I'm a hardware manufacturer. I decide I love some particular piece of open-source software, but when I sell my hardware, I want to make sure it runs only one particular version of that software, because that's what I've validated. So I make my hardware check the cryptographic signature of the binary before I run it ... The GPLv3 doesn't seem to allow that, and in fact, most of the GPLv3 changes seem to be explicitly designed exactly to not allow the above kind of use, which I don't think it has any business doing."

Comments

Linus is wrong

[Anonymous Coward]

I don't think manufacturers have any business preventing me from running my own code on hardware I purchased, at that stage I may as well be using MS Windows.

Re:Linus is wrong

[Spazmania]

This is a twisted and difficult issue.

On the one hand, the whole point of open source is that you can change it and then run your changed version. That shouldn't be suddenly untrue at the arbitrary border between hardware and software. Hardware that uses approved versions of open source while actively preventing my version from running violates the spirit of the thing.

On the other hand, most of us have spent the last decade saying that its OK to use both open source and closed source software on the same machine. No one argues, for example, that you can't run GCC on top of a closed-source unix kernel even though it requires that kernel in order to run. Nor does anyone argue that the processor and other chips used by the kernel must be an open, free design.

The real problem, I think, is that RMS (via the FSF) is trying to force it down our throats as usual. He's a strange bird in that he really gets the freedom issue at one level while it flies totally over his head at another.

I think I'd put the DRM stuff in GPL3 as an optional component and see what happens. Let us authors decide whether we want it. If it works for us, it can be made permanant in GPLv4.

So I'd do something like this: Software released under the GPL MAY designate (on either a file-by-file or full release basis) that it can not be used by any device which by design actively prevents its legitimate owner from adjusting the software or data. Distribution of code so designated would be fully compatible with distribution of any other interlinked GPLv3 code with the sole exception that binary forms of the portions so designated may not be distributed for use with the restricted systems.

But then I'm a vi guy. Maybe if I'd written emacs I'd see it differently.

on the other hand

[ptr2004]

Say I'm a hardware consumer. I decide I love some particular piece of hardware and buy it with my hard earned money. But when I try to run one particular version of open source software customized for me, it doesnt run because the hardware complains it is not validated.

Re:Of Course That's the Point

[Anonymous Coward]

And the copyright owner of the software has the right to restrict the use of that software on devices which perform that hardware check. What's your point?

Re:Of Course That's the Point

[Pausanias]

BlackGriffen wrote:

It's fine to have the hardware validate the software, I don't think anyone can rationally argue against that. What's not fine is to have the hardware refuse to run the software at all. If the user is conscious that the software is modified and therefor unsupported, then the user should have the ability to run any software he chooses. So, have a cryptographic check alongside a message or error light or something about running in unsupported mode, but don't completely cripple the hardware just because you want to avoid support headaches.

What you say makes sense; however, I don't think the current language of the GPLv3 draft is clear on this point. Here is the relevant passage, emphasis mine:

The Corresponding Source also includes any encryption or authorization keys necessary to install and/or execute modified versions from source code in the recommended or principal context of use, such that they can implement all the same functionality in the same range of circumstances. (For instance, if the work is a DVD player and can play certain DVDs, it must be possible for modified versions to play those DVDs. If the work communicates with an online service, it must be possible for modified versions to communicate with the same online service in the same way such that the service cannot distinguish.)

It seems that the first phrase in bold allows what you describe: "implement all the same functionality" does not seem to prohibit a pop-up warning that the code is unsigned. However, the second phrase in bold says that modified versions must be indistinguishible from the original source from the point of view of an outside device. This seems to prohibit that same pop-up warning. So, it seems that Moglen & Stallman still have some clarifying work to do.

The heart of the issue

[PostPhil]

The FSF's stance is controversial (as exemplified by the GPL 3) because it's about freedom, which for all of human history has been hardly understood.

Licenses like BSD/MIT have a view of freedom that is more like anarchy: the "do anything you want" style of so-called freedom (but at least give credit to who wrote the code). This stance doesn't actually create freedom because "anything you want to do" can also include taking freedom away from others. BSD people used to argue that you would still have freedom, only it's with the old code before the proprietary fork, etc. But DRM and other methods of preventing you from modifying and running software is not protected by BSD licensing. So, it is even more true today that BSD-like licensing in actuality has little to do with freedom and more to do with technological research without regard to the sustained openness that made studying that code possible.

Freedom must be preserved and encouraged in order to exist! It is not a spontaneous choice that can be made after neglecting its preservation. Once freedom is gone, once official mechanisms are in place to restrict you, you can't simply make a choice to be free again. When I think of the FSF, I believe they understand freedom as many others have realized throughout history...

"You can only protect your liberties in this world by protecting the other man's freedom. You can only be free if I am free." - Clarence Darrow

"None are so hopelessly enslaved as those who falsely believe they are free." -Goethe

"Liberty without learning is always in peril and learning without liberty is always in vain." - John F. Kennedy

...while the FSF would probably characterize false freedom as this:

"After I asked him what he meant, he replied that freedom consisted of the unimpeded right to get rich, to use his ability, no matter what the cost to others, to win advancement." - Norman Thomas

The more we are tempted by money to deprive others of freedom, the less freedom we all have in the end, and the less it's worth living in such a society even if you're rich. Don't worry about people crying about loss of profitability, etc. History has always shown that there will always be clever people that will find some way to make money, whether people are free or in chains.

Re:You are wrong

[illuminatedwax]

No, it fixes a very important issue.

You know the story of rms' printer driver: he wanted to be able to modify the printer driver so it would bloody work right or work better. He couldn't do that, so he made GNU.

Now let's say the new rms. smr wants to fix his printer which is running embedded GPL software. Great, he thinks, I have the source code to this, so I can just fix the source and make my printer work/better.

Oops! The printer doesn't allow you to do this. This is an awful loophole that restricts your freedom to modify the program. You can modify it, but you might as well write it on on a piece of paper for all that's worth. What the user needs to be able to do is modify the software and use it to really have that freedom. GPLv3 protects this. Linus is really being a stubborn idiot about this.