GPL concerns halt Kororaa live CD (NewsForge)

There are trade secrets contained within the video drivers themselves that either vendor cannot afford to have open for the competition to see and copy. I'm not sure how familiar with people are with modern 3d-accelerated video drivers, but according to an article I read on AnandTech (www.anandtech.com) years ago, video drivers contain what's called a "runtime compiler". From what I understand, this is the part of the video driver that allows for a fully programmable GPU by allowing 3rd party developers to write their own programs to utilize the hardware on the video card. When a developer writes a program (such as a pixel shading function utilizing the pixel shaders on the card's hardware), the source code for the program must be compiled by the runtime compiler on the fly at runtime. The overall performance of the video card relies heavily on the ability of the video driver to compile these programs as fast as possible.

Some of you may remember back in 2000 when nVidia released their Detonator 3 and 4 drivers. These drivers alone increased performance on the GeForce-based cards by a great amount. This was made possible by finding ways to optimize the runtime compiler contained in the driver itself.

As you can see, the runtime compiler contained in the video driver plays a very important role in the overall performance of the video card. We all know how cut-throat of an industry the video chipset maket is. Because of this, each company must protect the source to their runtime complilers found in their drivers, otherwise they lose their trade secret, which they've invested lots of time and money in to develope and protect.

This isn't intended to be an argument for either side, just an explaination of why nVidia and ATI will not open-source their drivers and their reasoning behind it.

Ultimately it'll take a court case to settle, there really isn't any case law on this particular point, but from what I've seen there's a pretty strong concensus among the lawyers that it doesn't matter a bit whether you load it as a dynamically linked module or statically compile it into the kernel - you're still creating a derivative work.

As I tried to make clear, this isn't a problem for most distributions, because they don't do either - they simply 'aggregate' the materials on the same disk, but actually choosing to link and run the blob is still a step the customer has to take on his own. That's the loophole that lets nvidia and ati get away with playing games here.

But on a live cd, that's not really practical.

brainless? hmm. I almost sunk to your level just now but thought better of it. While bignose did a clever logical inversion, it suffers from the same basic weakness that the other anti-nvidia posts do - namely the illusion that the tiny, cantankerous and demanding market segment of gpl-only desktop customers is in a position of strength, and that they can dictate terms to commercial vendors who sell 95% of their product to ms windows users.

There was no "logical inversion" by bignose. It was a nicely logical counter to an emotional, irrational rant.

I have no clue what "gpl-only [sic] desktop customers" to which you refer. However, there appears to be considerable market influence by those who choose and support the GPL. Perhaps the strongest example is when Sam Palmisano proclaimed that IBM couldn't match the community's efforts and still provide an acceptable return for their shareholders. Of course, that was 5 years ago and there is probably a bit more influence involved now than when IBM was swayed.

If you look at Jonathan Corbet's article, Kernel Summit 2005: The hardware vendors' panel, you will find this quote:

The next speaker was Andrew Vasquez from QLogic. His brief talk went over some of the hassles he has had to deal with. At the top of the list was firmware blobs. They create big patches and have GPL issues. Interestingly, he said that the firmware issues, along with pressure from "a major distribution," are motivating the company to move its firmware back into the device. If the driver does not have to load firmware to make the device function, these issues go away.

Since, as he put it, "a double-digit percentage" of QLogic's sales are for Linux systems, providing good support (and keeping the community happy) matters to the company.

[Emphasis added]. "A" generally means just 1. It seems reasonable to assume that all together have more influence than 1 will have, so I think maybe your characterization of the "strength" of GPL users is a bit off. The above is not an isolated incident. I seem to have heard about other little businesses, such as HP and Intel, being convinced that opening drivers can be a good thing.

Since nvidia provides the best linux drivers available anywhere, why should they be criticized for how they choose to develop them? BTW nvidia claims to have legal obligations not to disclose the intellectual property of other parties which is contained in the hardware specs and API. OK, sounds entirely possible to me.

Your opinion of the quality of nVidia's drivers for Linux has nothing to do with Kororaa's problem, so far as I can see. Maybe it was part of the motivation for your emotional response? Any alleged encumbrances nVidia suffers with their drivers is also not Kororaa's problem.

Kororaa's problem appears to be someone claiming the right to demand Kororaa cease distributing Kororaa's LiveCD, based on an alleged GPL violation. A copyright owner has the standing to demand a cessation of infringement. Is the one who emailed Kororaa a kernel developer? I didn't see that being claimed, yet the emailer bases the demand on violation of "the linux [sic] kernel license". Now the file COPYING that came with my Linux has this at the beginning:

NOTE! This copyright does *not* cover user programs that use kernel services by normal system calls - this is merely considered normal use of the kernel, and does *not* fall under the heading of "derived work". Also note that the GPL below is copyrighted by the Free Software Foundation, but the instance of code that it refers to (the Linux kernel) is copyrighted by me and others who actually wrote it.

That seems to cover permission to use binary modules with respect to Linux.

What about the GNU software on the CD? Take a look at What is the difference between "mere aggregation" and "combining two modules into one program"?:

Mere aggregation of two programs means putting them side by side on the same CD-ROM or hard disk. We use this term in the case where they are separate programs, not parts of a single program. In this case, if one of the programs is covered by the GPL, it has no effect on the other program.

That appears to make it necessary to combine those non-GPL drivers with something besides the Linux kernel to be a violation of the GPL for other programs on the LiveCD. Just having the programs on the same CD doesn't trigger. Just in case, though, ...

Does the email complainer hold a copyright in any of the rest of the GPL software on the LiveCD? I can't tell. It would not appear to be so, based, again, on the reference to Linux. If you look at Who has the power to enforce the GPL, you will see:

Since the GPL is a copyright license, the copyright holders of the software are the ones who have the power to enforce the GPL. If you see a violation of the GPL, you should inform the developers of the GPL-covered software involved. They either are the copyright holders, or are connected with the copyright holders.

Who is the emailing complainer and what copyright does he or she hold that provides the power to demand ceasing of distribution?

The whole thing looks bogus. I suspect it's either someone who thinks the unethical use of terror tactics is excusable when used to promote ethical software, or it is someone who is looking for a way to discourage the use of the GPL under the ruse that it is tyranny, rather than protective of the freedoms of software users.

Quoting http://news.zdnet.com/2100-3513-6061491.html :
> For Nvidia, intellectual property is a secondary issue. "It's so hard
> to write a graphics driver that open-sourcing it would not help," said
> Andrew Fear, Nvidia's software product manager.

While I disagree with the reasoning for not open-sourcing the driver, it
is indeed stated here that those legal obligations you are speaking of
are for naught.

They can't force me to buy their hardware. They'll either support my OS of choice, or I won't buy it, simple as that.

Now maybe they're perfectly happy without my business. But you know what, I don't think so. Because they sure go to a lot of trouble to put up this deceptive illusion of linux support. They work a lot harder, in fact, at creating that illusion than they'd need to work to actually be supported. It kind of makes you wonder...

But the point is that the libre OS market is obviously big enough for them to spend all this unecessary effort on providing fake support for, so it follows that it is indeed big enough for them to care about.

So what's the problem? They're more comfortable with faking the support, for reasons of ingrained old-think, and enough well-meaning but ignorant users buy the illusion that they get enough reïnforcement of their own ideas to sustain the illusion? It's a guess.

At any rate, the market is big enough to get their attention or they wouldn't be doing what they're doing. So I think the next step is to raise awareness with that market that what they're doing doesn't actually constitute supporting, or being supported with, libre software.

[...] namely the illusion that the tiny, cantankerous and demanding market segment of gpl-only desktop customers is in a position of strength, and that they can dictate terms to commercial vendors who sell 95% of their product to ms windows users.

I think it is within our rights to make such requests, since: we have bought their stuff, they promise Linux support, and the natural way to do that is to integrate the drivers into the kernel. Why is it that some people always picture gpl supporters as dictating other people's behavior? Why do we get called derogatory names when we request only what is seen as natural for every other vendor? (If you don't see the derogatory terms, "tiny" and "cantankerous" is bad enough, but being called a "market segment" is really beyond bounds.)

nVidia is criticized for how they develop their drivers, because they are Linux drivers, they are not libre, they make users go through needless hoops, they cannot be debugged, and all this just because "It's so hard to write a graphics driver that open-sourcing it would not help". If this does not merit some criticism, then we might as well buy our copy of XP and go play minesweeper.

maybe I'm missing something but I didn't see a clause added to the kernel COPYING file that allows binary modules, all I see is an explenation from Linus that he considers *user space* applications that use normal syscalls indepdent and not a derived work. But not a word about kernel modules....

I think the problem is more my lack of clarity than you missing something. I wasn't suggesting shipping the ready-compiled driver; just the separate parts in the same way that nVidia is allowed to do. The confusion appears to be from my misuse of the terms you had selected in your previous comment. Sorry about that.

Including the same things that nVidia ships on the same CD as the kernel and the rest of the operating system would be "mere aggregation". It would be unusable as shipped and require the user to decide to compile it. That compiling produces the "taint", which is a thing that is not distributable under GPL.