FAQ for project proposal just added to wiki [LWN.net]

FAQ for project proposal just added to wiki

Posted Feb 2, 2012 18:55 UTC (Thu) by tbird20d (subscriber, #1901)
Parent article: A tempest in a toybox

I've seen a lot of similar questions about the project. In an effort to avoid spamming the list with responses, and to clarify the project motivation and goals, I have added a FAQ section to the project proposal wiki page.

Please see: http://elinux.org/Busybox_replacement_project#FAQ

to post comments

FAQ for project proposal just added to wiki

Posted Feb 2, 2012 19:07 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link] (8 responses)

"Q. Is this being done to prevent the SFC from asking for the source to the Linux kernel?
A. No, although it would have that effect."

You are wrong about this. What you are doing is pushing the kernel developers to use SFC instead of using busybox as a proxy

https://lwn.net/Articles/479010/

"I'll be assigning the rights to the SFC to enforce all copyrights that I personally hold in the Linux kernel and any other project they ask me about. "

Even BSD license has a copyright attribution requirement. Enforcing that vigorously would cause enough pain if a vendor hasn't thought about compliance before.

FAQ for project proposal just added to wiki

Posted Feb 2, 2012 20:22 UTC (Thu) by dwmw2 (subscriber, #2063) [Link] (7 responses)

If they think it would have that effect, then the reimplementation does make a certain amount of sense.

There is a cost involved in ensuring compliance for every single product, and with the best of will companies do often get it wrong. A lot of us have violated the GPL at times, quite unintentionally. It's easily done.

The fact of the matter is, it is only busybox which is actively defending its copyright these days. By replacing busybox, they dramatically reduce the risk of sanctions when they fail to comply with the licence.

It doesn't mean that they're intending to break the law; they're just taking the natural step to reduce their risk.

The approach does seem to be based on the idea that only busybox actually holds manufacturers to account, so it's only busybox that they need to replace in order to be able to violate the licences safely.

Even if you don't attribute malice, and you believe there's no plan to intentionally violate the licence of other projects that don't defend themselves, it does seem very cynical.

It's also naïve, because it is effectively a call for copyright-holders in other projects, especially the Linux kernel, to wake up and start to stand up for themselves rather than letting others do it for them.

FAQ for project proposal just added to wiki

Posted Feb 3, 2012 4:26 UTC (Fri) by raven667 (subscriber, #5198) [Link]

That's probably the most clear and unbiased summary of the whole thing.

FAQ for project proposal just added to wiki

Posted Feb 3, 2012 5:26 UTC (Fri) by dlang (guest, #313) [Link]

it's not only busybox that's defending the GPL code, gpl-violations.org is still very active, but they have different triggers for action and (based on the comments) different approaches to the companies in question.

Two different "risks" here

Posted Feb 3, 2012 12:23 UTC (Fri) by thumperward (guest, #34368) [Link] (4 responses)

You've missed the most important point here. There are two different definitions of "risk" here:

* Risk of shipping violating code
* Risk of being sued for shipping violating code

Given that BusyBox is rarely modified to any real extent by the vendors in question, the first risk is only mitigated to the extent of a megabyte of code which probably hasn't been modified anyway. This endeavour is very much intended to counter the second risk. In reality, vendors will still violate the license for software they're shipping, but it will be more difficult to address that because the simplest vector for compliance has been removed.

This is a very pragmatic choice. It is the polar opposite of an ethical one.

Chris

Two different "risks" here

Posted Feb 10, 2012 18:20 UTC (Fri) by landley (guest, #6789) [Link] (3 responses)

Once again, someone implies that it's somehow immoral for me to write new code obsoleting my old code. Here's some context for that:

I personally started the busybox lawsuits. Erik Andersen had a "hall of shame" page during his 7 years of maintainership but, didn't do anything about it. He had his father send out cease and desist letters, but if they were ignored didn't follow up.

Erik did team up with Bradley Kuhn and Eben Moglen and so on to threaten to sue linksys circa 2003, although I believe _they_ approached _him_. (And possibly Harald Welte as well, I'd have to look it up; it was before I got involved.) That action was carefully targetted and gave us openwrt and derivatives. It also caused Linksys to cease all Linux development and migrate its devices to a proprietary embedded OS for several years, but this was considered worthwhile as part of a larger overall outcome. The "embedded home router niche" hadn't _existed_ before Linksys, and now it was a ground-breakingly cheap deployment platform for open source software.

But guess what? The team of people who were _threatening_ to sue Cisco over the GPL back in 2003? They didn't actually file suit:

http://www.forbes.com/2003/10/14/cz_dl_1014linksys.html

All the work was done _out_ of court, and Cisco cooperated to _keep_ themselves out of court. The threat of a GPL lawsuit was just saber rattling to get Cisco's attention, they negotiated the release of source code _without_ever_filing_suit_.

I decided to experiment with filing actual suits to deal with the "Hall of Shame" I inherited from Erik Andersen when I took over as busybox maintainer. There was an existing, specific backlog of reports, which we expected to yield busybox code we could add to the repository. I asked Erik to cosign because he'd collected the hall of shame and what records there were consisted of his inbox, and the versions in question predated my maintainership and often my participation in busybox development. But Erik spent years not doing it, and setting the wheels in motion for it was one of my first acts as maintainer.

The 2003 Cisco effort, settled out of court without ever filing suit and yielding a mountain of useful code that opened up a new ecological niche to Linux, was nothing _like_ the current round of "sue first, ask questions later, and demand tens of thousands of dollars in legal fees and a commitment to keep paying protection money on every new product for years afterwards" the SFLC and SFC efforts have devolved into.

I'm now obsoleting the busybox lawsuits not _just_ because they drive companies away from Linux, but because nothing that's happened since 2006 has come even _close_ to openwrt. We got NOTHING OUT OF IT. All the open drivers for chips from nvidia and ati and broadcom and such we've gotten since then happened either because a company _chose_ to be nice (like Intel, hiring the xfree86 guys to push back against microsoft a bit), or because we _reverse_engineered_them_:

http://bcm-specs.sipsolutions.net/announcement.pdf
http://liquidat.wordpress.com/2007/03/04/the-forcedeth-st...
http://r300.sourceforge.net/
http://nouveau.freedesktop.org/wiki/

And so on. Sometimes Linux got support for hardware _FROM_ BSD:

http://madwifi-project.org/wiki/About/History

All this talk about lawsuits being necessary for hardware support does not match reality, and certainly was not the goal of the guy who started the suits (I.E. me).

I can show evidence that the "sue first and ask questions later" approach of stringing together a series of lawsuits to form a self-financing perpetual lawsuit machine that needs to be fed... that does way, way, way more harm than good. Because I started it, I consider stopping it (and thus ending that harm) to be my responsibility.

You're welcome to believe differently, but the facts do not support your position. I have first person experience here: you do not. You have presented no evidence in support of your position. I've asked for people to show me the code that came out of a busybox lawsuit and got merged, nobody has pointed at any. Instead they choose to weigh a purely theoretical benefit vs actual demonstrable harm.

In insulating your beliefs from challenge by facts, you are taking a position based on ideology, which does not match reality, and thus are a religious zealot, not an engineer. As with "eating meat on a friday" or "we only allow these sexual positions", it's easy to insist things are a sin a priori, because religious questions of morality do not actually require an underlying logic. Thus when logic fails, zealots start talking about morality without being able to show evidence of concrete harm to any specific persons.

(I can debate philosophy with you. But if you don't understand why I believe that we should put a teapot in orbit around Mars in honor of Bertrand Russell, you need to do some reading first.)

Rob

P.S. tl;dr - Bwahahahahahaha.

Two different "risks" here

Posted Feb 10, 2012 18:45 UTC (Fri) by mjg59 (subscriber, #23239) [Link]

I've personally got the source code for the Nook Color and the Viewsonic Gtablet released, both of which developed vibrant developer communities who, thanks to actually being able to rebase to newer kernels by virtue of having the source, have got newer versions of Android running on the devices than the vendor chose to support.

Availability of kernel source for consumer devices is important, and the uses its been put to are surprising. There are even communities of people who develop replacement firmware for TVs. We've seen what happens without active enforcement of the GPL - most vendors simply never supply their code. Enforcement is a necessary part of giving users the freedoms that the copyright holders intended, and those freedoms make a real difference.

Two different "risks" here

Posted Feb 11, 2012 2:08 UTC (Sat) by Cyberax (✭ supporter ✭, #52523) [Link]

Uhm.

R300, R600 and R800 now have public documentation and open source drivers. There's a virtual SVGA driver that provides 3D acceleration to guests which is fully mainlined.

Then there's OpenSource Broadcom driver.

Basically, the only major 'bad' companies by now are NVidia and Imgtec (PowerVR developers).

Two different "risks" here

Posted Feb 11, 2012 12:21 UTC (Sat) by khim (subscriber, #9252) [Link]

But guess what? The team of people who were _threatening_ to sue Cisco over the GPL back in 2003? They didn't actually file suit:

http://www.forbes.com/2003/10/14/cz_dl_1014linksys.html

This is factually wrong. They actually did that.

All the work was done _out_ of court, and Cisco cooperated to _keep_ themselves out of court.

As you describe in other place this "cooperation" was in form "let's sweep it under the rug, continue with sweet talk and hope FSF will eventually go away". It didn't work: FSF is extremely patient, but this is more then compensated by it's perseverance.

The threat of a GPL lawsuit was just saber rattling to get Cisco's attention, they negotiated the release of source code _without_ever_filing_suit_.

Yes. Once. And it was incomplete. And then Cisco issued new versions of both software and hardware - again with GPL violation. And as your admit when they started new "proper" effort they had no plans to do anything with old violations.

All the open drivers for chips from nvidia and ati and broadcom and such we've gotten since then happened either because a company _chose_ to be nice (like Intel, hiring the xfree86 guys to push back against microsoft a bit), or because we _reverse_engineered_them_.

Yes, but can you explain just why companies suddenly decided to be nice when they dragged their feet for years? And why are you so sure lawsuits have nothing to do with that?

All this talk about lawsuits being necessary for hardware support does not match reality, and certainly was not the goal of the guy who started the suits (I.E. me).

You may have started it, but your certainly don't have power to stop it. Simply because you don't own rights for a lot of GPLed code.

I have first person experience here: you do not.

As Kleiner said: It's difficult to see the picture when you're inside the frame. You can only talk about direct consequences of all that activity, but how can you claim that Broadcom's change of heart is not related to SFLC and SFC efforts?

You are correct when you say Never think you're _irreplacable_. Nobody is. That's certainly true. And if your goal is to make software used by millions then these lawsuits may hurt. But Stallman's initial goal was completely different: the goal was to be able to hack on the software you "own", not to spread your own software as an act of goodwill. GPL was not a goal in itself, but means to this end: if someone uses your code then GPL guarantees that you get your own code back with appropriate changes! Clever hack - but it was broken by tivoization and by blatant ignorance of GPL conditions. Solution to tivoization is obvious: GPLv3. It makes sure that manufacturer can not forbid you to modify your own code. Solution to GPL violations are lawsuits. If manufacturer does not give you enough data to start hacking on your code then what's the point of the whole exercise?

From this POV lawsuits are quite successful: a lot of companies published enough sources to make the whole classes of devices hackable. The fact that GPL stopped being the most popular license in the world is unfortunate side-effect, but alternative (GPL continues to be most popular license while manufacturers effectively treat it as BSD license) was clearly worse from FSF's POV.

P.S. It's funny that you argue are so adamantly against SFC's efforts where the only "proven disaster" happened because of FSF. FSF certainly started it's efforts way, way before your involvement. In fact G++ story happened before busybox project was even started!