About the calculus for the project [LWN.net]

About the calculus for the project

Posted Feb 2, 2012 4:35 UTC (Thu) by tbird20d (subscriber, #1901)
In reply to: About the calculus for the project by Company
Parent article: A tempest in a toybox

Did I get that right?

No. We have extensive mechanisms in place for compliance, for both business and moral reasons.

to post comments

About the calculus for the project

Posted Feb 2, 2012 10:18 UTC (Thu) by jospoortvliet (guest, #33164) [Link] (40 responses)

So then the real problem is that you're so afraid of making a mistake in your compliance procedures (despite believing they are quite good) and getting sued the sh*t out of you by SFC that this project was started?

I personally find it very hard to believe that they would go after Sony, which, as you say, has a website with FOSS code, a contact point and proper procedures. There are much bigger fish to fry... But you could simply talk to them, ask what room you have for mistakes?!?

About the calculus for the project

Posted Feb 2, 2012 18:59 UTC (Thu) by tbird20d (subscriber, #1901) [Link] (39 responses)

I don't think they'll come after Sony. This seems to keep coming up, so I've addressed it in the FAQ for the proposal.

About the calculus for the project

Posted Feb 3, 2012 11:01 UTC (Fri) by fb (guest, #53265) [Link] (38 responses)

Hi,

I would like to point an issue which I missed in this discussion so far.

I will refrain from naming my current and previous employers because they have nothing to do with my opinions.

Regardless of whether SFC sues people over Busybox or not, as a proxy or not, I see the existence of a BSD/Apache licensed package as a positive thing. At my previous job, I wrote 'closed' source code[*]. The (rather large, think Sony-size) company had strict license compliance measures. At least in my division, any software not written internally had to get approval from 'legal' to ship. As it happens only BSD/Apache licensed code would get approval.

People may campaign and cheer the (L)GPLv[23] as much as they wish, but the *reality* today is that many companies will -for whatever reasons- simply not touch anything with GPL on it.

So I see Toybox as a positive development because strictly complying companies may end up having it as a resource, and open-source will IMO reach further than before. People may make a fuss, "oh, but there is no reason not-to-use-the-gpl". In reality writing Toybox seems more doable than changing some written-in-stone legal policies.

[*] I now work full time writing Apache licensed code :-P and yes at my previous job, we did submit every little patch we made on Apache code to "upstream".

[...]

On a personal note, I really never thought I would see a discussion at LWN where people would be arguing in favour of software scarcity.

About the calculus for the project

Posted Feb 3, 2012 18:53 UTC (Fri) by raven667 (subscriber, #5198) [Link] (1 responses)

> People may campaign and cheer the (L)GPLv[23] as much as they wish, but the *reality* today is that many companies will -for whatever reasons- simply not touch anything with GPL on it.

There is certainly some outreach that should continue to be done, sometimes the GPL is rejected because of fear and misunderstanding of what the license actually says and means. I get the sense that the Android core developers have some funny ideas about how the GPL works. Other companies prefer to contribute to GPL projects rather than BSD because the reciprocity means that they are always on a level competitive playing field. Very few business would be comfortable with all of their hard work feeding in a one-way direction to their largest competitor.

About the calculus for the project

Posted Feb 3, 2012 19:37 UTC (Fri) by khim (subscriber, #9252) [Link]

Very few business would be comfortable with all of their hard work feeding in a one-way direction to their largest competitor.

Yup. Google is the process of discovering that. Via Android (see Kindle Fire and Nook Color), via Chrome (see chrome.yandex.ru)... We'll see what it'll do.

About the calculus for the project

Posted Feb 3, 2012 22:11 UTC (Fri) by jimparis (guest, #38647) [Link] (34 responses)

> many companies will -for whatever reasons- simply not touch anything with GPL on it.
...
> In reality writing Toybox seems more doable than changing some written-in-stone legal policies.

Busybox only works with Linux. Linux is GPL. You can't have it both ways -- either you're violating the license of Linux, or your company's legal department is already fine with following the GPL.

About the calculus for the project

Posted Feb 3, 2012 22:38 UTC (Fri) by tbird20d (subscriber, #1901) [Link] (32 responses)

either you're violating the license of Linux, or your company's legal department is already fine with following the GPL

This is not a correct dichotomy. Many companies are fine following the industry consensus interpretation of the GPL, but have difficulty with the interpretation of the GPL by the busybox litigators.

A consistent meme in this discussion is that the motivation for this project is only because busybox is the only actively enforced component. I keep pointing out a different motive - and it keeps getting ignored and rejected, which is a bit frustrating. It's really because some of the interpretations of the GPL (like the death-clause and scope of remedies) by the SFC appear to me to be counterproductive.

I'm entirely comfortable dealing with gplviolations.org, or the kernel community - not because either of these other parties are lax in their enforcement (although admittedly the kernel community appears lax), but because I trust their interpretation of the GPL more, and their reasonableness about rectifying mistakes.

About the calculus for the project

Posted Feb 3, 2012 23:29 UTC (Fri) by jimparis (guest, #38647) [Link] (31 responses)

> A consistent meme in this discussion is that the motivation for this project is only because busybox is the only actively enforced component. I keep pointing out a different motive - and it keeps getting ignored and rejected, which is a bit frustrating. It's really because some of the interpretations of the GPL (like the death-clause and scope of remedies) by the SFC appear to me to be counterproductive.

It is a consistent meme because they hardly seem independent. You are asking us to believe that a difference in "interpretation" is the reason for Toybox, and its creation is in no way related to the fact that the SFLC will sue you when nobody else will. That's just hard to swallow, from the point of view of the open-source community that has traditionally had a very hard time getting companies to follow the license.

> I'm entirely comfortable dealing with gplviolations.org, or the kernel community - not because either of these other parties are lax in their enforcement (although admittedly the kernel community appears lax), but because I trust their interpretation of the GPL more, and their reasonableness about rectifying mistakes.

As you say, they are lax. Neither gplviolations.org nor the kernel community have ever taken anyone to a US court. I'm just not convinced that your trust in "their interpretation" and "their reasonableness" is not at all affected by the fact that they have demonstrated no legal teeth.

Let's say you have a habit of "borrowing" things without permission, never with bad intentions. Of course you'd rather borrow from your kind neighbor who, upon catching you, says "Hey, that wasn't nice" and asks for it back. You don't like borrowing from the other grumpy neighbor who calls the cops right away. You make a big fuss to make sure everyone knows that you got your own lawnmower and will no longer need to borrow the grumpy neighbor's, because hey, he was totally unreasonable to call the cops on you. Then the kind neighbor says "You know, you should just stop borrowing our stuff without permission in the first place" and you say "Come on! That grumpy guy was a total jerk. This has nothing to do with my borrowing, just his interpretation of property law."

If you want a new lawnmower, that's fine. But you shouldn't use it as an excuse to bash on the grumpy guy, and don't be offended when the nice neighbors also use this opportunity to complain about how they're being treated and how you're not doing them any favors.

I'm just a bystander. My point of view may be entirely wrong. Maybe the threat of being sued truly has no effect on your or Sony's views of Busybox. Maybe the SFLC does act unreasonably even when there was a honest mistake somewhere in the supply chain. But I've seen no evidence of that.

About the calculus for the project

Posted Feb 4, 2012 0:29 UTC (Sat) by tbird20d (subscriber, #1901) [Link]

You make a big fuss to make sure everyone knows that you got your own lawnmower

No. I didn't announce this project, Matthew Garret did, for his own purposes. He incorrectly attributed the motivation for this to my employer, which I felt needed correcting. People's own biases have led to a lot of name-calling and fingerpointing and outrage. I'm not trying to rub anyone's face in anything. I've tried to explain my own motivations, and I've been called a liar numerous times.

If you don't trust me, I get it. You can think whatever you like. I would address your mower metaphor, but it seems pointless since people are obviously disposed to distrust what I say, and I've been over it already.

About the calculus for the project

Posted Feb 4, 2012 0:41 UTC (Sat) by dlang (guest, #313) [Link] (27 responses)

> It is a consistent meme because they hardly seem independent. You are asking us to believe that a difference in "interpretation" is the reason for Toybox, and its creation is in no way related to the fact that the SFLC will sue you when nobody else will. That's just hard to swallow, from the point of view of the open-source community that has traditionally had a very hard time getting companies to follow the license.

you may want to actually go look at toybox and it's creator (Rob Landley, a former maintainer of Busybox, and one of the plaintifs in many of the SFLC busybox lawsuits)

Rob has been working on the project since 2006, and it's only in the last few months that he changed the license on it from GPLv2 to BSD. Go back through his blog entries, he has been very vocal about toybox being worked on because he thinks that it is better than busybox. Yes, part of the reason is that he didn't want to give Bruce any possible claim over the code, and he doesn't like GPLv3, but it was worked on because he wanted it to be better than busybox.

Since toybox was a GPLv2 project for several years, it couldn't have been created to make it easy to bypass the GPL

Since toybox was created while Rob was a plaintif represented by the SFLC, it couldn't have been created because he was unhappy with how the SFLC was handling the lawsuits.

About the calculus for the project

Posted Feb 4, 2012 0:46 UTC (Sat) by mjg59 (subscriber, #23239) [Link]

Tim contacted Rob asking him if he'd be interesting in working on a liberally licensed Busybox clone. Rob's clearly working on Toybox for his own reasons, and I think they're all perfectly justifiable. Tim's interest in it is another matter.

About the calculus for the project

Posted Feb 4, 2012 1:10 UTC (Sat) by khim (subscriber, #9252) [Link] (25 responses)

you may want to actually go look at toybox and it's creator (Rob Landley, a former maintainer of Busybox, and one of the plaintifs in many of the SFLC busybox lawsuits)

I have no problems with Rob's work.

Since toybox was a GPLv2 project for several years, it couldn't have been created to make it easy to bypass the GPL.

Of course not. It was only co-opted for that purpose.

The evidence is damning: On an engineering note, they never contributed a single line of code to BusyBox. NOT ONE. I thought they would, but they didn't, so I stopped pursuing them.

What does it say to you? Two things.
1. That future lawsuits will probably not bring new enhancements for busybox.
2. It's not simple to comply with busybox's license. It's trivial.

OEMs don't really change busybox code thus it's enough to just call busybox --version, download tarball from the official site and you can recreate binary.

Now, we have two pieces of software (busybox and Linux kernel). One is 100 times more complex and 1000 times more often modified in practice by OEMs. And someone want to replace the simplest half of the pair - with the explicit goal to STFU the guys who use it as a proxy to get sources for the more interesting part. What does it say about the possible goals?

In fact SONY and others stretched copyright so much that you can argue that project under discussion is doing is most probably flat out illegal. Because in 99% of cases it'll be used as a means to facilitate copyright violation. This fact was enough to kill Naster, why should it be ignored in this case?

About the calculus for the project

Posted Feb 4, 2012 1:25 UTC (Sat) by dlang (guest, #313) [Link] (23 responses)

you say

> I have no problems with Rob's work.

but then you say

> that project under discussion is doing is most probably flat out illegal.

but you are talking about exactly the same project, Rob's toybox project.

which is it, Rob's project is Ok, or Rob's project is illegal?

you can't have it both ways.

About the calculus for the project

Posted Feb 4, 2012 11:21 UTC (Sat) by khim (subscriber, #9252) [Link] (22 responses)

you can't have it both ways.

Yes, I can.

> I have no problems with Rob's work.

Check the first world: I. I, personally, have no problems with Rob's work - but then I'd prefer world where copyright is much less powerful (or where it does not exist at all, as I've already said).

> what that project under discussion is doing is most probably flat out illegal.

But Disneys and SONYs of our world prefer not just strong, but superstrong copyright. They lobby for extensions in congress, they constantly try to straighten it in lawsuits, etc. IMNSHO (as Napster story showed) they already inflated power of copyright so much that that what Tim is doing is most probably illegal because that feature would have no commercial value without the non-infringing product.

but you are talking about exactly the same project, Rob's toybox project.

Nope. Rob's project used GPL and as such for not usable for Napster-style copyright indictment. Because it was just practically impossible to use it without violating both license of kernel and license of toybox. What Tim did is convinced Rob to change license - and this made toybox mostly useful as a facilitator of infringement of Linux kernel copyright.

Right now it's a theory - and not very solid at that. We'll need more evidence before it can be tested in court. If most toybox users will nonetheless diligently ship sources of Linux kernel (this is what Tim explains what is the goal of project: industry guys are quite Ok with GPL compliance and only want bisybox replacement to avoid pesky SFC) - then this theory will evaporate in a puff of smoke. If most users will infringe on Linux kernel copyright then aforementioned theory will be straightened enough to question the legality of initial Tim's work, too.

I just want to point out that SONY, too, can not have it both ways: it can not push to expand copyright power endlessly and at the same time expect that it'll not be affected by said expansion on the other front where it's not plaintiff, but defendant. Right now the coin is in air and how it'll land depends on future actions of chip vendors and suppliers. As well as on the future actions of Tim itself: he can claim that toybox is just first step and chip vendors and suppliers plan to replace Linux kernel, too (this will close this problem once and for all), but then he should show come concrete steps in this direction to convince court that it's all is not just "smoke and mirrors".

BSD-relicensed (that's important) toybox looks very much like Napster-style copyright infringement facilitator.

About the calculus for the project

Posted Feb 4, 2012 19:04 UTC (Sat) by alankila (guest, #47141) [Link] (20 responses)

I'm going to eradicate the context: you are basically advancing a theory which states that relicensing one software from GPL to BSD is illegal because some completely unrelated GPL-licensed software's license can then be violated without legal consequences?

I think such a theory would be laughed out of any court. If licenses are being violated on some piece of software, then let those who have the authority sue if they choose to. If the kernel people do not feel like suing, then it does seem questionable to me that busybox people feel like they have the right to impose settlement conditions on the kernel because of violations on busybox.

Secondly, GPL does not have so much leverage as it once seemed to: there appears to be not-GPL replacements for most if not all the pieces of the system software stack today. I'm sure that changing from Linux to some other kernel would be a tough engineering effort, but scaring enough companies with GPL threats could make them band together and start a replacement effort on some other kernel (bsd?). In the end, if all GPL-licensed software becomes replaced then the legal theory you are advancing will fail too, because there is no GPL-licensed software left whose license could be violated.

About the calculus for the project

Posted Feb 4, 2012 21:22 UTC (Sat) by khim (subscriber, #9252) [Link] (19 responses)

I'm going to eradicate the context:

Which will make the further discussion pointless.

Eben Moglen said it best here:

It generally turns out, as I know from having spent almost a quarter of a century now as a lawyer for hackers, that when hackers pretend to be lawyers, there are certain predictable formulations that they come to; they assume a degree of consistence in legal rules that is not achievable; this is a primary problem which occurs particularly in US focused conversation, such is that in Debian Legal, where the libertarian demand for intellectual consistency, and the hacker belief that laws are form of code that are executed without errors or ambiguities, joins together to create a particular frame of analysis for legal questions.

It doesn't work very well for me as a lawyer, I think it doesn't work very well for lawyers elsewhere in the World, because the one thing which lawyers around the world all share is an awareness of the squishiness of law, it is by no means the hard arthropod carapace for internal soft organs that non-lawyers have a tendency to assume it is....

Context is everything when law is concerned. The very same actions can be interpreted radially differently in court - depending on the context. That's why you can not just grab random pieces of legal advice and use them in court and that's why legal advice is so costly in general: it must be fine-tuned for your particular case.

If licenses are being violated on some piece of software, then let those who have the authority sue if they choose to.

Oh, sure. Only copyright holders of the code in kernel can ever sue Tim. I and you can not (unless you own significant chunk of kernel code, that is). At least that's the situation today. It may change tomorrow.

If the kernel people do not feel like suing, then it does seem questionable to me that busybox people feel like they have the right to impose settlement conditions on the kernel because of violations on busybox.

Sure. But when they will finally decide to sue they may decide to sue not just people who violated their copyright, but the people who've facilitated infringement of said copyright. And if it will be found that toybox is mostly used by sleazy companies to continue to violate kernel's copyright... well, at this point the project in question can found itself in hot water. Note: it all will depend on how far companies will stretch copyright in their efforts to provide adequate legal protection and effective legal remedies against any person knowingly performing without authority any of the following acts knowing, or with respect to civil remedies, having reasonable grounds to know, that it will induce, enable, facilitate, or conceal an infringement of any copyright or related rights. Looks like Disneys and SONYs of the world are quite serious which means that while what Tim is doing today is still legal (even if morally questionable) it may become illegal tomorrow!

Secondly, GPL does not have so much leverage as it once seemed to: there appears to be not-GPL replacements for most if not all the pieces of the system software stack today.

Indeed. As I've said already if toybox will be mostly used in combination with some kind of BSD kernel then the aforementioned theory will evaporate in a puff of smoke. We'll see.

In the end, if all GPL-licensed software becomes replaced then the legal theory you are advancing will fail too, because there is no GPL-licensed software left whose license could be violated.

Sure. If the project in question will be expanded to cover all the GPL-licensed code typically used by shady contractors and if GPL code will be eradicated as result then it'll be different story entirely. But so far it does not look like port of toybox to other kernels is even considered - and this puts the whole effort squarely in the "copyright facilitator" bin.

About the calculus for the project

Posted Feb 5, 2012 11:12 UTC (Sun) by rahvin (guest, #16953) [Link] (10 responses)

An interesting argument I hadn't considered. In the interest of adding to your supposition I'll add the following. The theory of copyright infringement facilitation without engaging in actual infringement yourself has been used successfully in court several times in the US.

As you mention the first use was the Napster case that resulted in the entire company getting shut down. But there have been other cases, such as Limewire, Grokster and several other smaller players. In all the previous cases it was a civil lawsuit and involved the argument that centered around contributory copyright infringement and resulted in millions of dollars of damages. Most of the cases were appealed all the way to the Supreme and as such have established a fairly solid legal citation in future cases (baring a future reversal by the Supreme).

But more recently the MegaUpload cases presents a dramatic escalation in this with a criminal proceeding with racketeering charges tied in where the US government is seeking multiple years of jail time for the parties responsible. The theory you've presented is interesting because it's a well established precedent in US courts that contributory copyright infringement is a real and punishable offense.

The court verdicts to date have basically established the precedent that if the intent and use of the product was for the primary purpose of infringing copyright, even if there are legitimate uses that make up a small part of the use case, that the parties responsible are guilty of contributory copyright infringement for the behavior of their users. You have to wonder if these verdicts couldn't be stretched to extend to this situation. Although it'll be argued the intent wasn't to encourage GPL infringement (but to avoid a SFC legal action) by the people involved it should be fairly obvious that regardless, a significant number of people in the community and on LWN have interpreted the action to be an attempt to facilitate infringement and that this entire discussion would likely be used in the case, particularly the comments by those involved.

Would a GPL case extend to this level? I doubt it, but as you say history doesn't predict the future. With laws like SOPA in play an ambitious district attorney looking to get their name in the news could litigate a GPL case without the consent of copyright holder (one of the many scary clauses in the law).

An interesting hypothesis no doubt, thanks for posting it.

About the calculus for the project

Posted Feb 5, 2012 15:00 UTC (Sun) by khim (subscriber, #9252) [Link] (9 responses)

Now you can understand why I clearly distinguish Rob's work and Tim's work.

Rob just wrote yet another replacement for busybox - may be technically interesting, but legally pretty plain project. It used the same GPL which both busybox and Linux kernel are using thus it was not all that interesting from the legal perspective: yes, SFC had no way to enforce it's copyright, but thus was just Rob's goodwill, it was easy for him to change his mind at any time. Hardly a reason for a hypothetical sleazy company to be excited. And hardly anybody was excited: toybox was around for years but was mostly a curiosity not worth talking about much.

Tim, on the other hand, convinced Rob to change license and make it as close to public domain grant as possible. Now it becomes much easier to infringe on kernel's copyright. Tim even admits it's directly:
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. As part of their request to remedy a busybox GPL violation, the SFC does ask for source code unrelated to busybox. Personally, I believe this is improper. However, my main reason for proposing this project is to avoid having the SFC gain review authority over unrelated products produced by a company. The larger the set of Linux-based products that are produced by a company, the greater exposure there is for a possible mistake, and the greater potential costs that would incur in the event of litigation and/or settlement.
Of course he also continues with rhetoric that his goal is not to help copyright violators, etc, but... how much it helped other companies you've mentioned here?

About the calculus for the project

Posted Feb 10, 2012 0:18 UTC (Fri) by landley (guest, #6789) [Link] (8 responses)

No, Tim asked me if I knew anybody who'd want to work on extending Android's crappy Toolbox so people wouldn't need to replace it with software that can never ship as part of Android, due to Google's existing "no GPL in userspace" policy which most android vendors inherit:

http://source.android.com/source/licenses.html

_I_ am the one who brought up Toybox, and pointed out I already _had_ something that was way the heck ahead of Toolbox, and which I'd love an excuse to work on again.

And I have already been talking about the smartphone replacing the PC since BEFORE ANDROID EVEN EXISTED:

http://landley.net/notes-2010.html#09-10-2010

(The 2002 reference in that is about when I wrote an article on the topic for a now-defuct wesite called linux and main, I might be able to pull the article out of archive.org if I try. I note that Eric's objection was ergonomics would limit the usability of palm pilot and newton and such. There weren't any USB docking stations yet, but it seemed obvious to me that the problem would be solved, and that internet-through-cellphone was _compelling_. Yes, in 2002. And yes, I hit bad bufferbloat the first time I tried getting internet through my cell phone on Linux in 2001, I just had no idea what to _do_ about it and assumed I'd configured things wrong.)

I'm the one who made the connection that "Oh, if toybox was BSD licensed _that_ would give it a separate niche from busybox giving me an excuse to work on it without feeling like I was wasting my time, because android is a deployment platform that only bsd licensed code can fit in. And I could use that as leverage to steer android towards a more full-fledged posix environment, and from there it's a simple step towards leveraging my aboriginal linux project to make android self-hosting, and then android would have a leg up against the iPhone, which it's been losing to recently".

The enthusiasm was _mine_, because I saw a path forward I'd missed before, which Tim's agenda was just a tiny part of. Tim was actually going a different direction with it, by I got an idea talking with him.

It was an idea that FSF zealots really don't like, but they lost me in 2006:

http://landley.net/notes-2006.html#03-12-2006

And I've been drifting _away_ from the GPL ever since. Before you blame Tim for my "screw it, BSD license the sucker" decision, read this bit almost TWO YEARS EARLIER:

http://landley.net/notes-2010.html#10-01-2010

> Since uClibc's so obviously moribund, I'm vaguely pondering poking at
> Android's "bionic" project instead, and maybe even their "toolbox"
> project. The first is their libc (uClibc replacement) and the second is
> their coreutils (busybox replacement). They wrote their own because they
> decided they want all the userspace stuff to be BSD licensed, no GPL in
> userspace. Before GPLv3 came out I would have objected to this, but now
> it seems kind of reasonable. (If the FSF is really _that_crazy_, then
> you don't want to get any of it on you.)

And goes on to talk about the merits and drawbacks of BSD licensing.

This was me analyzing some more of Android's licensing decisions a year before I relicensed Toybox:

http://landley.net/notes-2010.html#09-10-2010

So by all means, consider me some kind of strange puppet for Tim that's unable to form any independent thoughts about licensing, let alone make informed decisions like an adult.

About the calculus for the project

Posted Feb 10, 2012 10:16 UTC (Fri) by khim (subscriber, #9252) [Link] (7 responses)

If this is all about Android then why the page was started with talks about GPL infringement facilitation? It still talks about this aspect first and about potential usage in Android second.

Better userspace for Android will be easy sell. GPL infringement facilitator is not. Even if technically it's the same project.

P.S. Actually userspace for android is slightly different project because to support Dalvik you need different set of tools then to support traditional POSIX userspace.

About the calculus for the project

Posted Feb 10, 2012 16:19 UTC (Fri) by landley (guest, #6789) [Link]

"We'd like to get as many patents as possible to protect ourselves from the insane patent system via mutually assured destruction and cross-licensing agreements."

"We'd like to get an alternative to BusyBox to protect ourselves from the insane out of control SFLC/SFC and their demands about third-party binary only kernel modules and vendor toolchains we've never even seen the source code to if we shipped a vanilla unmodified busybox but didn't SAY we shipped a vanilla unmodified busybox."

"We'd like to distance ourselves from this legal quagmire rather than try to navigate an endless shifting maze."

That's not _my_ motivation, but I can certainly see spraying for lawyers holding an appeal for people, especially when legal stuff isn't what they _do_. And if it makes 'em more likely to use the code I already _wrote_ and am still adding to (which is technically better than the alternative)? Sure, why not?

Rob

About the calculus for the project

Posted Feb 10, 2012 16:22 UTC (Fri) by landley (guest, #6789) [Link] (5 responses)

By the way, another analogy for all the insane GPL zealots to ponder:

Don't put lifeboats on your cruise ship. Make it unsinkable. By planning for what happens if the boat sinks, you're planning to sink the boat! Shame on you, lifeboats are EVIL!

Rob

About the calculus for the project

Posted Feb 10, 2012 17:21 UTC (Fri) by khim (subscriber, #9252) [Link] (2 responses)

You are not putting lifebots on a ship. You are adding couple of lifebots to Titanic.

This makes no sense whatsoever. If you don't expect for it to be sunk then why bother with lifebots at all? If you do expect crash and burn scenario then why do you plan to save just a handful of people, not all of them?

About the calculus for the project

Posted Feb 11, 2012 14:35 UTC (Sat) by Wol (subscriber, #4433) [Link] (1 responses)

Actually, that's not fair on Titanic ...

She had FAR more lifeboats than were legally required.

And could have saved a lot more people, if they hadn't messed up evacuating the boat. Plus a load of people didn't want to get on the lifeboats because they felt safer staying on board (pretty much the same as happened in the recent cruise-ship sinking in Italy ...)

(I think the law nowadays says you need enough spaces on EACH SIDE of the boat for the entire passenger/crew complement - ie 2 spaces per person. Most boats carry even more. It makes sense, actually, if the boat is listing you can lose half your usable lifeboat capacity in the space of seconds.)

Cheers,
Wol

About the calculus for the project

Posted Feb 11, 2012 15:39 UTC (Sat) by khim (subscriber, #9252) [Link]

Actually, that's not fair on Titanic ...

She had FAR more lifeboats than were legally required.

Sure. Law required just 16 lifeboats while Titanic had 20. With total capacity of 1,178 people on a ship with capacity of 3,547. Less then ⅓ space per person. Good idea. Not.

And could have saved a lot more people, if they hadn't messed up evacuating the boat. Plus a load of people didn't want to get on the lifeboats because they felt safer staying on board (pretty much the same as happened in the recent cruise-ship sinking in Italy ...)

That's another issue. Yes, even if you have enough lifeboats you still can lose a lot of lives in a case of crush. But if you don't have enough lifeboats to save all passengers then loss of life is inevitable.

I think the law nowadays says you need enough spaces on EACH SIDE of the boat for the entire passenger/crew complement - ie 2 spaces per person. Most boats carry even more. It makes sense, actually, if the boat is listing you can lose half your usable lifeboat capacity in the space of seconds.

Sure. More then one space per person may be good idea. How much more is subject of the debate. Less then one space per person is pointless - and this is exactly how Titanic was equipped.

About the calculus for the project

Posted Feb 12, 2012 15:51 UTC (Sun) by nix (subscriber, #2304) [Link] (1 responses)

btw, you might stand more chance of convincing people rather than repelling them if you didn't routinely call them insane. Just a thought.

About the calculus for the project

Posted Feb 13, 2012 8:27 UTC (Mon) by bronson (subscriber, #4806) [Link]

This is true. But I don't think Rob is calling all GPL zealots insane...? That's just a tiny subset.

About the calculus for the project

Posted Feb 5, 2012 12:25 UTC (Sun) by alankila (guest, #47141) [Link] (3 responses)

Thank you for a thoughtful reply. I personally do hope that people would not find this argument reasonable way to think about licenses, because in general case any piece of software that happens to eliminate another piece of GPL software could be argued to help facilitate GPL infringement on the remaining software.

But as you point out, law is essentially insane and random, and stranger things could happen. My personal feeling, however, is that this is not fair nor right, whatever the courts might find.

About the calculus for the project

Posted Feb 5, 2012 15:14 UTC (Sun) by khim (subscriber, #9252) [Link] (2 responses)

I personally do hope that people would not find this argument reasonable way to think about licenses, because in general case any piece of software that happens to eliminate another piece of GPL software could be argued to help facilitate GPL infringement on the remaining software.

Nope. Again: context matters. As rahvin already wrote: The court verdicts to date have basically established the precedent that if the intent and use of the product was for the primary purpose of infringing copyright, even if there are legitimate uses that make up a small part of the use case, that the parties responsible are guilty of contributory copyright infringement for the behavior of their users. Not only you should show that someone picked your piece of software because it's under BSD license (that's understandable: less rules to follow), you must show another piece of GPLed code which is commonly used with first one and show that most users of your software not only continue to use it, but also do it in violation of it's copyright. Tall order, indeed.

But as you point out, law is essentially insane and random, and stranger things could happen. My personal feeling, however, is that this is not fair nor right, whatever the courts might find.

No questions here. I also feel that copyright's power is inflated way, way, WAY beyond any reasonable limits. But... dura lex sed lex. And since Disneys and SONYs of the world were instrumental in said insane inflation it's only proper to apply no mercy, as much punishment as possible principle to them.

About the calculus for the project

Posted Feb 5, 2012 23:13 UTC (Sun) by anselm (subscriber, #2796) [Link] (1 responses)

Not only you should show that someone picked your piece of software because it's under BSD license (that's understandable: less rules to follow), …

What about the idea of someone picking the BSD-licensed software simply because it was better code than the GPL software (i.e., written by somebody with several years of experience with the GPL codebase, taking into account any design and implementation issues since discovered with the GPL codebase etc., and better targeted towards solving the actual problem that users of the software need solved)?

In that case it would be difficult to argue that the »intent and use« of a BSD-licensed Busybox workalike would be »for the primary purpose of infringing copyright«. Maybe its intent and use is to provide a technically superior all-in-one shell-type tool, and its users prefer it to Busybox on these grounds (with the more liberal license as a fringe benefit). This isn't Napster, after all.

Given this, it would be next to impossible (for someone like the SFC) to show that somebody picked the BSD-licensed software exclusively on licensing grounds, and claiming that this in turn was only done to avoid GPL compliance issues concerning the Linux kernel would border on a conspiracy theory. At the very least, someone like the SFC would have to prove that the workalike was clearly less suited, technically, for its intended use (by the defendant, not the FLOSS community in general) than the original GPLed Busybox, and that the defendant still went for the workalike just to be able to do an end-run round the SFC. Have fun.

About the calculus for the project

Posted Feb 6, 2012 0:33 UTC (Mon) by khim (subscriber, #9252) [Link]

I never said it's easy to probe illegality of this work. In fact just a dozen years this question was flat out crazy: it was obvious that if there no copyright violation then there are no material for lawsuit. And as I've already said: it's hard to say a-priori which way the development will go. In fact there already exist alternative BSD-licensed userspace fro Linux: Android's one (albeit in this case it's pretty clear that it's goal was to support Java-based environment and not to act as GPL-free replacement for GNU tools). And if companies that use toybox will be postly good citizens then we'll not a problem: as was shown many times already it's not about busybox or toybox.

But Disneys and SONYs of the world are hard at work - and they continue to extend the reach of copyright in the vain hope to squeeze more money from 100 years old creations.

If enough guys from enough companies will blab out that they are switching to toybox to solve "SFC problem" then you'll have materials for lawsuit. It'll not matter if toybox is inferior or superior. If proportion of GPL-violating companies will be high enough among toybox users and low enough among busybox users and you'll have direct evidence for a few "switchers" - you'll probably have enough for the court in the current world, where copyright importance is inflated beyond any reason. And in tomorrow's world it may be even easier.

About the calculus for the project

Posted Feb 9, 2012 23:44 UTC (Thu) by landley (guest, #6789) [Link] (3 responses)

I'm the ex-maintainer of busybox, who started toybox before I actually did the last release of busybox. I'm the one who _started_ the busybox license enforcement lawsuits (and recruited Erik Andersen in into it, and yes I still have my old email).

You're claiming that my new project might be considered an infringement facilitator AGAINST MY OLD PROJECT, and that stopping the lawsuits I _started_ might somehow legally do... something?

Your honestly believe that the legal system can and should forbid people from doing new things that compete with (and obsolete) their own previous work? And you take this position in defense of "freedom", as defined by the FSF?

I ask seriously: are you crazy? Do you need some kind of diagnosis and treatment? (I already _know_ the FSF does...)

About the calculus for the project

Posted Feb 10, 2012 10:06 UTC (Fri) by khim (subscriber, #9252) [Link] (2 responses)

You're claiming that my new project might be considered an infringement facilitator AGAINST MY OLD PROJECT, and that stopping the lawsuits I _started_ might somehow legally do... something?

Oh yea. Absolutely. This is what this lawsuit is all about. Oracle even explicitly claimed that the fact that the very same people are working on Dalvik that worked on Java decades ago increases Google's liability. Or perhaps you live in some other world where such things never happen?

Your honestly believe that the legal system can and should forbid people from doing new things that compete with (and obsolete) their own previous work?

That's different question. I know that the legal system can and does forbid people from doing new things that compete with their own previous work. I'm not saying it's good thing - that's another thing entirely.

And you take this position in defense of "freedom", as defined by the FSF?

Law exist without FSF's freedom. Government changes it, not FSF.

I've pointed that what you are doing is probably illegal and that you can be sued for doing it in the future. This is almost certainly true. But I'm not saying that this is what FSF should do. Just because you can sue someone does not mean you should. That's different story altogether.

I ask seriously: are you crazy? Do you need some kind of diagnosis and treatment?

Me? I'm absolutely sane. The copyright world is crazy - but sadly I don't know how to cure it.

About the calculus for the project

Posted Feb 10, 2012 17:15 UTC (Fri) by landley (guest, #6789) [Link] (1 responses)

> That's different question. I know that the legal system can and does
> forbid people from doing new things that compete with their own previous
> work.

Not in California or Texas it doesn't. ("Right to work" states.) There's rather a lot of case law on this. Silicon Valley was _built_ on the "traitorous eight" from Shockley Semiconductor.

http://en.wikipedia.org/wiki/Traitorous_eight

Honestly, if you work in the industry you get to know this stuff...

Rob

(Oratroll is suing over patents, sounds like they're attempting to establish treble damages for wilful violation, ala http://www.ipwatchdog.com/patent/advanced-patent/patent-i... . Which is in fact the reason Linus gives for not reading software patents.)

About the calculus for the project

Posted Feb 10, 2012 17:38 UTC (Fri) by khim (subscriber, #9252) [Link]

http://en.wikipedia.org/wiki/Traitorous_eight

Honestly, if you work in the industry you get to know this stuff...

Sure. These tales are very well-known. But that happened half-century ago. Companies worked diligently to rectify "the problem" all that time. Today the outcome of such lawsuit is quite different to predict.

Oratroll is suing over patents

Nope. Copyrights are very much in dispute, too. And outcome is not yet certain. And I'm sure if Oracle will be repelled this time then it'll cooperate with Disneys and SONYs of the world to change the law to make sure next time they'll win.

In fact the whole SFC "problem" have one trivial and quite logical solution: change the law, reduce copyright power and make what SFC is doing illegal. Sadly Disneys and SONYs of the world reject this solution out of hand because it'll reduce their copyright power, too. Well, as long as that's the case don't expect me to feel any sympathy to their plight. They brought it on themselves.

About the calculus for the project

Posted Feb 9, 2012 23:36 UTC (Thu) by landley (guest, #6789) [Link]

> BSD-relicensed (that's important) toybox looks very much like
> Napster-style copyright infringement facilitator.

"It's not impossible, I used to bullseye womp rats in my T-16 back home. They're not much bigger than two meters."

I wrote the toybox code, I own the copyright on the code, I issued a second license to my copyrighted code. (A license is a permission statement. The older versions were released under GPLv2 and that was the only set of conditions you had permission to use it under. I have since granted additional permission to use my copyrighted code under less restrictive terms. Legally, the two are unrelated.)

As the guy who _started_ the Busybox licence enforcement suits, as someone who spent a couple years paid by IBM's lawyers as a domain expert to help defend them against legal claims, as someone who had an email exchange with Richard Stallmand in the 90's about nuances of the GPL for a "copyright and licensing HOWTO" that eventually turned into a week-long series of articles on the five types of intellectual property for The Motley Fool:

http://www.fool.com/portfolios/rulemaker/2000/rulemaker00...
http://www.fool.com/portfolios/rulemaker/2000/rulemaker00...
http://www.fool.com/portfolios/rulemaker/2000/rulemaker00...
http://www.fool.com/portfolios/rulemaker/2000/rulemaker00...
http://www.fool.com/portfolios/rulemaker/2000/rulemaker00...

As somebody who has studied intellectual property law as a hobby since the 1990's, as someone who has spoken on panels discussing the GPL when the other panelists were professional practicing lawyers, as someone who was once flown to new York to speak on a panel with Eben Moglen (the lawyer co-author of the GPL) about GPL enforcement.

As all that and more, I tell you:

No, it's not "important". You have absolutely no _clue_ what you're talking about, and I don't appreciate the transparent FUD.

About the calculus for the project

Posted Feb 9, 2012 23:17 UTC (Thu) by landley (guest, #6789) [Link]

> What does it say to you? Two things.
> 1. That future lawsuits will probably not bring new enhancements for
> busybox.

No, it says that past ones didn't. I cut my ties at the end of 2008, I haven't seen anything new show up in the busybox repository or on the mailing list that's attributable to lawsuits since, but to be honest I haven't been paying _that_ close attention. Outside of submitting my patch implementation, fixing busybox vi's horrible serial handling, trying to convince Denys to adopt some of toybox's better infrastructure (such as having all information about a command be in one file per command), and some trivial cleanups to demonstrate the the code doesn't HAVE to be #ifdef salad (I.E. "this is what I added the ENABLE_BLAH and USE_BLAH() macros for 5 years ago, you're using them wrong")... I haven't really paid that close attention to the busybox list. I catch up when I move to a new release, get involved in a thread or two, and then wander away again.

I still seem to have an impact when I do show up. My last exchange was:

http://lists.busybox.net/pipermail/busybox/2011-November/...

Which led to me fairly randomly advocating "some things are better as as shell scripts, not in C":

http://lists.busybox.net/pipermail/busybox/2011-November/...
http://lists.busybox.net/pipermail/busybox/2011-November/...
http://lists.busybox.net/pipermail/busybox/2011-November/...

And then I wandered away again, but this morning I noticed, in year 13 of the project we now have:

http://git.busybox.net/busybox/commit/?id=d0222503ff9ff26...

(My rsync script which syncs my laptop up to my web server also does a git pull on various repositories, and busybox is still in there...)

Sheer coincidence, of course. But given a choice, I'd rather put my energy into Toybox than Busybox. When I worked on busybox, I wanted it to spread up into the desktop and become the standard command line of things like Knoppix while remaining small and simple. Neither happened. Now I want toybox to spread down into android and ride the smartphone up as it displaces the PC the way the PC displaced the minicomputer, and that displaced the mainframe before it.

And before you say "that's impossible", google "Toshiba Dynadock" and imagine what plugging a USB docking station into an android phone containing a native compiler actually _means_...

Rob

About the calculus for the project

Posted Feb 4, 2012 14:30 UTC (Sat) by nix (subscriber, #2304) [Link]

As you say, they are lax. Neither gplviolations.org nor the kernel community have ever taken anyone to a US court.

Uh, that would be because gpl-violations.org operates in Germany. Harald Welte is a kernel hacker, so members of the kernel community certainly have taken people to court -- just not US court.

About the calculus for the project

Posted Feb 9, 2012 22:54 UTC (Thu) by landley (guest, #6789) [Link]

> You are asking us to believe that a difference in "interpretation" is the
> reason for Toybox, and its creation is in no way related to the fact that
> the SFLC will sue you when nobody else will.

No, Toybox started in 2006, because a troll showed up to take credit for all my work ten years after he last wrote any code (for anything), and disgusted me so hugely I left busybox entirely rather than let him take credit for one more LINE of code I wrote:

http://landley.net/notes-2006.html#26-09-2006

But I still wanted to work on small, simple utilities (I had several half-finished, and I write this sort of code for fun), so I took the excuse to start over from scratch without the decade of legacy design decisions BusyBox had accumulated. Instead of just sulking, after I withdrew from BusyBox I started a new project and sat down to write better code:

http://landley.net/notes-2006.html#01-10-2006

I continued to work on it, on and off, for the next 3 years, until I'd solved all the interesting design problems and knew how the new project should be implemented. But I had a bunch of other demands on my time (aboriginal linux, my tinycc fork, day job), and I had the problem with Toybox: the niche it was going for was already filled. By BusyBox. I'd appointed its current maintainer and handed years of my work off to him, I was going up against something with a 10 year head start, led by a smart active guy, including years of my own work. It was "good enough", merely being technically _better_ wouldn't necessarily draw users away from it. It had a greater "project gravity", and is what people would reach for first to solve the problem both projects addressed.

I started to consider rejoining BusyBox after about 2 years:

http://landley.net/notes-2008.html#30-06-2008

And eventually I mothballed ToyBox and started pushing my work "upstream" into busybox (which was actually a huge port since the point was to base from-scratch rewrites on completely different infrastructure):

http://landley.net/notes-2010.html#11-03-2010

The biggest chunk of infrastructure was my new patch implementation, because BusyBox's was more or less a stub. But I also met denys at CELF 2010 and explained toybox's config and help infrastructure (where all you do to add a command is add a toys/mynewcommand.c file and then re-run the build and it picks everything up from that; in BusyBox you had to touch 5 different files), and so on.

But I liked the toybox infratructure much better than busybox's, even with what cleanups I managed to convince Denys to do. And unfortunately, BusyBox had lost focus on "simple", and moved it down below "small binary size", "fast execution time", and "lots of features". The resulting code was #ifdef salad full of magic annotations and control flow that hid the program's actual entry point in libbb/appletlib.c line 771. The result was just no fun to work on:

http://landley.net/notes-2011.html#08-06-2011

For a couple years I vaguely tried to learn Lua to rewrite toybox in that as a way of giving it a reason to exist separate from busybox (and thus a reason for me to still work on it):

http://landley.net/notes-2009.html#26-02-2009

But lua turns out not to have a standard posix binding, making rewriting toybox in it kinda hard. (If I have to include native C code to extend lua, I might as well write the whole thing in C. Being in a higher level language eliminates the need for cross compiling, but being in a higher level langauge with C extensions _doesn't_.)

So toybox was in limbo for a couple years, with the bits I really _needed_ (for my aboriginal linux project, creating the smallest Linux development environment capable of rebuilding itself from source) pushed into busybox, but busybox being no fun to work on and not seeing much point in pushing toybox up a hill busybox was already at the top of...

Then when Tim contacted me in November, what he did was point out that there _was_ a niche for ToyBox: on Android, which by policy forbids GPL in userspace but has only a stub command line that's crying out to be replaced.

I've considered Android very important for a while now, because it _will_ take over from the PC:

http://landley.net/notes-2011.html#26-06-2011

And Linux won't:

http://landley.net/notes-2010.html#10-03-2010

I just hadn't played with its native development environment much after a bad early experience:

http://landley.net/notes-2010.html#10-01-2010

So when Tim went "android" (and pointed out the AOSP is much less stupid now than it was 2 years ago) I went "D'oh" and started scheming:

http://landley.net/notes-2011.html#13-11-2011

I'd never have left GPLv2 behind if it wasn't for GPLv3. (I seriously doubt Android would have forbid GPL in userspace either, v2 got tarred with the same brush as v3.) I loved GPLv2:

http://sf.geekitude.com/content/pros-and-cons-gnu-general...

But I'm not the one who left the GPL, it left _me_:

http://landley.net/notes-2006.html#03-12-2006

Rob

About the calculus for the project

Posted Feb 3, 2012 22:40 UTC (Fri) by raven667 (subscriber, #5198) [Link]

Or the position they seem to be taking, IIUC, that GPL enforcement for Linux is NotTheirProblem(tm).

About the calculus for the project

Posted Feb 3, 2012 23:56 UTC (Fri) by marcH (subscriber, #57642) [Link]

The reality today is that every other company in the world is using GPL software, most notably Linux

http://www.h-online.com/open/features/We-won-and-we-didn-...

Companies which exclude any piece of GPL'ed software without even considering it are obviously restricting their options, which is unlikely to make them more competitive.

> ... and open-source will IMO reach further than before.

Whether BSD or GPL is more efficient at producing open-source has been debated 5 zillions times already, and the answer is certainly not as obvious as you make it sound.