Winning the copyleft fight [LWN.net]

For humans, by humans
Every article on LWN.net is written for humans, by humans. If you've enjoyed this article and want to see more like it, your subscription goes a long way to keeping the slop at bay. We are offering a free one-month trial subscription (no credit card required) to get you started.

By Jonathan Corbet
February 12, 2016

linux.conf.au 2016

Bradley Kuhn started off his linux.conf.au 2016 talk by stating a goal that, he hoped, he shared with the audience: a world where more (or most) software is free software. The community has one key strategy toward that goal: copyleft licensing. He was there to talk about whether that strategy is working, and what can be done to make it more effective; the picture he painted was not entirely rosy, but there is hope if software developers are willing to make some changes.

Copyleft licensing is still an effective strategy, he said; that can be seen because we've had the chance to run a real-world parallel experiment — an opportunity that doesn't come often. A lot of non-copyleft software has been written over the years; if proprietary forks of that software don't exist, then it seems clear that there is no need for copyleft; we just have to look to see whether proprietary versions of non-copyleft software exist. But, he said, he has yet to find a non-trivial non-copyleft program that lacks proprietary forks; without copyleft, companies will indeed take free software and make it proprietary.

As an example, he noted that the version of OpenStack running behind his Rackspace account clearly has features that are not in the free version. He does not know of a single OpenStack product that does not contain at least a few features that have not been pushed back upstream.

There is more and more non-copyleft software out there, he said, but copyleft itself remains the most viable strategy to bring about software freedom. Other strategies can be employed, but we need to continue with copyleft as the centerpiece. Unfortunately, it is not working as well as we would like for one simple reason: there is less copyleft software being produced.

Corporate opposition to copyleft

There is not much support for copyleft in the corporate world; Martin Fink's recent LinuxCon Europe talk was the first positive words he's heard from a company about copyleft in quite a while. In general, though, companies are embracing free software; even Apple is, to an extent. They have discovered that they cannot beat the free-software community, so they have now switched to an approach of coopting that community. The result is the emergence of groups claiming to speak for the community, saying that copyleft is no longer wanted. He has heard developers say that users prefer permissive licenses and that it's impossible to get funding for work on copyleft-licensed software. Why are they saying that? In the end, Bradley said, if you hear things enough times you may well start to believe them.

As one might imagine, he sees the lack of enforcement for the code that is under copyleft licensing as a big problem. The only people who can change that, he said, are developers who hold their own copyrights in the code. Companies, which hold most copyrights at this point, will never do enforcement. Or, on the rare occasion when they do, it is to support their own goals rather than to defend the freedom of the software. One example is MySQL, which was happy to use the GPL as a way to force companies to buy proprietary licenses. Red Hat alleged GPL infringement against Twin Peaks Software as a way of defending itself against a patent troll; in the end, Red Hat got its patent license, and Twin Peaks Software's code remains proprietary.

The Software Freedom Conservancy recently published a set of principles that, they say, should govern GPL enforcement; the first of those is that the primary objective is to bring about license compliance. Since then, no trade association or company has endorsed those principles.

That is the case, he said, because lawyers in the corporate world have gained control over policy with regard to the GPL, and they are actively coordinating to try to push the GPL aside. They hold "secret meetings" in which to work through scenarios of possible enforcement actions and how to prevail against them. There have been "two meetings" with the purpose of finding ways to increase corporate ownership of GPL-licensed software as a way of curtailing enforcement by individuals.

The result, he said, is a crisis for the concept of copyleft. If copyleft is the right strategy, then some way has to be found to deal with these challenges. We need a plan.

Defending copyleft

The copyleft strategy works best when copyrights are held by people and institutions that hold software-freedom principles. That means individuals and charities, he said; companies and trade associations do not subscribe to those principles. Thus, he said, developers who support software freedom should go to their employers and insist on being allowed to hold their own copyrights. Perhaps developers should unionize to demand this right. Then they should join one of the Conservancy's coalitions for enforcement.

He also called for developers to go back to volunteer coding — off hours, it is not necessary to quit one's job to do this (though he did note that some companies will claim ownership of even an employee's off-hours work). Early free software was written this way. Bradley said that he works nights and weekends for the cause; he would like for the development community to do the same. Developers should also be willing to fork permissively licensed software under copyleft when necessary.

But copyleft won't work if we don't enforce the terms of the license, he said. Enforcement is getting harder: companies are actively working to replace copyleft-licensed software. Toybox and LLVM are two examples there. Given that, which essential copyleft-licensed software remains?

The one thing that companies haven't tried to rewrite, he said, is the Linux kernel. Thus Linux is the battleground over which the future of copyleft licensing will be determined. An area of immediate interest is proprietary modules, and whether they are derived works of the kernel; he has not yet seen one that didn't look that way to him. There are corporate lawyers who want to fight over this issue; it is a fight that the community will have to take on. If we shrink from this particular confrontation, we will no longer have effective copyleft licensing.

Once upon a time, he assumed that most license violations were innocent mistakes. Those days are coming to an end. Most violations, he alleged, are done deliberately by companies that hire lawyers in advance to plan strategies should it come to a fight. The community is helping these companies with its openness. This is, he said, a zero-sum game; the two sides have diametrically opposed goals.

GPL enforcement requires more than copyright holders, though; it also requires funding. That is a problem; enforcement as a profit-making enterprise is a path to corruption. Violators will always ask how much it will cost to buy permission, but if owners take money without demanding full compliance, they are just another player in the proprietary relicensing business. Companies will not support this work, so the only way is for the community to support it. He reminded the audience of the Conservancy's funding drive, which continues through February.

He concluded by saying that his plan can easily be described as simple-minded. It depends on a community of individuals coming together to stand up for software freedom and the GPL. But this should be possible: uniting and working toward a common goal is just what the community is good at doing.

The video for this talk is available on the LCA site.

[Your editor thanks LCA for assisting with his travel expenses.]

Index entries for this article
Conference	linux.conf.au/2016

to post comments

Winning the copyleft fight

Posted Feb 12, 2016 20:55 UTC (Fri) by xtifr (guest, #143) [Link] (44 responses)

I am somewhat surprised that more companies don't see the benefits of the GPL, when applied to their own code! It may not be quite the same set of benefits that I see as an individual developer, but the ability to keep your competitors from leveraging your own code and competing against you without sharing their improvements to (again) your own code has to be something worth considering.

Frankly, instead of complaining about the way companies are leveraging the GPL, I think Kuhn should point to those as GPL success stories! No, they're not all-focused on making all software free, but they're gaining clear benefits from GPLing their code, which still benefits us all in the longer run. These companies should be encouraged and praised; they provide a clear example of the positive nature of the GPL/copyleft, even to those who don't necessarily share our ideals.

As for the companies (particularly Apple), that are starting to use "pushover" licenses to attack the GPL, I think it's worth noting that software under an MIT/BSD/Apache license can be sublicensed under the GPL as easily as it can a proprietary license! For years, there's been an uneasy detente between the copyleft folks and the BSDers, and some people will claim its unethical to sublicense BSD code under the GPL. In general, I would agree, but when the BSD or similar license is being used to attack copyleft, I think that changes things, and I would have no problem contributing to a GPL'd fork of LLVM. In fact, I'd be proud to do so!

Winning the copyleft fight

Posted Feb 12, 2016 23:14 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link] (32 responses)

A nice way to split community.

Anyway, it's also a nice way for people to waste time and for a project to die. Case in point: https://github.com/mailpile/Mailpile/graphs/contributors?...

Mailpipe switched to AGPL in Jan 2015. And after that new contributions not made by its principal author pretty much died away. If the author loses interest, then the project is going to be 0xDEAD.

Winning the copyleft fight

Posted Feb 13, 2016 1:40 UTC (Sat) by xtifr (guest, #143) [Link] (4 responses)

A nice way to split community.

What Apple's doing? I agree, except that I think "nice" may not be quite the right word. Dickish seems more appropriate.

Anyway, it's also a nice way for people to waste time and for a project to die.

The way LibreOffice died the moment AOO came along? ;) I wasn't necessarily making a serious suggestion, but I would like to point out that such a project, if it did appear, would have the big advantage that LibreOffice also had—patches can only flow one way.

Winning the copyleft fight

Posted Feb 13, 2016 1:44 UTC (Sat) by bronson (subscriber, #4806) [Link] (3 responses)

You're claiming that Apple is splitting the GCC community?

Winning the copyleft fight

Posted Feb 17, 2016 1:10 UTC (Wed) by rahvin (guest, #16953) [Link] (1 responses)

Apple has a history of taking BSD licensed code and continuing to improve it while creating proprietary forks. For example, the Apple kernel is mostly permissively licensed but forms the core of a proprietary operating system. There are number of people that believe Apple is trying to damage GCC with LLVM.

Personally I think LLVM is good for GCC and vice versa and is a rotten example. They keep each other honest and progressing.

But I don't think Apple is a good steward of FOSS code and they choose BSD so that they can drop a nuke on the community if it becomes in their interest to do so, and I truly believe they would if they believed it was in their interest to do so. GPL does one thing really well, it keep all participants on equal footing. I personally don't believe the BSD license would be a viable competitor were it not for GPL and Linux opening up private enterprise to the idea BSD licensing as a lessor evil to GPL.

Without the GPL, BSD licensed code would be practically non-existent and virtually irrelevant in the marketplace. It's the very strength of Linux and the GPL ecosystem that developed that allowed BSD licensed code to become acceptable to companies that want to subvert FOSS. I personally believe that the corporate effort to replace major GPL ecosystem components with BSD licensed clones is an attempt to subvert FOSS. Ultimately an effort I think will be futile as the code can be forked GPL and the improvements can only flow one direction in such a case. The only thing that keeps these efforts going is strong corporate backing from companies like Apple that clearly want to subvert the GPL. In 20 years or so we'll know how successful they were.

Winning the copyleft fight

Posted Feb 17, 2016 4:06 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]

> Without the GPL, BSD licensed code would be practically non-existent and virtually irrelevant in the marketplace.
Really? Right now we have pretty much achieved a "market without GPL". Yet I see OpenSource development to be bigger than ever before.

That's actually easy to explain - the Open Source development model is clearly superior for "plumbing-style" projects. And this days 90% of computing is "plumbing".

Winning the copyleft fight

Posted Feb 22, 2016 0:27 UTC (Mon) by mlopezibanez (guest, #66088) [Link]

Apple's GCC developers did move to work exclusively on LLVM, but in the last 10 years, most contributions to GCC (C/C++) have come from Mentor/Codesourcery, Redhat, Suse and Google (roughly in comparable amount).

Then, a couple of years ago, Google moved all its GCC developers to work on LLVM/Clang (and probably the same is going on with Google's GDB developers now mostly working on LLDB). Mentor has also moved significant resources from GCC to LLVM, but a few of its developers keep contributing to GCC, for now. Some of the work on GCC done by researchers is now exclusively done on LLVM. I'm not saying there is a nefarious plan going on, but GCC has lost and is losing developers to LLVM.

Winning the copyleft fight

Posted Feb 13, 2016 4:23 UTC (Sat) by spaetz (guest, #32870) [Link] (8 responses)

> Case in point: https://github.com/mailpile/Mailpile/graphs/contributors?...
> Mailpipe switched to AGPL in Jan 2015. And after that new contributions not made by its principal author pretty much died away. If the author loses interest, then the project is going to be 0xDEAD.

Really? THIS is your case in point? There only ever were 3 committers with more than a handful commits, even when the thing was Apache licensed. In _May_ 2015 they started a vote on the licensing, in _July_ 2015 they announced the AGPL as the winner on their blog. In August, the founder blogged about impeding burnout and later about paternity leave and declared code feature freeze until 1.0 is released.

So, any effect from beginning of 2015 that you see, cannot be license related.

This does not exactly sound like a case where a switch to thge AGPL killed a vibrant community...

Winning the copyleft fight

Posted Feb 16, 2016 1:11 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link] (7 responses)

Yes, you're right. I somehow messed up the date.

Winning the copyleft fight

Posted Feb 18, 2016 5:52 UTC (Thu) by branden (guest, #7029) [Link] (6 responses)

I think you misspelled "confirmation bias".

You saw what you wanted to see.

Winning the copyleft fight

Posted Feb 18, 2016 21:38 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (5 responses)

Probably. But do you know of _any_ AGPL-ed successful product that doesn't have a dual license?

Winning the copyleft fight

Posted Mar 5, 2016 3:25 UTC (Sat) by jospoortvliet (guest, #33164) [Link] (4 responses)

Why the 'no dual license' part? OwnCloud very explicitly choose AGPL to keep the code free when it started and the company works well with that, deriving a significant portion of its revenue from companies that want a proprietary licensed version.

Totally freedom-compattible, promoting even, and core to our business model. (yes, it would be nicer to do pure support but many companies would simply not want the code if it was agpl only. Yeah, crazy)

And plenty successful, we trippled our user base last year :-)

Winning the copyleft fight

Posted Mar 5, 2016 4:17 UTC (Sat) by andresfreund (subscriber, #69562) [Link] (1 responses)

Because it essentially means nobody else can contribute on an equal footing. That might me the only viable approach sometimes, but it's quite different from projects without such asymmetry.

Winning the copyleft fight

Posted Mar 17, 2016 18:28 UTC (Thu) by jospoortvliet (guest, #33164) [Link]

It's true that the company can do something other contributors can't: sell this with a proprietary license.

On the other hand, that is ONLY the case for the core, almost all other apps have no copyright assignment and the core contributions come for >80% from the company. Of course, without this licensing model, there would be less contributions (as there was less opportunity to make money) and less users (because, as I said, a proportion of users, about 1/3rd of our customers, only wants the enterprise edition for the non-copyleft license, a privilege they pay for).

In other words, as I said - ownCloud would never have been as successful as it is without the company - and I'm only talking code and users here, not the effects of marketing and quality and infrastructure and everything else.

And the other way around - the company would never have worked without the community. Perfect symbiosis if you ask me.

Winning the copyleft fight

Posted Mar 9, 2016 15:07 UTC (Wed) by nye (guest, #51576) [Link] (1 responses)

>Why the 'no dual license' part? OwnCloud very explicitly choose AGPL to keep the code free when it started and the company works well with that, deriving a significant portion of its revenue from companies that want a proprietary licensed version.

>many companies would simply not want the code if it was agpl only

So what you're saying is that the proprietary version is successful and the AGPL version is not. That sounds like a strong argument in favour of Cyberax's point.

Winning the copyleft fight

Posted Mar 17, 2016 18:24 UTC (Thu) by jospoortvliet (guest, #33164) [Link]

Not sure how you read that - what I said is that ownCloud would get less users (to be precise, less enterprise users) without the option of a proprietary license.

It makes no difference for the AGPL users, of course, how could it? The AGPL version is extremely successful - the vast majority of users use the community version, not the proprietary version.

Winning the copyleft fight

Posted Feb 25, 2016 9:04 UTC (Thu) by Del- (guest, #72641) [Link] (16 responses)

> A nice way to split community.

But Apple's take on LLVM is not splitting community? What about the AOO ordeal? I could go on ... and on. Copyleft (with distributed copyright) is about building communities, permissive licenses is about splitting communities. Permissive licenses are specifically created for that single purpose. Haven't you heard the new buzz-word, "we need to differentiate"? Maybe what you really meant is that you want us all to give you our code so that you can fork it for your own profit without needing to share back? I believe it is high time we start the forking.

Winning the copyleft fight

Posted Feb 25, 2016 10:20 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (15 responses)

> But Apple's take on LLVM is not splitting community?
Nope. It's a completely _different_ project.

You're certainly free to start a new GPL-ed compiler infrastructure project. That would be the correct move.

> Haven't you heard the new buzz-word, "we need to differentiate"?
Like, you know, by taking a BSD licensed project and releasing it under GPL?

> Maybe what you really meant is that you want us all to give you our code so that you can fork it for your own profit without needing to share back?
Not exactly. I don't want to share back _all_ my code if I use just ONE GPL-ed package. And lots of which has no business being open.

I'm perfectly OK with contributing or funding work on OpenSource software. For the record, my favorite license would be LGPL 2.1 with a static link exception.

Winning the copyleft fight

Posted Feb 25, 2016 10:40 UTC (Thu) by Del- (guest, #72641) [Link] (14 responses)

> > But Apple's take on LLVM is not splitting community?

>Nope. It's a completely _different_ project.

LLVM (and Clang btw.) is trying to attract developers from gcc contributors (Google, I am looking at you), and from what I hear having some success doing it. Do you have some special definition of "splitting community" maybe? That does not involve bringing developers from one over to the other?

> Like, you know, by taking a BSD licensed project and releasing it under GPL?

Yes exactly, why do you have a problem with that all of a sudden?

> I'm perfectly OK with contributing or funding work on OpenSource software. For the record, my favorite license would be LGPL 2.1 with a static link exception.

I am very pragmatic, I welcome all open code, although I prefer contributing to copyleft projects. However, I hate to see corporations out to destroy the very same communities that gave them life.

Winning the copyleft fight

Posted Feb 25, 2016 11:30 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (13 responses)

> LLVM (and Clang btw.) is trying to attract developers from gcc contributors (Google, I am looking at you)
Have they forked GCC under another license? Nope. They developed a compelling alternative and other companies are moving to it.

> Yes exactly, why do you have a problem with that all of a sudden?
Because it can _irrevocably_ split a community. Do you remember XEmacs vs. Emacs story? The same deal here.

Code from GPL-ed branch can NEVER flow back to a non-GPL-ed branch.

> However, I hate to see corporations out to destroy the very same communities that gave them life.
It's interesting, the cognitive dissonance in your words is huge. On the very first line of your reply you mention how Apple is attracting corporate contributors to gcc, and then you rail against corporations.

Shouldn't gcc be developed by engineers working during their spare time? Perhaps, funded by a non-profit? It's supposed to be the flagship free project, after all!

Winning the copyleft fight

Posted Feb 25, 2016 11:52 UTC (Thu) by Del- (guest, #72641) [Link] (12 responses)

> It's interesting, the cognitive dissonance in your words is huge. On the very first line of your reply you mention how Apple is attracting corporate contributors to gcc, and then you rail against corporations.

Please avoid strawmen. I dislike how Apple splits the GCC community by deliberately attracting developers away from the project, when they could have continued collaborating with GCC instead. Especially considering the history, where Apple were dependent on GPL software to survive against Microsoft. Where I come from, you are not supposed to back-stab those who help you.

Moreover, I certainly do not rail against corporations, never have.

> Code from GPL-ed branch can NEVER flow back to a non-GPL-ed branch.

Yes it can, just like proprietary code can be opened up. But I see your point.

> Because it can _irrevocably_ split a community. Do you remember XEmacs vs. Emacs story? The same deal here.

Cheery-picking? Have you had a look at the BSDs lately? Yes, plural is deliberate. I am questioning why you have a problem with forking, when you argue strongly for permissive licenses. The only reason to pick a permissive license over copyleft is if you want to allow forks. I have no problem in seeing the downsides of splitting a community, that is why I want copyleft. I thought that was clear?

Winning the copyleft fight

Posted Feb 25, 2016 11:59 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (10 responses)

> Please avoid strawmen. I dislike how Apple splits the GCC community by deliberately attracting developers away from the project, when they could have continued collaborating with GCC instead.
There's a smaaaaal issue here - GCC decided to switch to GPLv3.

Yes, gcc could have collaborated with Apple by not forcing a hostile fork to a more restrictive license.

Hmm... That reminds me of something related to this thread....

> Yes it can, just like proprietary code can be opened up. But I see your point.
Proprietary code usually has one copyright owner. If you have several dozens of contributors then getting them all to agree to relicense the code might be impossible.

> Cheery-picking? Have you had a look at the BSDs lately?
Yes, I actually had a look a couple of months ago. Companies that use BSD internally are contributing changes back. And that even includes such ungentle giants as Sony.

> Yes, plural is deliberate.
Sure. Because there's just one Linux Desktop, as everybody knows.

Winning the copyleft fight

Posted Feb 25, 2016 15:32 UTC (Thu) by aggelos (subscriber, #41752) [Link] (4 responses)

Yes it can, just like proprietary code can be opened up. But I see your point.
Proprietary code usually has one copyright owner. If you have several dozens of contributors then getting them all to agree to relicense the code might be impossible.

IIRC "our code base integrates code from an unspecified third-party who is not willing to change their license" has often been brought up as a showstopper from companies who were being pressured to open up some of their stuff. So taking that "usually" with a grain of salt.

That said, I expect that copyleft code remains useful and accessible for the vast majority of people or entities who would want to use it. The two obvious exceptions being entities who want to distribute proprietary derivatives and entities who want to integrate the software in a setup that is locked down (as far as their users/customers are concerned).

With that in mind, it seems to me the split of a copyleft fork (which doesn't happen that often anyway, see below) only disenfranchises a few contributors to a project (steering clear of the word 'community' here; it's debatable whether developers of proprietary forks at least should be considered well-behaved members of an open source / free software project community).

Granted, those contributors might be economically powerful and could be employing many of the most active contributors. Which is one of the reasons not to fork lightly. AFAICT, copyleft forks of permissively-licensed projects don't happen often (evidence: all the projects out there that could have been forked but haven't). The examples that come to mind ATM were actually forked for very good reasons, e.g. LibreOffice and Kallithea. The earlier reference to Xemacs vs GNU Emacs above doesn't seem directly relevant to this discussion BTW, as IIUC, the main contention point was the copyright assignment requirement; both projects are copyleft-licensed. When such a fork does happen, it seems to be because the prior arrangement is deemed unreasonably beneficial to the parties left behind.

So, while casually forking projects is not a good idea (and AFAIU nobody claimed that it was), a copyleft fork is not necessarily "community splitting" either. In fact, the word "community" brings with it a narrative of entities happily working together, whereas it's very often the case that contributors to an open source project will have different or even opposing interests. In that case, the benefits of collaborating need to be considered in the larger context of "how can we best achieve our goals?"

Your reference to apple and gcc's switch to gplv3 seems disingeous to me. The anti-tivoization clause should be pretty much irrelevant for a compiler. And the posturing of having apple-employed clang developers presenting benchmarks against the latest gplv2 version of gcc (b/c they weren't allowed to /use/ any later gcc apparently) seems indicative of fanaticism at best. Hopefully that's changed since I last saw such a presentation...

Winning the copyleft fight

Posted Feb 25, 2016 23:11 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (2 responses)

> The examples that come to mind ATM were actually forked for very good reasons, e.g. LibreOffice and Kallithea. The earlier reference to Xemacs vs GNU Emacs above doesn't seem directly relevant to this discussion BTW, as IIUC, the main contention point was the copyright assignment requirement; both projects are copyleft-licensed. When such a fork does happen, it seems to be because the prior arrangement is deemed unreasonably beneficial to the parties left behind.
LibreOffice had a fairly good reason for forking - its upstream is pretty much dead.

And XEmacs vs. Emacs essentially replicates the effects of the BSD -> GPL fork, except that in this case it was caused by copyright assignment requirement and not by a license.

> So, while casually forking projects is not a good idea (and AFAIU nobody claimed that it was), a copyleft fork is not necessarily "community splitting" either.
People actually DO advocate exactly that.

Winning the copyleft fight

Posted Feb 26, 2016 11:48 UTC (Fri) by Wol (subscriber, #4433) [Link]

> > The examples that come to mind ATM were actually forked for very good reasons, e.g. LibreOffice and Kallithea. The earlier reference to Xemacs vs GNU Emacs above doesn't seem directly relevant to this discussion BTW, as IIUC, the main contention point was the copyright assignment requirement; both projects are copyleft-licensed. When such a fork does happen, it seems to be because the prior arrangement is deemed unreasonably beneficial to the parties left behind.

> LibreOffice had a fairly good reason for forking - its upstream is pretty much dead.

LibreOffice actually forked LONG before that ...

LO is Go-OO under a different name. And I don't know when that started but it was during the Sun days.

Okay, Go-OO was a patch set to be applied to Sun-OO, while LibreOffice is a complete project on its own, but the name-change / "fork" simply set up Go-OO as a full-blown project in its own right rather than a long standing add-on.

Cheers,
Wol

Winning the copyleft fight

Posted Feb 26, 2016 23:02 UTC (Fri) by aggelos (subscriber, #41752) [Link]

So, while casually forking projects is not a good idea (and AFAIU nobody claimed that it was), a copyleft fork is not necessarily "community splitting" either.
People actually DO advocate exactly that.

Trying to cut to the point here... (a) Where are those people? The article's summation of bkuhn's position is that

fork permissively licensed software under copyleft /when necessary/.

(emphasis added). I've skimmed the comments (ugh) and haven't found anyone advocating forking a permissively-licensed project under the GPL for kicks. Even if a commenter here did suggest that, I still fail to see the big deal. At the very least I'd expect to find /some/ examples of a fork (i) happening on general principle and (ii) actually impacting development of the permissively-licensed code base, before observing this kind of backlash.

(b) I'd like to repeat the point that your reply failed to address, for the benefit of anyone trying to follow this thread:

"Community splitting" is a term loaded with the connotation that somehow the people contributing on a project are all in this together, whereas it is common and acceptable for people to be contributing to a project with different (or, again, opposing) goals in mind.

Further, it is entirely reasonable for contributors to a project to decide that the current licensing arrangement undermines their goals to such a degree that they lose more than they gain from the increased collaboration.

Finally, it is conceivable that people contributing to a permissively licensed project, who are doing so because they want to produce useful and popular free software, might feel that their goals are undermined. E.g. by a proprietary version which integrates all the functionality of the permissively-licensed project, while adding a set of additional features, extensions (possibly some that could lock users in), what have you. One does not need to agree with this position to find it defensible, if not reasonable. Whether it is a good decision for any given project and set of people, is something that depends on the specifics, of course.

Winning the copyleft fight

Posted Feb 27, 2016 12:50 UTC (Sat) by ms_43 (subscriber, #99293) [Link]

LibreOffice is not a relevant example, since it is a fork of OpenOffice.org, which was licensed LGPLv3 at the time, not under a permissive license (until 2005 or so it was dual-licensed LGPLv2 / SISSL, and you might call SISSL permissive although it's a really odd license...).

Winning the copyleft fight

Posted Feb 25, 2016 15:38 UTC (Thu) by Del- (guest, #72641) [Link] (4 responses)

> There's a smaaaaal issue here - GCC decided to switch to GPLv3.

As if that had any practical importance for Apple. It might have accelerated the process, but Apple showed contempt for GPL before the v3 issue came up (recall the khtml fork that is now webkit). It was probably only a matter of time.

> Proprietary code usually has one copyright owner.

Really? Larger proprietary projects I have been involved in all have had multiple copyright holders. In fact, that has caused much of the problems related to opening up graphics drivers.

> Sure. Because there's just one Linux Desktop, as everybody knows.

There are three, none of them forks. Your point being?

Winning the copyleft fight

Posted Feb 25, 2016 20:12 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (3 responses)

> As if that had any practical importance for Apple.
Yes, it is important for them.

> It might have accelerated the process, but Apple showed contempt for GPL before the v3 issue came up (recall the khtml fork that is now webkit). It was probably only a matter of time.
They were strictly following the license for WebKit. It's the _way_ they did it (code drops) was not helpful.

> Really? Larger proprietary projects I have been involved in all have had multiple copyright holders. In fact, that has caused much of the problems related to opening up graphics drivers.
Really. Most proprietary projects don't involve more than a handful of copyright owners. Third-party code is usually in the form of external libraries, it's extremely rare for contractors or employees to retain the copyright to the work they're doing.

3D drivers are held mostly by DRM support - it simply can not be released without violating NDAs. Even public documentation has to be vetted for possibility of DRM workaround leaks.

> There are three, none of them forks. Your point being?
That's not any different from the BSD situation.

Winning the copyleft fight

Posted Feb 25, 2016 20:44 UTC (Thu) by Del- (guest, #72641) [Link] (2 responses)

> They were strictly following the license for WebKit.

It was khtml, and no they definitely did not follow the license. Their contributions when ultimately released after years was permissively licensed. You need to read up on this story, I believe I have even provided it to you on an earlier occasion here. That could also have been khim though.

> Really. Most proprietary projects don't involve more than a handful of copyright owners. Third-party code is usually in the form of external libraries, it's extremely rare for contractors or employees to retain the copyright to the work they're doing.

Seems you are moving the goal post again, but still my experience differs from yours. It need not be libraries, and it is actually quite common to license code to eachother without signing over copyright.

> That's not any different from the BSD situation.

Of course it is, BSD has no desktops at all. Btw. I saw you mentioned Sony contributing code to BSD, have you actually seen how they do it? They toss some code over the fence, like to some hungry dogs in the yard. A world apart from the effort developers put in to get code into linux. Just go have a look on their mailinglists.

Winning the copyleft fight

Posted Feb 25, 2016 22:36 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (1 responses)

> It was khtml, and no they definitely did not follow the license.
Yes, they did follow the license by providing source code.

> Their contributions when ultimately released after years was permissively licensed.
Incorrect. The main issue with Apple was the "throw stuff over the wall" development approach, and they fixed it by providing a public repository with their ongoing work.

> Of course it is, BSD has no desktops at all.
It has: Dragonfly BSD, for example.

> have you actually seen how they do it? They toss some code over the fence, like to some hungry dogs in the yard.
And you're proposing spreading the GPL cancer by allowing it to metastasize inside healthy Apache/BSD code tissue. Your desired end result is a dead uniform stinking mass of GPL-ed cancerous tumors. See? I also can invent metaphors to suit my arguments.

And no, Sony did a good job by submitting patches to LLVM and FreeBSD.

Winning the copyleft fight

Posted Feb 25, 2016 23:32 UTC (Thu) by Del- (guest, #72641) [Link]

> It has: Dragonfly BSD, for example.

Thanks for giving me a laugh.

> And no, Sony did a good job by submitting patches to LLVM and FreeBSD.

You know, I spent more time finding a single patch from Sony to FreeBSD than it took me to dig out twenty one healthy GPLv3+ projects. This is what I found:
http://lists.freebsd.org/pipermail/freebsd-amd64/2011-Mar...
In your parallel universe I am sure this is all fine and dandy. Have you even checked out this stuff yourself, or are you making it up as you go along?

and with this you wore me out, I need a break..

Winning the copyleft fight

Posted Mar 9, 2016 15:41 UTC (Wed) by nye (guest, #51576) [Link]

>Especially considering the history, where Apple were dependent on GPL software to survive against Microsoft.

Apple has never been dependent on GPL software to survive against Microsoft, though I think a case could be made for the *exact opposite*. From at least the mid 90s, up until the last few years, Apple was heavily dependent upon Microsoft to survive, on the one hand in the form of making their software available for Mac OS, and on the other hand in the form of financial investment. Microsoft gained from this by having a straw man 'competitor' in the market, that was never a threat to anything but they could point to to claim they didn't hold a monopoly.

>The only reason to pick a permissive license over copyleft is if you want to allow forks

Copyleft licenses do not make forking any harder (how could they? that claim doesn't even make sense - the freedom to fork is considered fundamental), but they do make *merging* a great deal harder, because it's not usually possible to combine two pieces of copyleft code unless they use they exact same version of the exact same license[0]. Case in point: the recent discussions over ZFS and Linux, two great projects with apparently irreconcilable copyleft licenses.

>Have you had a look at the BSDs lately? Yes, plural is deliberate

There are probably hundreds of Linux distributions, mostly vanity projects, almost all terrible, with communities split in some cases pretty vehemently. How well is copyleft doing at preventing all this needless forking? The answer is that it's a silly question; you might as well ask how well bread is doing, because copyleft vs permissive is completely orthogonal to this point.

[0] With a few exceptions such as LGPLvX specifically permitting relicensing to GPLvX

Winning the copyleft fight

Posted Feb 26, 2016 11:41 UTC (Fri) by Wol (subscriber, #4433) [Link]

> Mailpipe switched to AGPL in Jan 2015. And after that new contributions not made by its principal author pretty much died away. If the author loses interest, then the project is going to be 0xDEAD.

Not making any comment on the MERIT of the argument, but the AGPL is very contentious. I've come across various reports of GPL-friendly lawyers who have concluded that, if part of your stack is AGPL, you pretty much have to AGPL *EVERYTHING* that that web-server touches ...

Simply put - there are a lot of GPL-friendly people out there who won't touch AGPL with a barge-pole.

Cheers,
Wol

Winning the copyleft fight

Posted Feb 12, 2016 23:16 UTC (Fri) by viro (subscriber, #7872) [Link] (6 responses)

You might be, but do you have anything valuable enough to contribute? I would expect at least some contributions to some compiler work first and I don't see anything from you in gcc tree... Do you believe that clang is that much better than gcc? If not, and if you are capable of contributing something worthwhile *and* prefer GPL so much, where's your gcc work?

Winning the copyleft fight

Posted Feb 12, 2016 23:41 UTC (Fri) by nix (subscriber, #2304) [Link]

Oh come on, this is the Internet. On the Internet, 'I would be proud to X' *always* means 'I will talk about X'. It takes ever so much less time, after all.

Winning the copyleft fight

Posted Feb 13, 2016 1:43 UTC (Sat) by xtifr (guest, #143) [Link] (4 responses)

You might be, but do you have anything valuable enough to contribute? I would expect at least some contributions to some compiler work first

Not recently, so I might be a little rusty, but I was on the team that developed what was almost certainly the first compiler for the IBM PC—while the PC itself was still a top secret project. Whether I'd still have good enough chops...I dunno. Be cool to find out, though.

Winning the copyleft fight

Posted Feb 13, 2016 2:25 UTC (Sat) by viro (subscriber, #7872) [Link] (3 responses)

Now you've got me really curious - 8086-targeted compilers existed _before_ IBM PC got even started (PL/M, for example), so what exactly are you talking about? PC-hosted one? Which language, which hosting system and where had that been happening?

All that aside, you are aware of that large GPLed compiler suite, right? Called GNU Compiler Collection or somesuch... IDGI - if you think that gcc is being harmed by clang, starting a fork of the latter is exactly the wrong thing to do; codebase differences alone will seriously complicate (to put it _very_ mildly) a migration of improvements from that fork to gcc. If clang sucks developers away from gcc, an extra fork will only exacerbate the problem...

The only way your idea might make sense is if you've concluded that gcc is hopelessly bad, clang/llvm is going to be a clear winner and it's better to accelerate the process, creating a fork with acceptable license, driving it past any remaining issues where gcc is still better and hoping that gcc developers jump ship, killing gcc off and making your fork outpace the clang mainline. I *really* doubt that they'll agree with the assessment above and desirability of such plan in general.

Another interpretation is "let's start a fork, no matter how moribund will it be, and try to split llvm developers' community; viability of fork be damned, it'll slow the mainline down", but I very much doubt that such political games would succeed in anything other than making participants (deservedly) shunned by everyone.

Winning the copyleft fight

Posted Feb 13, 2016 3:47 UTC (Sat) by xtifr (guest, #143) [Link] (1 responses)

PC Hosted. Although we started with a cross-compiler to get us started, since we were coming from the 6502! Once we got the compiler self-hosted on the PC, it was pretty straight-forward to finish porting EasyWriter. I mostly worked on the cross-compiler, since I knew the guts of the one we used on the 6502.

As for the rest of it: I have no strong opinions on whether gcc or clang is better, but I have certainly developed the impression that clang is intended (by Apple at least) as a direct attack on the GPL and copyleft. But you're right; my idea (which was really nothing more than a throwaway that popped into my mind) is probably quite premature at present. To put it mildly. So, file it away as something to think about if clang ever does start to threaten gcc, and otherwise, forget it for now.

The main point of my post (and the part I hoped people would respond to) was that the benefits of copyleft to companies is something we should be promoting and endorsing, rather than dismissing the way Kuhn seemed to be doing.

Winning the copyleft fight

Posted Feb 15, 2016 16:58 UTC (Mon) by JanC_ (subscriber, #34940) [Link]

So that was a Forth compiler then?

Winning the copyleft fight

Posted Feb 13, 2016 19:26 UTC (Sat) by vonbrand (guest, #4458) [Link]

This is exactly what happened to gcc, with the egcs fork (which in the end took over).

Winning the copyleft fight

Posted Feb 15, 2016 9:43 UTC (Mon) by mjthayer (guest, #39183) [Link] (3 responses)

> I am somewhat surprised that more companies don't see the benefits of the GPL, when applied to their own code! It may not be quite the same set of benefits that I see as an individual developer, but the ability to keep your competitors from leveraging your own code and competing against you without sharing their improvements to (again) your own code has to be something worth considering.

I have already heard that argument from people commercially writing GPL software, who put what they considered as commodity code under the GPL and licenced the rest commercially.

That said, I think the talk was not about GPL but about copyleft, and specifically about GPLed code which is not entirely owned by a single company. I still think it would be very sensible to think through the possible advantages that copyleft might have for companies. The first which comes to my mind is that if a competitor does try to misuse your GPLed code then it is more likely to be spotted and stopped, hopefully without your having to pay much for the privilege.

And if Bradley Kuhn is right about the conspiracy of corporate lawyers, then I presume that at least one of them is following this thread. In that case I will put the question directly to them - we are told that you see disadvantages in copyleft, but can you also see any potential advantages to you and your customers?

Winning the copyleft fight

Posted Feb 15, 2016 23:26 UTC (Mon) by rfontana (subscriber, #52677) [Link] (1 responses)

It's a feature of the conspiracy that none of the corporate lawyers read LWN.

Winning the copyleft fight

Posted Feb 17, 2016 1:19 UTC (Wed) by rahvin (guest, #16953) [Link]

Huh? Any good conspiracy theory includes an omnipresent, omnipotent opponent. They have to be monitoring all community discussion on the topic so as to feed contrary evidence and wage disinformation campaigns. Without the sinister ever present villain of the conspiracy theory the story just isn't as compelling.

You need to brush up on how conspiracy theories work. The formula is pretty well established.

LWN, corporate lawyers, and (non)-conspiracy

Posted Feb 23, 2016 21:09 UTC (Tue) by bkuhn (subscriber, #58642) [Link]

FWIW, I've had what I've said on LWN quoted back to me during conference calls with GPL violators lawyers. Some of them are reading this. Also, in the talk, I said it's not a conspiracy, they're just working together on a common goal: to limit the scope of strong copyleft.

The mistake our community makes is assuming that those who want to see copyleft's scope reduced are not working together to figure out how.

Winning the copyleft fight

Posted Feb 12, 2016 20:58 UTC (Fri) by andresfreund (subscriber, #69562) [Link] (1 responses)

I don't think the existence of closed forks of open project necessarily means that free software is failing (nor the contrary). In some cases, e.g. postgres, these forks are essentially what finances a lot of the developers on the open project. And a good portion of the features in the closed forks are things that realistically wouldn't be integrated in the open source version anyway (say Oracle compatibility). Don't get me wrong, it sometimes grates.

I do however strongly think that copyleft + assignment are nearly always worse than a non copyleft license would have been in the first place. It can kinda work if the collective owner has history and is deemed trustworthy. But most often it's a company that sells that software under s different license, with some features added. And that's IMO way worse than a non copyleft license, because only that one entity has the ability to fork.

Winning the copyleft fight

Posted Feb 13, 2016 12:24 UTC (Sat) by aggelos (subscriber, #41752) [Link]

FWIW, I thought it was an uncontested point by now that most people on the "external" side of a CLA consider it unfair; surely the only point of contention there would be on whether permissive is preferable to copyleft held by different entities?

Winning the copyleft fight

Posted Feb 12, 2016 22:32 UTC (Fri) by mattdm (subscriber, #18) [Link] (1 responses)

> As an example, he noted that the version of OpenStack running behind his Rackspace account clearly has features that are not in the free version. He does not know of a single OpenStack product that does not contain at least a few features that have not been pushed back upstream.

Red Hat's Open Stack implementation is entirely non-proprietary. While "upstream, upstream, upstream!" is certainly the mantra, there are probably some things which haven't made the mainline upstream — but any of those would be available as part of RDO, available from https://www.rdoproject.org/

(Disclaimer: I work for Red Hat, but not on OpenStack or RDO.)

Red Hat and OpenStack

Posted Feb 23, 2016 21:11 UTC (Tue) by bkuhn (subscriber, #58642) [Link]

mattdm, thanks for pointing that out. When I give versions of this talk in the future, I'll note that Red Hat is an exception.

Winning the copyleft fight

Posted Feb 12, 2016 22:56 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link] (1 responses)

> But copyleft won't work if we don't enforce the terms of the license, he said. Enforcement is getting harder: companies are actively working to replace copyleft-licensed software. Toybox and LLVM are two examples there. Given that, which essential copyleft-licensed software remains?
Nothing of interest, really. And it's entirely FSF's fault.

Winning the copyleft fight

Posted Feb 13, 2016 11:37 UTC (Sat) by aggelos (subscriber, #41752) [Link]

The Westboro Baptist Church is taking on digital society issues now?

Embedded enforcement?

Posted Feb 13, 2016 5:00 UTC (Sat) by Tester (guest, #40675) [Link] (19 responses)

When walking through the electronic gadgets alley at BestBuy, most of the products there are based on embedded Linux, but a large number of them don't even acknowledge that in their documentation, much less include the mandatory offer for the source code. I'd love for the conservancy/FSF to attack that problem up front, I wonder how retailers would react to C&D letters, wouldn't they be likely to just pull those products from the shelves? That would probably give the enforcer a pretty strong negotiating hand.

And it's nice to be pure and honest and everything, but GPL enforcement needs to happen on a much larger scale and the current funding model is just not scaling. Getting money as part of the settlements extracted from the most recalcitrant violators seems like a perfectly fair way to do it.

Embedded enforcement?

Posted Feb 13, 2016 19:29 UTC (Sat) by vonbrand (guest, #4458) [Link]

The FSF has no standing with regards to the Linux kernel, it doesn't hold any (significant) copyright on it.

Embedded enforcement?

Posted Feb 13, 2016 21:53 UTC (Sat) by pabs (subscriber, #43278) [Link]

SFC already asks for legal expenses from violators.

https://sfconservancy.org/copyleft-compliance/principles....

Embedded enforcement?

Posted Feb 15, 2016 10:30 UTC (Mon) by epa (subscriber, #39769) [Link] (16 responses)

Even if they did comply and included a CD with the source code, what would that achieve? You usually have no way to load the modified code onto the device, so it doesn't help you make better use of the device you bought. The vanilla Linux kernel sources are widely available anyway. And who wants to reuse or even read the ugly write-and-throw-away driver code and kludges specific to one particular device from one manufacturer?

Embedded enforcement?

Posted Feb 15, 2016 13:30 UTC (Mon) by pizza (subscriber, #46) [Link] (15 responses)

> And who wants to reuse or even read the ugly write-and-throw-away driver code and kludges specific to one particular device from one manufacturer?

*raises hand*

Because the source code with those ugly kludges at least imparts the information necessary to enable *someone* to support the device later.

Without that source code, the likes of DD-WRT and cyanogen would not exist; both produce a far superior "product" than what the manufacturer originally released.

Embedded enforcement?

Posted Feb 15, 2016 21:04 UTC (Mon) by Tester (guest, #40675) [Link]

Also, if we don't punish the cheaters, then no one will comply and the GPL will not be worth the bits it's written on and we may as well give up on copyleft.

Embedded enforcement?

Posted Feb 16, 2016 9:38 UTC (Tue) by epa (subscriber, #39769) [Link] (13 responses)

Because the source code with those ugly kludges at least imparts the information necessary to enable *someone* to support the device later.

That's not the case for some random electronic gizmo which lacks any way for you to replace the software - either because it has nonupgradeable firmware, or because it requires signed firmware images. That's really the case I am thinking of. I agree, for a computing device where you can install your own kernel it is necessary to have the source code. My point is that the source code by itself, without any means to run it on the device you bought, doesn't give you much.

Embedded enforcement?

Posted Feb 16, 2016 13:25 UTC (Tue) by pizza (subscriber, #46) [Link] (11 responses)

> My point is that the source code by itself, without any means to run it on the device you bought, doesn't give you much.

I disagree, quite vehemently.

Regardless of the practicalities of running modified versions, under the GPLv2 they are *legally obligated* to provide the complete corresponding source code. Even if the device puts everything in mask ROM.

Embedded enforcement?

Posted Feb 20, 2016 4:21 UTC (Sat) by giraffedata (guest, #1954) [Link] (10 responses)

under the GPLv2 they are *legally obligated* to provide the complete corresponding source code

No, they aren't. A copyright license cannot require anyone to do anything. A license is permission. It lets a licensee do things; it doesn't prevent the licensee from doing things.

When people use phrases like "comply with the license" they confuse everyone about this. A license isn't something you comply with. Copyright law is.

In this case, where code is not licensed by its copyright holder except under GPLv2, the manufacturer who doesn't offer source code is legally obligated, by copyright, not to use the code in its product. And if the manufacturer does use the code in its product, it's then legally obligated to pay royalties, or sometimes other money, to the copyright holder. Most copyright holders offer to waive those payments in exchange for the manufacturer licensing its additions to the code to the public, and the manufacturer usually takes the offer, so it may make it look like the manufacturer was forced to license its code.

Embedded enforcement?

Posted Feb 20, 2016 14:43 UTC (Sat) by pizza (subscriber, #46) [Link] (7 responses)

> No, they aren't. A copyright license cannot require anyone to do anything. A license is permission. It lets a licensee do things; it doesn't prevent the licensee from doing things.

Reading your response, a certain quote comes to mind: "You are technically correct, and that is the best kind of correct."

So while perhaps I should have said "complying with the law to avoid massive statutory penalties that would arise if I chose to perform actions that are the exclusive right of the copyright holder without first obtaining permission from said copyright holder in the form of a license and supplying whatever considerations and adhering to whatever conditions the license stipulated," Instead I'll take a page from my employer's in-house IP counsel and use the same shorthand "comply with the license" language they used not two days ago.

Embedded enforcement?

Posted Feb 21, 2016 17:41 UTC (Sun) by giraffedata (guest, #1954) [Link] (6 responses)

If you're trying to rationalize using the shorthand "comply with the license," you don't have to look any further than copyleft expert Bradley Kuhn, who uses the phrase a lot, including in this very article. But if you want to be understood and not spread misconceptions about how copyright works, you'd be better off using the slightly longer phrase, "comply with the license conditions."

Likewise, another oxymoron Kuhn and others like to use is "violate the license," which can be replaced with the more accurate and equally short, "violate the copyright."

Embedded enforcement?

Posted Feb 21, 2016 17:59 UTC (Sun) by pizza (subscriber, #46) [Link]

> But if you want to be understood and not spread misconceptions about how copyright works, you'd be better off using the slightly longer phrase, "comply with the license conditions."

Good point, I will try to do that going forward.

Comply with the license is reasonable terminology

Posted Feb 23, 2016 21:15 UTC (Tue) by bkuhn (subscriber, #58642) [Link] (4 responses)

I stand by the term "comply with the license". Copyright law is of course the governing mechanism. To cease infringing copyright, one must comply with the copyright license they were given. Similarly true with violating the license. If you fail to comply with the license, you violate the license. This does not contradict the fact that copyright law is legal framework within which that license operates.

Comply with the license is reasonable terminology

Posted Feb 23, 2016 22:53 UTC (Tue) by giraffedata (guest, #1954) [Link] (3 responses)

I'm making a grammatical point. You cannot comply with a license. Those words do not go together. You comply with something that demands something of you, like a law or an order or a request. A license does not do that; it gives permission. Complying with a license makes as much sense as complying with a gift.

This misnomer does cause people to mistake the mechanism of copyleft, and conditional copyright licenses in general. People especially draw the wrong conclusions about the consequences of having copied something without having met the conditions of some conditional copyright license.

It is not hard to use plainer English that makes this complex law more clear.

Comply with the license is reasonable terminology

Posted Feb 24, 2016 10:16 UTC (Wed) by paulj (subscriber, #341) [Link]

Would "honour the licence" be an acceptable substitute, to "honour" the licence is more concordant with a copyleft licence as a gift - while being shorter than "comply with the licence conditions"?

Comply with the license is reasonable terminology

Posted Feb 24, 2016 14:39 UTC (Wed) by nybble41 (subscriber, #55106) [Link] (1 responses)

> I'm making a grammatical point. You cannot comply with a license. Those words do not go together. You comply with something that demands something of you, like a law or an order or a request. A license does not do that; it gives permission.

The license gives *conditional* permission. Complying with the license means complying with those conditions, in order to receive permission. It's a perfectly clear and reasonable use of the words. One can comply with a license but not copyright law (for example, by meeting the terms of the license but using the work in ways copyright law restricts and the license does not permit), or comply with copyright law but not a particular license (under the terms of a different license, or fair use, or in ways copyright law does not restrict); or one can comply with both, or neither.

Comply with the license is reasonable terminology

Posted Feb 25, 2016 3:08 UTC (Thu) by giraffedata (guest, #1954) [Link]

The license gives *conditional* permission.

The license is the actual permission. They're legal synonyms. Conditional permission is a conditional license. (And by the way, don't make the mistake many do of thinking that the document is the license; that's just a memorandum of the license).

It is clear to you and me that "comply with the license" means meet the conditions of the license because we know how copyright works and literally complying with the license per se wouldn't make any sense. But I'm concerned about people who don't know how something like copyleft works, and are trying to infer it based on the way "comply" is used in the rest of English. Think of all the ways outside of the open source world you hear "comply." It always implies something someone's trying to make you do, not something that's merely made available for you to exploit.

For the first year or two that I was trying to follow conversations about GPL, I was convinced there was some kind of contract involved, because people seemed to be talking about publishers forcing other people to release their code. Someone finally explained to me that it's just about a conditional copyright license. Terms such as "comply," "violate," and "enforce" surely contributed to my confusion.

Embedded enforcement?

Posted Feb 26, 2016 13:41 UTC (Fri) by Wol (subscriber, #4433) [Link] (1 responses)

>> under the GPLv2 they are *legally obligated* to provide the complete corresponding source code

> No, they aren't. A copyright license cannot require anyone to do anything. A license is permission. It lets a licensee do things; it doesn't prevent the licensee from doing things.

Actually, they are.

EITHER they are distributing the code under the GPL (which is what the GP said) in which case they ARE legally obligated

OR they are distributing without a licence.

Cheers,
Wol

Embedded enforcement?

Posted Feb 26, 2016 16:38 UTC (Fri) by giraffedata (guest, #1954) [Link]

It sounds like you might have the very cause and effect confusion that I worry about spreading with the imprecise use of words like "comply."

EITHER they are distributing the code under the GPL (which is what the GP said) in which case they ARE legally obligated

The GP didn't actually say they are distributing under the GPL (I took the reference to GPL to be that it is "GPL code"), but if they are, then they have already provided (or at least offered) source code. It does not make sense to say someone is legally obligated to do something he has already done. If they have not provided source code, the GPL does not apply, so they have not distributed under that.

For anyone who is confused by this, consider this: one does not have to accept GPL or make any declaration in advance that one is relying on it. One simply copies (distributes). If the author has a problem with that, he accuses the copier of copyright infringement. The copier offers the defense of, "you gave me permission to copy: GPL." The author may then say, "no I didn't, because you didn't meet my conditions." If the copier never brings up GPL, then the author cannot either.

I worry that people think when you throw a copy of GPL in your tarball, you create the situation, "If you distribute the object code, you MUST distribute the source code," when in fact, it is the other way around: "If you distribute the source code, you MAY distribute the object code."

Embedded enforcement?

Posted Feb 20, 2016 4:25 UTC (Sat) by pabs (subscriber, #43278) [Link]

You can also audit the source code for jailbreak^Wsecurity vulnerabilities more easily, which can lead you to buy a more secure device (or install different firmware).

Winning the copyleft fight

Posted Feb 13, 2016 14:54 UTC (Sat) by alankila (guest, #47141) [Link] (27 responses)

> The one thing that companies haven't tried to rewrite, he said, is the Linux kernel. Thus Linux is the battleground over which the future of copyleft licensing will be determined.

This sounds counterproductive. BSD kernels exist, for instance. I bet a few of them are perfectly serviceable replacements for Linux.

BSD

Posted Feb 13, 2016 17:31 UTC (Sat) by tialaramex (subscriber, #21167) [Link] (21 responses)

Well, it depends what your problem is

For example if your problem involves hardware that BSD doesn't offer a driver for, you're going to need to develop the BSD driver to make BSD a replacement for Linux. The situation today is that the hardware vendor offers a Linux driver because their customers all want Linux and they need to prove it works with _something_ or nobody will buy. Possibly the Linux driver is shit. But moan as we might, people prefer "shit" to "non-existent".

Of course if everybody selling to consumers systematically jumps ship to, say, FreeBSD that would offer an impetus to component vendors and maybe five years from now there would be BSD drivers for everything new on the market. But for those five years what's the rationale? "If we keep this up, assuming we don't go out of business, we get to thumb our noses at the GPL" ? Any company that decides not to get on board gets a huge competitive advantage now, yet they can still choose to switch later if the plan comes off, so there's no incentive to make the big sacrifice.

I think if you look around at new products coming onto the market with Linux inside them, it's often possible to imagine how a BSD hobbyist could build something similar, but for perhaps a bit more money and spending some period of time writing their own software to make it work. But it's rarely obvious how to build the _same_ product, with the _same_ materials budget and _same_ time to market, except by using Linux. And that's the difference between a successful, profitable ground-breaking product and arriving last to market with an also-ran product which costs more but does less.

BSD

Posted Feb 17, 2016 22:04 UTC (Wed) by alankila (guest, #47141) [Link] (20 responses)

You are right in that a switch away from Linux would be a serious engineering problem of nontrivial cost in terms of time and manpower. This cost is also not infinite. If the commercial world starts to get the impression that use of Linux tends to mean problems down the road, then they will look harder into alternatives, and may latch on to another choice if it suits their needs.

Going back into history, it seems to me that FSF overestimated the hold they had of the world with GPLv2, the thought was that they could force a specific behavior through change in the licensing terms, and would outright remove certain business models that they don't like such as those that involve locked-down devices. It seems that the commercial world has consequently reduced its support for GPL software. It is no longer their first choice (except maybe for the cases when it is some copyright holder that sells something like AGPL3 software also under some other commercial license. But this is an exception that proves the rule: they know full well that this license is unpalatable to most companies).

BSD

Posted Feb 17, 2016 23:57 UTC (Wed) by pizza (subscriber, #46) [Link] (19 responses)

> ...and would outright remove certain business models that they don't like such as those that involve locked-down devices.

As has been said here many times, the FSF didn't fire the first shots in that battle.

> It seems that the commercial world has consequently reduced its support for GPL software.

They never liked it to begin with; but put up with it as it cost them less than commercially licensing something else. Once there was a viable alternative that cost them even less, they jumped the first chance they could. But that sort of thing is obvious to anyone with three brain cells; businesses will always seek to get more for less.

What the FSF didn't anticipate is that they'd also see so many *developers* actively undermining their own interests by gleefully jumping over to more permissively-licensed projects.

BSD

Posted Feb 18, 2016 0:05 UTC (Thu) by bronson (subscriber, #4806) [Link] (18 responses)

> As has been said here many times, the FSF didn't fire the first shots in that battle.

Well, it wasn't much of a battle until the FSF started firing shots.

(heck, even now it isn't much of a battle)

BSD

Posted Feb 18, 2016 0:34 UTC (Thu) by pizza (subscriber, #46) [Link] (17 responses)

> Well, it wasn't much of a battle until the FSF started firing shots.

Yeah, how dare they fight back when attacked!

I suggest you look up the events that led to the term "tivoization" being coined.

While you're at it, look up the Java Railroad Model Interface (JMRI) folks went through.

BSD

Posted Feb 18, 2016 2:07 UTC (Thu) by bronson (subscriber, #4806) [Link] (16 responses)

Meh, nobody attacked the FSF. And, despite their divisive effort, they haven't put a dent in tivoization. Turns out it's no big deal.

JMRI was the artistic license, wasn't it? Not sure how it applies here.

BSD

Posted Feb 18, 2016 2:59 UTC (Thu) by pizza (subscriber, #46) [Link] (5 responses)

Au contraire -- Tivo used glibc, a core FSF project. While what Tivo did wasn't a direct attack on the FSF, it was a slap in the face and exposed a major vulnerability.

And as for JMRI -- The GPLv2's lack of an explicit patent grant meant that it is vulnerable to the same sort of attacks that JMRI faced.

In both cases, there was a demonstrated threat to the FSF's core mission (ie Software Freedom) and the licenses that enabled that mission. The (very) public consultation and drafting process that followed culminated with the GPLv3 license family.

BSD

Posted Feb 18, 2016 4:35 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (3 responses)

> Au contraire -- Tivo used glibc, a core FSF project. While what Tivo did wasn't a direct attack on the FSF, it was a slap in the face and exposed a major vulnerability.
They can switch to musl these days. It's not a big deal anymore.

> The (very) public consultation and drafting process that followed culminated with the GPLv3 license family.
Yes, and it very publicly forced companies like Apple to invest billions in replacing GPL software.

BSD

Posted Feb 18, 2016 13:36 UTC (Thu) by pizza (subscriber, #46) [Link] (2 responses)

> Yes, and it very publicly forced companies like Apple to invest billions in replacing GPL software.

Yes, poor Apple, forced to actually spend money in order to have software it can use on its own terms instead of freeloading on someone else's work.

BSD

Posted Feb 18, 2016 16:41 UTC (Thu) by raven667 (subscriber, #5198) [Link] (1 responses)

They weren't freeloading, they were contributing as well, including the whole ObjC language, if under a bit if protest. If you are trying to create a worldwide software commons where copyleft is the new default and proprietary is the exception you want to attract the worlds largest software developers and their billions of dollars worth of time, not push them away. We can build a self-sustaining copyleft community without them, but if you want more than basic survival you need to grow the development community.

BSD

Posted Feb 18, 2016 20:28 UTC (Thu) by pizza (subscriber, #46) [Link]

> They weren't freeloading, they were contributing as well, including the whole ObjC language, if under a bit if protest.

"A bit under protest" is a bit of an understatement -- NeXT only contributed their Objective-C frontend back to GCC after the lawyers got involved.

In a bit of irony, the fallout from that led to GCC deliberately making it harder to add proprietary extensions, which also made it harder to extend in any way, which led to the creation of LLVM.

BSD

Posted Feb 26, 2016 19:39 UTC (Fri) by Wol (subscriber, #4433) [Link]

> And as for JMRI -- The GPLv2's lack of an explicit patent grant meant that it is vulnerable to the same sort of attacks that JMRI faced.

Unfortunately, especially in America, ANY project is vulnerable to attack. JMRI was especially egregious, though.

For those who don't know, JMRI was written by a hobbyist. A company started distributing the code, patented it, then sued the hobbyist for breach of patent. A patent on the hobbyist's own code!!!

A typical example of American "abuse of process", imho.

It all turned out "well" in the end, but at considerable cost ...

Cheers,
Wol

BSD

Posted Feb 25, 2016 9:20 UTC (Thu) by Del- (guest, #72641) [Link] (9 responses)

> Meh, nobody attacked the FSF. And, despite their divisive effort, they haven't put a dent in tivoization.

All my devices are open (unfortunately still some closed linux drivers on the phone though). It would never have happened without FSF and their contribution of GPL combined with the law cases years back forcing vendors to open up devices. I cannot fathom how you can argue against that. Maybe you meant the GPLv3 hasn't made a dent? I beg to differ here to, haven't you noticed the latest attempts at locking down the PC?

BSD

Posted Feb 29, 2016 19:15 UTC (Mon) by bronson (subscriber, #4806) [Link] (8 responses)

> haven't you noticed the latest attempts at locking down the PC?

Of course. And none of those articles have mentioned GPLv3. Care to share a bit more?

BSD

Posted Feb 29, 2016 22:06 UTC (Mon) by Del- (guest, #72641) [Link] (7 responses)

> Of course. And none of those articles have mentioned GPLv3. Care to share a bit more?

Grub is GPLv3+ and is in use for all GNU/Linux distributions (worth mentioning), and linux has a non-trivial market share (yes, even on the desktop, but workstations and servers are also important markets in this respect). Lilo lost. The anti-Tivo clause is in effect for boot-loaders. You may of course argue that PCs would not be tivotized regardless, but then you will be neglecting all statistical evidence. Which would be a bit tedious to list all here. However, on important example is the U-boot bootloader, it is available under GPLv2, I cannot think of any platform using uboot that has not been tivotized. Feel free to make your own conclusions though. And yes, opening up locked down routers and phones is a very tedious exercise, those who do it never stops to amaze me.

BSD

Posted Feb 29, 2016 22:41 UTC (Mon) by viro (subscriber, #7872) [Link] (6 responses)

Yeah, I just can see it - evviiil corporation hell-bent on locking everything down, planning to use something Linux-based on their hardware and... damn, foiled by that pesky grub and its license. Use lilo or elilo? Nah, can't do - it has "lost", whatever that means. And users of that thing (willing to live with locked-down system, etc.) won't tolerate the lack of grub features; yes, sir - that'd be the last straw. Why, just to think - need to rerun that thing after installing the self-built kernel! In-fucking-tolerable. Mind you, being unable to install said self-built kernel (and defeat whatever lockdown they had forced on them) wouldn't be a problem, but the knowledge that having accomplished that feat they would have to run that... that thing - that would've scared them off. And staying quiet while booting the only kernel provided (no menus, etc.) would also be a deal-breaker. Or being unable to switch the root filesystem type from what's provided by the install given by Duh Evvviiill Corp, for that matter - nevermind that being able to do that would mean having defeated the lockdown anyway...

</sarcasm> Seriously, what the hell are you smoking? Scratch that, what the hell do you expect your readers to be smoking in order to buy that load of horse manure of yours?
<sarcasm>

BSD

Posted Mar 1, 2016 7:19 UTC (Tue) by Del- (guest, #72641) [Link] (5 responses)

> Yeah, I just can see it - evviiil corporation hell-bent on locking everything down,

Corporations typically try to maximize profits. Practically all device manufacturers lock down devices as a means to control their product and earn more money. Evil has nothing to do with it. You cannot sell a locked down device with Red Hat or Suse, I don't even believe any of them support alternative boot-loaders anymore. On a related note, Canonical was specifically asked to use GPLv2 on upstart to allow lock down, and they complied. Sony's actions on PS3 is a good study on open/locked down computers, ever thought about why they ended up locking it down completely? HTC for a period supplied tools to unlock their phones after popular demand. Guess what? When they discovered how little the tinkering community was, they removed the tool. Why shouldn't they, they lost practically no business, and ensured that all HTC phones continue to run their version of Android in the process. Ever heard of vertical integration? Short story, we need something more than just the goodwill of manufacturers to keep computers open, and GPLv3 is today our only means to do it. You are of course free to reach another conclusion, but please try to behave civilized.

BSD

Posted Mar 1, 2016 9:00 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link] (4 responses)

Unlocked (i.e. not sold through carriers) HTC phones can be rooted with fastboot.

But don't let that dissuade you from ranting...

BSD

Posted Mar 1, 2016 12:45 UTC (Tue) by Del- (guest, #72641) [Link] (3 responses)

> Unlocked (i.e. not sold through carriers) HTC phones can be rooted with fastboot.

No it can not. You probably mean to say that HTC has re-introduced their unlock tool, but some carriers still lock them down. Indeed, so it seems (I was unaware):
http://www.htcdev.com/bootloader
I am very happy to see that, and I hope more manufacturers will follow. Not holding my breath though.

> But don't let that dissuade you from ranting...

Now that is beneath you. If you have a case, just state it. I know you enjoy a good discussion just as much as I do.

BSD

Posted Mar 1, 2016 18:21 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link] (2 responses)

> No it can not.
Yes, you can. I have an HTC phone bought at a retail store, it's lying on my desktop right now. I simply unlocked it with fastboot.

> You probably mean to say that HTC has re-introduced their unlock tool, but some carriers still lock them down.
Nope. Fastboot is the standard tool in Android-world for bootloader manipulation.

Carriers definitely can lock it down if you lease your phone, not buy it.

> Now that is beneath you. If you have a case, just state it. I know you enjoy a good discussion just as much as I do.
Sure, a little bit of trolling is good for health.

BSD

Posted Mar 1, 2016 23:11 UTC (Tue) by mathstuf (subscriber, #69389) [Link] (1 responses)

As little(?) as 14 months ago, a Moto X from Verizon was not unlockable. Motorola would not gives the codes for Verizon phones. Now, they might have changed thos since, but last I heard the line was being neglected for updates. Guess I made the right choice in waiting for the new Nexus line to come out (stuck on Verizon; every other carrier has shit coverage where I need it).

BSD

Posted Mar 1, 2016 23:29 UTC (Tue) by dlang (guest, #313) [Link]

14 months can be a long time in this business.

Winning the copyleft fight

Posted Feb 13, 2016 21:54 UTC (Sat) by pabs (subscriber, #43278) [Link]

*cough*MacOS*cough*

Winning the copyleft fight

Posted Feb 17, 2016 1:28 UTC (Wed) by rahvin (guest, #16953) [Link] (3 responses)

>BSD kernels exist, for instance. I bet a few of them are perfectly serviceable replacements for Linux.

Were that actually true you would see BSD kernels in these electronics. You don't. The market has clearly determined that BSD Kernels are not a serviceable replacement for Linux, even if you don't agree.

Winning the copyleft fight

Posted Feb 17, 2016 10:01 UTC (Wed) by pabs (subscriber, #43278) [Link]

What about Juniper and MacOS?

Winning the copyleft fight

Posted Feb 17, 2016 10:40 UTC (Wed) by hummassa (guest, #307) [Link] (1 responses)

What about Sony PlayStation4?

Winning the copyleft fight

Posted Feb 17, 2016 16:26 UTC (Wed) by flussence (guest, #85566) [Link]

IIRC, *all* of that product line runs BSD to some extent.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 14, 2016 1:23 UTC (Sun) by b7j0c (guest, #27559) [Link] (82 responses)

Outside of the FSF, I typically find those licensing code with GPLV3/AGPL are only doing so because they realize these licenses are unpalatable to most, for whom they offer differently-licensed code...for a fee. Was this the goal of the FSF? To create a license that served only to motivate users to adopt corporate licensing while still allowing authors to ostensibly claim to be "free and open"?

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 14, 2016 9:27 UTC (Sun) by gioele (subscriber, #61675) [Link] (81 responses)

> Outside of the FSF, I typically find those licensing code with GPLV3/AGPL are only doing so because they realize these licenses are unpalatable to most, for whom they offer differently-licensed code...for a fee.

I like how the Qt devs see argue this point (<from http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-...>)

> LGPLv3 in essence prevents the creation of locked-down devices. So any Qt user will then either help the ecosystem by creating an open device that allows Qt developers to target a larger market, or help the ecosystem by funding further development of the product.

And the FSF is historically happy with people offering GPL exceptions for a fee: <https://www.fsf.org/blogs/rms/selling-exceptions>.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 14, 2016 17:09 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link] (80 responses)

> I like how the Qt devs see argue this point (<from http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-...;)
Well, they need money to eat. There's nothing wrong with it.

However, making your license toxic and offering a commercial option (via copyright assignment) is definitely not what I'd call "Free Software". It's at most a form of evaluation or shareware.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 14, 2016 21:09 UTC (Sun) by b7j0c (guest, #27559) [Link] (30 responses)

"shareware" licensing is exactly what it is. all of it enabled by the good name of the FSF and GPLv3/AGPL.

the problem is, if the FSF attempts to "fix" this with an amendment or new version of the GPL, they will eventually just paint themselves into a corner and they will have to admit that it is too late to undo the damage resulting from their bipolar approach that attempts to promote a militant version of "freedom" while also somehow accommodating corporate users

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 14, 2016 22:43 UTC (Sun) by rahulsundaram (subscriber, #21946) [Link]

>the problem is, if the FSF attempts to "fix" this with an amendment or new version of the GPL

No such "fix" is possible. Copyright holders will always have the legal right to license their code under any number of licenses they see fit.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 15, 2016 16:03 UTC (Mon) by mordocai (guest, #71668) [Link] (28 responses)

I may have missed something, but for the most part the FSF doesn't care about corporate users (for good reasons, corporate interests are generally opposed to FSF interests). The LGPL is the closest thing I've seen the FSF to come out with that seems to cater to corporate use cases.

In any case, many other free sofware supporters and I consider permissive licenses the ones causing the damage.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 15, 2016 18:21 UTC (Mon) by Cyberax (✭ supporter ✭, #52523) [Link] (27 responses)

> I may have missed something, but for the most part the FSF doesn't care about corporate users
And that's the reason GPL supporters are now begging for developers' time.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 15, 2016 18:48 UTC (Mon) by mordocai (guest, #71668) [Link] (26 responses)

The free software movement as a whole has always begged for developers' time.

I don't see a problem with this. It is the fallout of not being controlled by corporate greed.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 16, 2016 2:28 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link] (25 responses)

Not really. There was a good period during the early 2000-s when multiple companies (even Apple!) had been contributing to GPL projects like GCC, binutils, Samba and others.

That all has changed after the GPLv3 crusade.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 17, 2016 1:59 UTC (Wed) by rahvin (guest, #16953) [Link] (22 responses)

Though GPLv3 was contentious, IMO it had little to no effect on the issues you claim. LLVM would have developed regardless as Apple and other commercial companies wanted/needed a compiler they could cut commercial versions of because GCC destroyed vast swaths of compiler companies. LLVM has more to do with GCC destroying commercial companies who simply couldn't compete against a free compiler that was often better in every respect. These private companies needed a compiler on an equal footing with GCC technically that they could close source and offer extended features and plugins against. LLVM exists because of GCC's success, not because of a licensing issue. Have you ever thought about all the compiler companies that went under in the late 90's and 00's? The list is quite large. LLVM developed to fill the void created by all those specialty compilers going under.

There has been no effort (that I'm aware of) to replace Samba with a BSD licensed version and their likely never will, it's contribution profile has changed little since the switch. It serves only one purpose and that's inter-operation with Windows, hopefully something that won't be needed forever.

Personally I've seen little impact from GPLv3 beyond all the initial hullabaloo by all the corporate interests that didn't like the explicit patent license which was intended to subvert the main extortion avenue on GPL licensed code. You'd have a hard time listing more than a couple FOSS programs that switched to GPLv3 and were subsequently "replaced" in marketshare by a BSD licensed alternative. I personally can't think of a single example.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 17, 2016 2:34 UTC (Wed) by raven667 (subscriber, #5198) [Link] (12 responses)

> LLVM developed to fill the void created by all those specialty compilers going under.

Isn't this the same as saying GCC couldn't (or wouldn't) actually meet the needs that all those specialty compilers were fulfilling? Maybe I'm wrong but I don't think there is a large secondary market selling expensive LLVM plugins, instead LLVM is just being integrated where optimizing JITs are useful, like OpenGL shaders, something the GCC project intentionally made difficult during a critical time. It's a means to an end and a market where GCC is no longer competitive.

> I've seen little impact from GPLv3 beyond all the initial hullabaloo by all the corporate interests

Those corporate interests and the resources and developers they represent moved on to other non-copyleft licenses. If the goal of the GPL license sphere is to have fun, self-sustaining software platform for a few cognoscenti to use and improve then we are achieving it, if the goal is to change the default way that software is developed across the entire industry to copyleft, then we are failing, massively.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 9:34 UTC (Thu) by Del- (guest, #72641) [Link] (11 responses)

> Isn't this the same as saying GCC couldn't (or wouldn't) actually meet the needs that all those specialty compilers were fulfilling?

Yes, GCC fails to allow proprietary forks and extensions. Apple seems to have no interest in sharing the source code of xcode. LLVM almost became part of GCC, but Apple at the time felt strong enough to fight the same community that helped them back from the dead.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 10:23 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (10 responses)

> Yes, GCC fails to allow proprietary forks and extensions.
It's also incompatible with pretty much anything but command-line compilers. For example, LLVM is also used to compile shaders in Mesa - and GCC would be useless for this purpose due to its license.

> Apple seems to have no interest in sharing the source code of xcode.
Why should it? XCode is an IDE, after all.

And Apple shares their changes to LLVM, like the new Swift language.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 10:55 UTC (Thu) by Del- (guest, #72641) [Link] (9 responses)

> For example, LLVM is also used to compile shaders in Mesa - and GCC would be useless for this purpose due to its license.

Are you referring to Mesa's permissive license or GCCs technical abilities? Both can be changed you know. After all, LLVM was initially slated for inclusion in GCC, nothing prevents GCC from gaining such capabilities. By now, the Mesa developers should realise that they have nothing to gain from permissive licenses. They possibly have quite a bit to gain from copyleft though.

> Why should it? XCode is an IDE, after all.

Are you asking me why Apple should share code? This is indeed an unexpected question. Short answer, to stay ahead. Give me one freaking reason they need to keep xcode closed? Fewer people buying iphones maybe? I believe Apple's hostility to open communities will be their downfall. Both Apple and Microsoft can stay on top due to dominating market positions, but that will not be sufficient forever. It seems Microsoft is slowly learning this, while Apple is too full of themselves. To me, Clang/LLVM looks like a classic embrace, extend, extinguish.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 11:36 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (8 responses)

> Are you referring to Mesa's permissive license or GCCs technical abilities? Both can be changed you know.
Mesa needs to link with other applications directly, including proprietary ones. That means GPL is right out. If gcc wanted to change their license to BSD then yes, it might have been useful.

Meanwhile, GCC has released their JIT library and it's used by.... erm... nobody - exactly because of this.

> They possibly have quite a bit to gain from copyleft though.
No, they don't. They'll become pretty much irrelevant the moment they switch to GPL.

> Are you asking me why Apple should share code? This is indeed an unexpected question.
I'm asking why they should share code for XCode, which is their private IDE.

> To me, Clang/LLVM looks like a classic embrace, extend, extinguish.
Apple has been releasing their LLVM/clang changes for about 10 years by now. Including the changes that they could have kept private.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 15:17 UTC (Thu) by mlopezibanez (guest, #66088) [Link] (1 responses)

Meanwhile, GCC has released their JIT library and it's used by.... erm... nobody - exactly because of this.

I think you are trolling us (why Apple should share code? You know, because that is the whole ****ing point of FOSS? The four freedoms? Ethics? Co-operation? Social responsibility? Ah, you think the point of FOSS is maximising Apple's profits? Then, why should anyone who believes otherwise care about your opinion or advice? RMS is completely right when he says: Why take advice on pursuing your goals from people who are against those goals!?).

But I could not let the above lie stand. The reason nobody uses GCC JIT in production code yet is that:

libgccjit is currently of “Alpha” quality; the APIs are not yet set in stone, and they shouldn’t be used in production yet.

(Second line at https://gcc.gnu.org/onlinedocs/jit/)

I would not answer to any of your other comments, but any readers should be aware that you are thoroughly misinformed (or misinforming) on this topic.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 20:08 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

> I think you are trolling us (why Apple should share code?)
Why should Apple share code for Xcode? It won't really add significant benefit for users.

On the other hand, they DO release stuff that benefits other projects (clang support for autocomplete, for example).

Case in point: JetBrains released their community edition source code under GPL and nobody cares.

> But I could not let the above lie stand. The reason nobody uses GCC JIT in production code yet is that:
> libgccjit is currently of “Alpha” quality; the APIs are not yet set in stone, and they shouldn’t be used in production yet.
Well, that's because nobody is really using it. It's been in that form since late 2013. It's been officially released since Sep 2015, and it's still not being used for anything that is more than a code sample.

Want to be that it's not going to change?

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 15:52 UTC (Thu) by Del- (guest, #72641) [Link] (5 responses)

> Mesa needs to link with other applications directly, including proprietary ones. That means GPL is right out.

Linking exceptions are common. As for use of GCC, that should be no problem. GCC already compiles and links a plethora of proprietary code, and that is specifically allowed by the license. I see absolutely no advantage of permissive licensing in Mesa, and I have not seen you produce any argument to the contrary.

> Apple has been releasing their LLVM/clang changes for about 10 years by now. Including the changes that they could have kept private.

I believe that is not correct. I believe there were some ios related bits kept away, but I am not inclined to dig that up now. I am also pretty sure that the continued dominance of GCC keeps Apple in check here, so again, you should be grateful for the existence of a copyleft alternative.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 19:40 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (2 responses)

> Linking exceptions are common. As for use of GCC, that should be no problem. GCC already compiles and links a plethora of proprietary code, and that is specifically allowed by the license.
Well, yes. And relicensing GCC under BSD would also have helped.

But that defeats the very copyleft argument.

> I see absolutely no advantage of permissive licensing in Mesa, and I have not seen you produce any argument to the contrary.
Ask Mesa developers. Really, go to mesa-dev and ask them what they are thinking about switching to GPL.

> I believe that is not correct. I believe there were some ios related bits kept away
Nope. You can compile iOS applications with a stock LLVM. However, you're also going to need private headers and libraries that have nothing to do with LLVM itself.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 20:57 UTC (Thu) by Del- (guest, #72641) [Link] (1 responses)

> But that defeats the very copyleft argument.

You lost me, are you not aware that GCC is used to compile proprietary code, and that the license specifically allows it?

> Ask Mesa developers. Really, go to mesa-dev and ask them what they are thinking about switching to GPL.

GPL with an exception to allow, e.g., proprietary games to run? Sure, if I was involved in the project I would. Not being involved it is none of my business.

> Nope.

Are you by any chance working at Apple with LLVM? Otherwise I find your statement here provocative, how can you be so sure? In any case, not sharing bits from Clang and its standard libraries suffices for me to conclude.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 22:15 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

> GPL with an exception
You're defeating your own argument. GPL with exceptions is not a GPL. It's even incompatible with other pure GPL projects.

> Are you by any chance working at Apple with LLVM? Otherwise I find your statement here provocative, how can you be so sure? In any case, not sharing bits from Clang and its standard libraries suffices for me to conclude.
No, I don't work at Apple and I don't even use an iPhone. Apple shares all the relevant clang bits, the unshared bits are not related to the compiler or its base language libraries.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 1:54 UTC (Fri) by zlynx (guest, #2285) [Link] (1 responses)

> Linking exceptions are common. As for use of GCC, that should be no problem. GCC already compiles and links a plethora of proprietary code, and that is specifically allowed by the license.

Are you saying that the GCC JIT engine itself is allowed to be statically linked into proprietary code? I haven't looked, but I doubt the FSF would have done that.

One of the things LLVM is great for is embedding it so that binary code can be produced as needed from a cross-platform compiled intermediate form. And people don't want to GPL their game engines in order to do that.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 9:10 UTC (Fri) by Del- (guest, #72641) [Link]

> Are you saying that the GCC JIT engine itself is allowed to be statically linked into proprietary code?

I don't know, I haven't followed that project. From a quick glance I cannot find any linking exceptions, so this may be more involved than I thought.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 17, 2016 4:02 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (8 responses)

> These private companies needed a compiler on an equal footing with GCC technically that they could close source and offer extended features and plugins against. LLVM exists because of GCC's success, not because of a licensing issue
While there are proprietary compilers based on LLVM, they are a small niche. And before LLVM and clang they just used stuff like http://www.edg.com/

Nope, the move to LLVM+clang was very clearly caused by GPLv3. And before that, Apple had no problem with GPLv2 - they had timely support for gcc in their tools, up to the last GPLv2 version.

The same with Samba - Apple has just switched to a homegrown SMB implementation rather than using GPLv3 version of Samba.

> There has been no effort (that I'm aware of) to replace Samba with a BSD licensed version and their likely never will, it's contribution profile has changed little since the switch.
Yep. It's a niche software that becomes less relevant even in the Windows world.

From what I see, there are no interesting GPLv3 applications anymore. And by "interesting" I mean something non-trivial that has no liberally licensed counterpart.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 19, 2016 2:22 UTC (Fri) by rahvin (guest, #16953) [Link] (7 responses)

>Nope, the move to LLVM+clang was very clearly caused by GPLv3.

I won't agree with that. The market was ripe for another OSS compiler, far too many specialty compilers and compiler companies had gone under due to GCC. Though I might agree it was related I won't assign causality as I believe something like it would have developed either way due to the changing market eliminating options.

You simply couldn't write a closed source compiler anymore unless you could spend hundreds of millions on development to stay ahead of GCC. LLVM+clang levels the playing field for a lot of companies and takes the heavy lifting off. Hell in 5 years I wouldn't be surprised to see the Intel compiler or even the Visual studio compiler end up as forked LLVM, that is if they aren't already.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 19, 2016 3:06 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link]

> I won't agree with that. The market was ripe for another OSS compiler, far too many specialty compilers and compiler companies had gone under due to GCC.
How two of these are related?

There were several backends and frontends for proprietary compilers (Open64, EDG and others).

> You simply couldn't write a closed source compiler anymore unless you could spend hundreds of millions on development to stay ahead of GCC. LLVM+clang levels the playing field for a lot of companies and takes the heavy lifting off.
And why would Apple want to write a closed-source compiler in 2007? They are not in the business of selling compilers. Even their proprietary language (Swift) is open.

The causality is pretty clear on this point - Apple wanted to have a GPLv3-free system due to the anti-TiVo-clause. Just look at the recent Mac OS - there's literally NOTHING under GPLv3 there, even utilities like bash and rsync are pre-GPLv3.

And timing works out perfectly. clang development was started after the final drafts of GPLv3 had been made available.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 19, 2016 9:27 UTC (Fri) by excors (subscriber, #95769) [Link] (5 responses)

> Hell in 5 years I wouldn't be surprised to see the Intel compiler or even the Visual studio compiler end up as forked LLVM, that is if they aren't already.

The Intel C++ Compiler already has a Clang front-end for OS X (http://llvm.org/devmtg/2014-04/PDFs/Posters/ClangIntel.pdf), in addition to its original EDG front-end.

Visual Studio 2015 also has an optional Clang front-end (https://blogs.msdn.microsoft.com/vcblog/2015/12/04/clang-...), seemingly aimed at iOS and Android projects.

Both claim to have offered their Clang changes upstream, but I don't know what their current status is.

Both are still using their proprietary backends though, not LLVM's, presumably because they're better in some ways (likely performance). For frontends, I think "better" usually means "more compatible with my existing code and libraries", so standardising on Clang seems sensible.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 20, 2016 2:12 UTC (Sat) by rahvin (guest, #16953) [Link] (4 responses)

I was aware of the front end changes from Microsoft but not Intel. As I said I wouldn't be surprised to see them at some point drop their proprietary backends and just drop LLVM in with their own proprietary optimizations because they can't keep up with the OSS community. This is the problem I've been talking about, the OSS compilers, between GCC and LLVM are often way way ahead of all the proprietary compilers to the point of it not even being close.

Does Visual studio's proprietary compiler even support C99 at this point let alone C11? GCC and LLVM had C11 support before it was even official. Even the standards organizations are using them to develop the standards at this point. There reaches a point where you have to spend millions developing a compiler and you can't even keep up with GCC/LLVM on basic stuff let alone your own proprietary optimizations that are your selling point. Its at the point that there is little point trying to compete because you simply can't and your effort is better spent on plugin's, extensions or support on the OSS compiler.

I think it reached that point several years ago and the only real competitors to the OSS compilers at this point are the companies that produce compilers not for revenue but for other reasons, like Microsoft producing a compiler that is targeted specifically at windows or Intel using it to differentiate their processors. But even in those cases there is going to be some point in the future where these companies just decide it's easier to fork LLVM/clang and track their changes on top of the OSS codebase.

At least that's my opinion on the matter.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 20, 2016 14:13 UTC (Sat) by madscientist (subscriber, #16861) [Link] (3 responses)

Visual Studio 14 (2015) supports C99 (finally!) and virtually all of C++11. Visual Studio 13 supported a non-trivial subset of C++11 but still had significant holes, and didn't support C99 completely.

https://msdn.microsoft.com/en-us/library/hh567368.aspx

FWIW.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 20, 2016 18:31 UTC (Sat) by lsl (subscriber, #86508) [Link] (2 responses)

AFAIU, it doesn't support C99 at all. It only has the parts of C99 that are required for C++11 (the CPP, a long long type and most of the stdlib) and *supports them only in C++ mode*. When compiling C code you still get C89 and nothing else.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 20, 2016 20:17 UTC (Sat) by madscientist (subscriber, #16861) [Link] (1 responses)

You could well be right. I avoid Windows and Visual Studio as much as possible: I only know this much because I just made the effort to update our Windows development prerequisites at work to VS 14 so we could get better C++11 support on all our platforms. In doing so I had to undo some of the "hacks" we needed to get older VS versions to work, for example VS 14 finally supports snprintf() properly.

However our code is C++, not C. I assumed if it was supported in C++ it would be supported in C as well but maybe that was naive.

GPLV3 has been a boon for bait-and-switch licensing

Posted Mar 9, 2016 16:45 UTC (Wed) by nye (guest, #51576) [Link]

Basically MSVC is a C++ compiler that happens to have some support for compiling C in cases where that was something they could easily chuck in as an extra feature by reusing their existing C++ work.

Alright, technically that's not strictly correct, but I think it's a pretty close approximation.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 17, 2016 14:26 UTC (Wed) by mathstuf (subscriber, #69389) [Link] (1 responses)

> That all has changed after the GPLv3 crusade.

If GPLv2 had had GPLv3's patent and tivoization clauses when it was first released, would it have been as unwanted? Or is it more that the delta between GPLv2 and GPLv3 is what deterred these companies?

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 17, 2016 20:29 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link]

Who knows? Playing "what if" with history is rarely productive. Perhaps tivoization wouldn't have happened at all. Or perhaps TiVo would have used the BSD kernerl as the base catapulting it onto the first place, sidelining Linux.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 17, 2016 19:52 UTC (Wed) by ksandstr (guest, #60862) [Link]

The difference is that shareware is non-free software. It follows that those who'd pay for a non-copyleft license will receive none of the benefits of free software (e.g. third-party contributions that cannot be relicensed), and will likely suffer effects of non-free software in the licensing they receive instead.

That's to say: just as copyleft is copyright turned to serving software freedom, dual licensing is "try before you buy" turned to serving development of free software. Indeed there'll always be companies that won't ever use a piece of free software (generally because of management quirks), so why not make a dime off them instead of driving them to (presumably inferior) born-proprietary software?

On the downside, projects demanding authorization to relicense from contributors, in order to sell exceptional license to those contributors' work, are putting their monetary gain before their own gain from software freedom. Some go so far as to stop offering the GPL'd version after e.g. a change of ownership. This is where the GPL's freedom-to-fork comes in: often such companies won't even dare.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 18, 2016 0:27 UTC (Thu) by pizza (subscriber, #46) [Link]

> However, making your license toxic and offering a commercial option (via copyright assignment) is definitely not what I'd call "Free Software". It's at most a form of evaluation or shareware.

You really should clarify "commercial option" a bit -- Commercial licensing runs the gamut from "do whatever" on one end to being quite restricted (physically isolated from other code, dedicated developers who can't even so much as look at anything else that may be related, per-unit royalties, binary obfuscation requirements, DRM, mandatory auditing, etc etc) that the supposedly-toxic GPL would be lept upon in a heartbeat.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 22, 2016 1:06 UTC (Mon) by mlopezibanez (guest, #66088) [Link] (43 responses)

What you call "toxic" are terms to designed to protect the freedom of the users.

The FSF reasoning here is flawless: copyleft+selling-exceptions has the same effect as a non-copy-left license in terms of freedom (in both cases, a company creates non-free software from free software). But copyleft+selling-exceptions means that users can still get (and distribute) a copyleft version of the software if so they desire. Thus, in terms of freedom, it is strictly better than non-copy-left software. It also has the benefit of encouraging the development of copyleft software, since the developers get paid for it. It is still worse than only-copyleft software, since the exceptions still encourage non-free software.

https://www.fsf.org/blogs/rms/selling-exceptions

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 22, 2016 2:53 UTC (Mon) by Cyberax (✭ supporter ✭, #52523) [Link] (42 responses)

> What you call "toxic" are terms to designed to protect the freedom of the users.
And they're just as effective in doing this as these inane "Coexist" bumper stickers in achieving world peace.

> Thus, in terms of freedom, it is strictly better than non-copy-left software. It also has the benefit of encouraging the development of copyleft software, since the developers get paid for it.
And this is a straightforward doublespeak from Orwell. No, GPL doesn't encourage paid work. It discourages it, since in most cases companies can not use the result.

Selling exceptions is absolutely the opposite of what is important in Open Source software - collaborative work.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 23, 2016 0:02 UTC (Tue) by mlopezibanez (guest, #66088) [Link] (41 responses)

No company uses Linux, Qt, glibc, GCC? Nobody is paid to develop them? Companies do use GPL software and do pay for GPL software (for support, features, fixes). In which sense a company that wants to create non-free software from free software (either by means of GPL+exception or non-copyleft) is "collaborating"? Orwellian indeed.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 23, 2016 1:49 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link] (40 responses)

> No company uses Linux, Qt, glibc, GCC? Nobody is paid to develop them?
Linux is very special (it's firmly in GPLv2-only land). GCC is slowly withering. glibc is fairly minor, _and_ it's being replaced by other libcs like bionic and musl.

And QT, of course, is dual-licensed.

So yes, to a first approximation not many people are paid to work on GPL. Meanwhile, in the BSD/Apache world we're seeing an explosion of new projects, from LLVM+clang to Docker.

Personally, I've recently ditched my zoopark of build containers and instead replaced them with Alpine Linux which allows me to build GPL-free fully static binaries. I can't be happier.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 9:55 UTC (Thu) by Del- (guest, #72641) [Link] (39 responses)

> Linux is very special (it's firmly in GPLv2-only land).

I deeply disagree. The distinction between v2 and v3 is minor. Who benefits from exaggerating the differences?

> not many people are paid to work on GPL

By what metric? I believe there has never been more people paid to work on GPL code ever. The fact that software houses are trying to counter with open core models (be it permissive or dual license), and are employing even more developers in the attempt, only proves that GPL is succeeding. You should think long and hard on how the world would look without copyleft.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 10:28 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (38 responses)

> I deeply disagree. The distinction between v2 and v3 is minor.
No, it's not. Just ask Apple why they don't have _ANY_ v3 packages in Mac OS X. They even wrote their own SMB client rather than continue using Samba.

> By what metric? I believe there has never been more people paid to work on GPL code ever.
I highly doubt this. Outside of Linux development, not many GPL-ed projects are seriously active. Can you name a handful of new major GPL-ed projects?

While the rise of liberally-licensed projects is indisputable.

> and are employing even more developers in the attempt, only proves that GPL is succeeding.
"We're not retreating! We're just advancing in a different direction!"

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 11:05 UTC (Thu) by Del- (guest, #72641) [Link] (1 responses)

> Just ask Apple why they don't have _ANY_ v3 packages in Mac OS X.

Are you cherry-picking unconsciously? As I am sure you know, Apple do not allow GPLv2 software in their app store, and they have done all they can to remove gpl code from webkit totally independent of the v3 discussion. They did not share the webkit code willingly you know. Apple do not like GPL, they probably never did. It has very little to do with v3. Why do you try to exaggerate the difference between v2 and v3? Your gripe seems to be with both of them anyway.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 11:39 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

> Are you cherry-picking unconsciously? As I am sure you know, Apple do not allow GPLv2 software in their app store
I thought I was quite precise. Apple does not have any GPLv3 packages in _Mac OS X_.

Yet they have plenty of GPLv2 packages there. Some packages like bash and rsync are frozen at their last GPLv2 versions.

So yes, while v2->v3 changes might seem trivial to you, they are definitely NOT trivial for a lot of other people.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 11:36 UTC (Thu) by Del- (guest, #72641) [Link] (35 responses)

> Outside of Linux development, not many GPL-ed projects are seriously active.

You mean like git, samba, gcc, mariadb, pulseaudio, systemd, bash, debian, grub, glibc, mediawiki, wordpress, drupal, joomla, kolab, kde, gnome, and almost all user facing code on linux desktops? Have you noticed how core developers on KDE and Gnome are paid these days? How do you think that picture looked ten years ago? I see a lot of smaller companies making a good living out of copyleft these days, at least from my end of the universe it is growing quickly.

BTW, did you see how Suse these days have fantastic growth (around 20% y-o-y I believe), already the master piece for their owner? Seems some people think only Red Hat can earn money from GPL.

> While the rise of liberally-licensed projects is indisputable.

Yes it is, but it grew out of the success of the free software movement. Software companies trying to find their way to harness that success. Moreover, I do not see it as displacing copyleft for the most part. It is the latest attempt from software houses to compete against free software (if you can't beat them join'm or something like that). Yes I see it as a win for free software, at least as long as copyleft is doing well too. After all, as Bradley noted, we can fork it at any time.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 11:50 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (34 responses)

> You mean like git, samba, gcc, mariadb, pulseaudio, systemd, bash, debian, grub, glibc, mediawiki, wordpress, drupal, joomla, kolab, kde, gnome, and almost all user facing code on linux desktops?
Some of these projects you listed are LGPLv2.1 (systemd, wordpress, pulseaudio, drupal, joomla). Others are GPLv2 (like mediawiki and git).

I think only samba, gcc and bash are GPLv3. GCC is being replaced as a result and not many people seriously care about bash or Samba.

> BTW, did you see how Suse these days have fantastic growth (around 20% y-o-y I believe), already the master piece for their owner? Seems some people think only Red Hat can earn money from GPL.
RedHat controls the development of the GPL-ed projects they own. And it uses GPL to lock out other competitors from making it easy to compete. Certainly a viable business model, but there can't be many redhats on the market.

> Yes it is, but it grew out of the success of the free software movement.
And so?

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 16:13 UTC (Thu) by aggelos (subscriber, #41752) [Link] (8 responses)

You mean like git, samba, gcc, mariadb, pulseaudio, systemd, bash, debian, grub, glibc, mediawiki, wordpress, drupal, joomla, kolab, kde, gnome, and almost all user facing code on linux desktops?
Some of these projects you listed are LGPLv2.1 (systemd, wordpress, pulseaudio, drupal, joomla). Others are GPLv2 (like mediawiki and git). I think only samba, gcc and bash are GPLv3. GCC is being replaced as a result and not many people seriously care about bash or Samba.

Come now, this was in response to your

Outside of Linux development, not many GPL-ed projects are seriously active.

This is both shifting of the goal posts ("but they're not gplv3") and no true scotsman ("not many people seriously care about samba"). It pains me to have to point that out, as a reasonable discussion could be had on the issue of license choices the past decade. But no.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 19:48 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (7 responses)

> This is both shifting of the goal posts ("but they're not gplv3") and no true scotsman ("not many people seriously care about samba"). It pains me to have to point that out, as a reasonable discussion could be had on the issue of license choices the past decade. But no.
Ok. A simple challenge - list 10 major GPLv3 projects that:
1) Are non-trivial in size.
2) Are not dual-licensed.
3) Are seeing their development accelerating.

The ones I know of: gcc, Samba, emacs, glibc. That's pretty much it.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 22:41 UTC (Thu) by Del- (guest, #72641) [Link] (6 responses)

> list 10 major GPLv3 projects

and you for some reason want to disqualify the GPLv2+ projects, eventhough you know that they specifically provide the GPLv3 (and are FSF future proof with the +) for those of us who wants it? The fact that KDE and Gnome goes with GPLv2+ typically for both desktop and apps of course rules out many projects.

Still in addition to already mentioned grub and kolab, a web-search turned up octave, qcad, deluge, rsyslog, lftp, GNU make, GNU parallel, GNU mailutils, citadel, openmediavault, encfs, pts, hypercube, openshot, lives,openchange, darktable, ipfire and gimp. That's twenty one I believe. They are all doing well, either with stable or accelerated development. They are non-trivial in size, and afaik none of them follow a dual license path.

Why not ask for AGPLv3 while you are at it, then of course owncloud is the new kid on the block (yes, with paid developers), Kolab is also there with a couple of web-interfaces and paid developers. Both doing well, none with dual license model afaik.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 23:53 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (4 responses)

> and you for some reason want to disqualify the GPLv2+ projects, eventhough you know that they specifically provide the GPLv3 (and are FSF future proof with the +) for those of us who wants it?
GPLv2+ are semi-OK. They are protected from "pray-I-don't-alter-it-further" FSF, but extremely weakly. One wrong accepted patch and the project is toast.

> Still in addition to already mentioned grub and kolab, a web-search turned up octave, qcad, deluge, rsyslog, lftp, GNU make, GNU parallel, GNU mailutils, citadel, openmediavault, encfs, pts, hypercube, openshot, lives,openchange, darktable, ipfire and gimp.
Nope. Not good enough.

For example, OpenChange: https://github.com/openchange/openchange/commits/master - last commits are from year ago. It's dead, Jim.

GNU Make is not exactly _dead_, it just looks this way: http://ftp.gnu.org/gnu/make/ - last release in 2014. Ditto for parallel, mailutils and others.

For comparison, look at Hadoop. It alone has more changesets than your all other projects (-qcad) put together.

> They are non-trivial in size, and afaik none of them follow a dual license path.
At least Qcad does: http://www.qcad.org/en/license

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 9:47 UTC (Fri) by Del- (guest, #72641) [Link]

> For example, OpenChange: https://github.com/openchange/openchange/commits/master - last commits are from year ago. It's dead, Jim.

Just because I said I need a break, doesn't mean you can say anything you know. Facts is a difficult thing to dispute, here is a commit from december:
https://github.com/openchange/openchange/commit/03bae8ff2...
If you look through all branches, you will find 58 commits the last thirty days with three new contributors. Define dead please.

> GNU Make is not exactly _dead_, it just looks this way: http://ftp.gnu.org/gnu/make/ - last release in 2014. Ditto for parallel, mailutils and others.

No, make, mailutils and parallel are not dead, they are well maintained. All three sees activity. However, a program like make is mature, and already does what it is supposed to do, so yes, it is not the accelerated project you were looking for. And it has less frequent commits than mailutils and parallel. You really like cherry-picking, don't you? They are small projects, not dying, but small. If you keep it at that, I am happy. Next time around please be more specific when you state your qualifiers. Some of the others are rather large projects that are also doing well, so no, there is no ditto.

> For comparison, look at Hadoop. It alone has more changesets than your all other projects (-qcad) put together.

In terms of commits and lines of code, Hadoop scores well. Almost double of Octave, about a two thirds of Owncloud. In number of contributers the picture looks very different, here Hadoop has a comparatively low number. Can you please document how you did the additions to reach you conclusion? Yes, I am specifically looking for documented numbers, I am well capable of adding myself.

> At least Qcad does: http://www.qcad.org/en/license

You are right, my apologies, it should not have been on the list.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 16:03 UTC (Fri) by madscientist (subscriber, #16861) [Link] (1 responses)

> GNU Make is not exactly _dead_, it just looks this way: http://ftp.gnu.org/gnu/make/ - last release in 2014.

GNU make is the Abe Vigoda of software... except that doesn't work so well anymore :-(.

Maybe the classic: "reports of GNU make's death have been greatly exaggerated"... :-)

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 29, 2016 19:25 UTC (Mon) by bronson (subscriber, #4806) [Link]

I keep wishing it would die. I try out alternatives every few years, waste a bunch of time figuring out workarounds for oddball limitations.... and then keep reaching for gmake whenever spinning up a new C/C++ project. Argh. I'm not sure if the problem is with me or the universe.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 11:58 UTC (Sat) by paulj (subscriber, #341) [Link]

Hadoop may score better on your "current rate of change" metric. Just not sure that's a relevant one. On my "stuff I actually use on a daily basis" metric, GNU make, Octave score 100% and Hadoop 0%. There's a whole bunch of GPL software that way outscores the likes of Hadoop for me...

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 8:55 UTC (Fri) by tao (subscriber, #17563) [Link]

GPLv2+ is fine as long as no one forks it and makes it GPLv3.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 16:14 UTC (Thu) by Del- (guest, #72641) [Link] (24 responses)

> Some of these projects you listed are LGPLv2.1 (systemd, wordpress, pulseaudio, drupal, joomla).

Ah, I see. That can explain some of your statements. Actually, none of them are LGPLv2.1. Pulseaudio is a combination of GPL and LGPL, so you actually need to remove some of the functionality to get an LGPL version. Systemd is LGPLv2.1+. As for the other three, they are all released under GPLv2+, that is, both v2 and v3 with FSF given a prerogative to relicense at will. Checking out licensing information from Drupal is quite telling:
https://www.drupal.org/about/licensing#q1
they are quite specific in stating compatibility with AGPLv3.

> I think only samba, gcc and bash are GPLv3

Are you neglecting the GPLv2+ usually kept for historic reasons? Very few have followed Linus example in using GPLv2 only you know. From the corporate side of things my impression is that the corporate world want what ever patent protection available (except for patent trolls like Microsoft and Apple). Even so, both grub and kolab are also strict GPLv3+ in the short list I gave you. Actually, the web-interfaces in Kolab are AGPL licensed quite successfully. Yes they have a number of paid developers. I am on their paying customer list and have also contributed a little there in the past. Yep, there are more companies than Red Hat and Suse making good money on GPL software. I would love to see more pop up.

> RedHat controls the development of the GPL-ed projects they own.

AFAIK, Red Hat follows a "patches are welcome" policy with distributed copyright. Please document your claim or retract it. Maybe you were mistaking them for Canonical?

>> Yes it is, but it grew out of the success of the free software movement.
> And so?

I believe you still need free software to keep the permissive guys from locking things more and more up. While you seem to believe in the ideal that everybody understands the benefits of sharing, I am far less optimistic on behalf of humanity. Accordingly, I believe it is in your best interest to stop attacking copyleft. Please feel free to contribute or use any code you will though.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 19:58 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link] (23 responses)

> Ah, I see. That can explain some of your statements. Actually, none of them are LGPLv2.1.
Please, do your due diligence.

systemd is LGPL2.1 (no +) with static linking exception - https://github.com/systemd/systemd/blob/master/LICENSE.LG...

> As for the other three, they are all released under GPLv2+, that is, both v2 and v3 with FSF given a prerogative to relicense at will.
That's not true. All the contributions to wordpress are required to have GPLv2+ license, so FSF can't unilaterally change the terms since everybody always have the option to use GPLv2.

> Are you neglecting the GPLv2+ usually kept for historic reasons? Very few have followed Linus example in using GPLv2 only you know.
Actually, quite a lot of them did, even from your own list.

> AFAIK, Red Hat follows a "patches are welcome" policy with distributed copyright. Please document your claim or retract it.
No, they are quite ruthless in rejecting patches that don't square with their control policies and making it harder for downstream projects to use their code: https://lwn.net/Articles/430098/

> Maybe you were mistaking them for Canonical?
Let's talk about Canonical, then. Unity is released under GPL. Why it's not available in all other distributions? Isn't GPL supposed to be a magic pill against BSD-style fork-proliferation?

> I believe you still need free software to keep the permissive guys from locking things more and more up.
Sure. Go on and fork every BSD project on github and slap a GPL license on it. It'll help a lot.

Not.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 25, 2016 23:14 UTC (Thu) by Del- (guest, #72641) [Link] (22 responses)

> Please, do your due diligence.

I did, both wikipedia and freedesktop.org disagrees with you:
https://www.freedesktop.org/wiki/Software/systemd/
https://en.wikipedia.org/wiki/Systemd
if you even bothered to look at any of the actual systemd files on github, you would be less embarrased, see here, the + is clearly stated:
https://github.com/systemd/systemd/blob/master/src/core/a...
I belive you are confused by the relicensing from GPL to LGPL with you reference to the exception. From what I gather there is no exception to the LGPLv2.1+

> That's not true

Anybody can relicense a GPLv2+ code to GPLv3+, the FSF can relicense it to whatever they see fit. You should get your facts straight. I believe many of us are well aware of the trust we put in FSF when the + is there.

> Actually, quite a lot of them did, even from your own list.

What are you talking about now? Git is the *only* example on my list with only GPLv2, and guess ten times who chose that license. You really need to check your facts better. Yes there are other projects, after all he is quite influential. There are few and far between though. GPLv2+ is abundant on the other hand. Why do you even care, you hate both?

> they are quite ruthless in rejecting patches that don't square with their control policies and making it harder for downstream projects to use their code

Aha, so you are of the opinion that Red Hat is too restrictive in what patches to accept on projects their employees are maintaining. Your goal posts are moving faster than light.

> Sure. Go on and fork every BSD project on github and slap a GPL license on it. It'll help a lot.

I may, if I see it beneficial. Libreoffice and Jenkins are good examples of beneficial forks. Neither of them chose GPL, but they just as well could have.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 0:06 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link] (1 responses)

> if you even bothered to look at any of the actual systemd files on github, you would be less embarrased, see here, the + is clearly stated
I actually _linked_ the license: https://github.com/systemd/systemd/blob/master/LICENSE.LG...

> Anybody can relicense a GPLv2+ code to GPLv3+, the FSF can relicense it to whatever they see fit. You should get your facts straight. I believe many of us are well aware of the trust we put in FSF when the + is there.
You can't change the license on the existing code.

> Aha, so you are of the opinion that Red Hat is too restrictive in what patches to accept on projects their employees are maintaining. Your goal posts are moving faster than light.
RH does not always behave better than Apple.

> I may, if I see it beneficial. Libreoffice and Jenkins are good examples of beneficial forks. Neither of them chose GPL, but they just as well could have.
They forked dead projects. It's certainly OK to do that.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 13:50 UTC (Sat) by lsl (subscriber, #86508) [Link]

> I actually _linked_ the license

Yes, that's the LGPL v2.1 license text. At your option, you may also pick any later version, though.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 9:05 UTC (Fri) by tao (subscriber, #17563) [Link] (19 responses)

Sorry, but what are you smoking? The FSF can only relicense projects in two ways -- if the project is a FSF-project (and thus has contributor-licenses set up that reassigns the copyright to the FSF), or by issuing a new version of the GPL that supersedes GPLv3. And they explicitly promise (though didn't really live up to it with the GPLv3, in my opinion) that the new license is similar to the previous. So no, they cannot relicense to "whatever they see fit".

Let's say that the FSF "sees fit" to relicense a GPLv2+ project to "You have to kill small children to use this project". If it's a project with the copyrights assigned to the FSF, then yes, they can of course do whatever they want. But if it's someone else's GPLv2+ project, they cannot, because the new license would definitely not be in the spirit of the old one...

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 15:31 UTC (Fri) by Wol (subscriber, #4433) [Link] (18 responses)

> > Anybody can relicense a GPLv2+ code to GPLv3+, the FSF can relicense it to whatever they see fit. You should get your facts straight. I believe many of us are well aware of the trust we put in FSF when the + is there.

Actually, that is called FRAUD - tricking someone else into believing they can do something they aren't allowed to ...

> Sorry, but what are you smoking? The FSF can only relicense projects in two ways -- if the project is a FSF-project (and thus has contributor-licenses set up that reassigns the copyright to the FSF), or by issuing a new version of the GPL that supersedes GPLv3. And they explicitly promise (though didn't really live up to it with the GPLv3, in my opinion) that the new license is similar to the previous. So no, they cannot relicense to "whatever they see fit".

And this is wrong, actually, too. The arrival of GPL 4 *will* *not* change the licence on any existing GPL code, be it v2, v3, v2+, v3+ or whatever.

If you are the copyright holder, the licence(s) on the code is irrelevant, you can do what you like.

If you own the copyright on the licence (eg in this case you are the FSF) it gives you ABSOLUTELY NO RIGHTS WHATSOEVER to any code that may use said licence. So no, the FSF has no power whatsoever to relicence ANY code licenced under ANY version of the GPL.

If you are not the copyright holder, then you cannot change the licence on any code unless you have a licence that gives you permission. I am unaware of any Free or Open Source licence that gives you that permission ...

The reason people keep talking (and getting it horribly legally wrong) about relicensing is that

(1) The GPL gives an explicit guarantee that, if you distribute according to the terms of the GPL, you will also be complying with the terms of any other compatible licence that may be involved, and

(2) the GRANT OF LICENCE that is "v2+" or "v3+" or any other plus, gives the *distributor* the right to choose which version of the GPL they use. If I give you a grant of "LGPL v2.1+", then that gives you the right to use the GPLv2, GPLv3, LGPLv2.1, or LGPLv3 (have I forgotten any?). YOU HAVE TO PICK ONE - but it is your choice as distributor, not mine as licensor.

Which means they think they've relicensed the code, when in fact they haven't. They have only changed either (1) the set of licences that can be used, or (2) the choice of licence that THEY have used.

And as has been mentioned, the FSF is legally constrained that any new GPL must respect the principles of previous GPLs - namely that it has to support and fight for *user* freedom. (The FSF can't renege on this for a host of reasons, not least that all its copyright assignments would become null and void and land it in a huge copyright infringement mess :-)

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 15:42 UTC (Fri) by Wol (subscriber, #4433) [Link] (1 responses)

Let's give a quick example.

User A releases a program under BSD. Anybody can distribute it under BSD.

User B adds some mods and puts them in the public domain. Distribution is still BSD.

User C adds some code and grants an LGPL2.1+ licence. Distribution is now only possible under (L)GPL-style licences, but A's and B's code is still BSD and PD respectively.

User D adds some more code, under a GPL3+ licence grant. The set of licences is now reduced to GPL3 or any replacement.

User E adds some more code, under a GPL2+ licence grant. The set of licenses doesn't change, but if D is reverted the set of licenses would be GPL-style.

User F adds some code under a GPL2 licence grant. The set of licences is now empty so the code can't be distributed.

User G replaces user F's code under a GPL3 licence grant. The set of licences is now GPL3 and that's all.

Note that AT NO POINT WHATSOEVER have any licences been changed. The only thing that's changed is the licence choice available to the distributor.

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 15:44 UTC (Fri) by Wol (subscriber, #4433) [Link]

Whoops - I've been sloppy - I'm assuming when user F adds code, user E's changes have NOT been reverted.

If they have, the set of licences is reduced to GPL2, not null.

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 9:40 UTC (Sat) by Del- (guest, #72641) [Link] (15 responses)

> Actually, that is called FRAUD - tricking someone else into believing they can do something they aren't allowed to ...

My wording was unfortunate. If I had said fork rather than relicense, it would have been precise. My apologies.

> YOU HAVE TO PICK ONE

Hm, I am not sure about this one, so a clarification here is helpful (at least to me). My interpretation of GPLv2+ is that I as a recipient of the code can distribute it only providing the license text of GPLv3+. That is, I opt to have the code under GPLv3, and later, and that grants me the right to distribute it under said license. Of course, only copyright holders can enforce the license, so third parties may choose to ignore the relicensing if they know the history of the code.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 12:28 UTC (Sat) by Wol (subscriber, #4433) [Link] (14 responses)

No - this is the classic confusion.

The copyright holder has given you the OPTION of distributing EITHER under GPLv2, OR GPLv3. You have to choose. There is no v2+ or v3+ *licence*. (If GPLv4 comes out, then you can choose that.)

Your recipient then receives a LICENCE GRANT direct from the copyright holder, which means he gets v2+, same as you did. If it's got a + in it, then it's a GRANT not a LICENCE.

Which means you can give the code to him citing v3 as your licence, but he can then pass the code on to someone else citing v2 as the licence.

And if you tell him you have changed the licence grant to v3+ you are actually in breach of the GPL yourself! - you are obliged to pass on all the rights you received, including the right to distribute under v2, but you've just told him you've taken away that right!

There are three concepts you need to get clear in your head - the licence DOCUMENT, ie your copy of v2 or v3 or whatever. The GRANT OF LICENCE - the statement by the copyright holder telling you what licence(s) you can use, and the LICENCE ITSELF, which is the legal permission you use to distribute. This thread is full of confusion about which is what.

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 12:34 UTC (Sat) by Wol (subscriber, #4433) [Link]

NB - If you fork a v2+ project and add v3+ code OF YOUR OWN, of course now you can say that the PROJECT is v3+, because it is.

You can't, however, stop the next forker from reverting your 3+ changes and going back to v2+.

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 13:38 UTC (Sat) by paulj (subscriber, #341) [Link] (12 responses)

Is there a meaningful distinction between "grant" and "licence"?

The way GPLvX+ works is that it is two stage - and there's a level of indirection:

- The source contains a brief licence, stating the licence is as per a published licence document, e.g. GPLvX "or any later version" (which is a term defined by the GPLvX).

- The licence document itself then defines what that "licence" to use "any later version" means exactly.

These are all conditional permissions granted by the copyright holder - and "permission" is the root meaning of "licence". Without following the conditions necessary for such permission, then one risks infringing the copyright held by the author(s), who may then sue you for damages.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 14:42 UTC (Sat) by Wol (subscriber, #4433) [Link] (11 responses)

Yes there is a meaningful distinction between "grant" and "licence". We muddy the water horribly when we use the two words interchangeably.

Like in photography, another bugbear of mine. The terms "aperture" and "f-stop" are used interchangeably, but f-stop is a dimensionless ratio, and *everyone* *else* measures apertures in units of distance.

To be pedantic, the "grant" (usually in the COPYING file) tells you which licence(s) you can use, while the "licence" (in the LICENCE or GPL-LICENCE-V2" file) tells you what you can/must do.

And it's actually very important, because if I grant you the right to use "v2 or later", while you are free to use v3 you MUST tell your recipients "v2 or later" because v3 tells you to!

By the way, can you point me to the text of the GPLv2+ licence? Or the GPLv3+ licence? NO YOU CAN'T. Because it only appears as an advisory paragraph (ie it has no legal force whatsoever) at the ?end? of GPLv2 or GPLv3.

At the end of the day, nobody really cares that much about "the v2+ licence" or "the v3+ licence" because, usually, we have no need to care. But if you want to UNDERSTAND what is actually going on, that's when you need to care, and that's when the distinction is important.

And as I said, it's actually legally quite serious - it is a *breach* *of* *the* *GPL* if you change v2+ to v3+ !!! (on someone else's code, that is - not if you add new code that changes the overall licence on the derived work.)

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 16:31 UTC (Sat) by Del- (guest, #72641) [Link] (6 responses)

> And it's actually very important, because if I grant you the right to use "v2 or later", while you are free to use v3 you MUST tell your recipients "v2 or later" because v3 tells you to!

Can you point me to where this is stated. I do believe you, but it would be nice to see the writing.

> And as I said, it's actually legally quite serious - it is a *breach* *of* *the* *GPL* if you change v2+ to v3+ !!! (on someone else's code, that is - not if you add new code that changes the overall licence on the derived work.)

This is a really minor one though. If the recipient is aware of this, then the recipient can easily add a couple of lines to the code base just to do the relicensing. As such, for all practical reasons, my original statement that anybody can relicense actually stands. In other words, bypassing any commitment to keep GPLv2 when you re-distribute is trivial. Enforcing the GPLv3+ license is still left to the copyright holders of course, but that is actually a different matter.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 19:14 UTC (Sat) by Wol (subscriber, #4433) [Link] (5 responses)

> > And it's actually very important, because if I grant you the right to use "v2 or later", while you are free to use v3 you MUST tell your recipients "v2 or later" because v3 tells you to!

> Can you point me to where this is stated. I do believe you, but it would be nice to see the writing.

Fifth paragraph of the preamble to GPL version 3 ...

"For example, if you distribute copies of such a program, whether gratis or for a fee, you must pass on to the recipients the same freedoms that you received. You must make sure that they, too, receive or can get the source code. And you must show them these terms so they know their rights."

To emphasise - "you must pass on to the recipients the same freedoms that you received" - ie the freedom to distribute under v2.

> If the recipient is aware of this, then the recipient can easily add a couple of lines to the code base just to do the relicensing.

Which is likely to fall foul of copyright law and be classed as "too trivial to be copyrightable". It pretty much falls in the same category as a brain-dead update of the licence statement and copyright date in all source files - which is legally almost the same thing AT&T did when they messed about with all the copyright statements in Unix way back when - and which a Judge ended up ruling pretty much that they had abandoned their own copyright interest thanks to their own stupidity.

> Enforcing the GPLv3+ license is still left to the copyright holders of course,

Said copyright holder - singular, only copyright holder - being YOU! And as I said above, the change is probably not copyrightable if it's that trivial...

You are presuming that you have the right to alter someone else's grant of licence on their code, and as I have said elsewhere, I do not know of ANY Free or Open Source licence that gives you that power.

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 28, 2016 10:09 UTC (Sun) by Del- (guest, #72641) [Link] (4 responses)

> To emphasise - "you must pass on to the recipients the same freedoms that you received" - ie the freedom to distribute under v2.

Yes, this was the part I thought, thank you. I don't mean to be difficult, but freedom is open for wide interpretation. Actually, I have seen some argue that permissively licensed code is more free than copyleft. For this reason, the FSF has been very specific in stating a definition of freedom in relation to software, as provided by, e.g., wikipedia:
https://en.wikipedia.org/wiki/The_Free_Software_Definition
it does in no way imply GPLv2. GPLv3 preserves the same freedoms found in GPLv2. Hence, I am afraid your argument falls apart.

> Which is likely to fall foul of copyright law and be classed as "too trivial to be copyrightable".

Indeed, again my precision was way off, my apologies. The added code lines should be sufficient addition to warrant copyright, and trivial additions do not apply here. I still stand by my claim though. For any software you yourself want to distribute, it is trivial to add a non-trivial patch (non-trivial enough to warrant copyright), and hence enable relicensing (within the context discussed here, i.e., from GPLv2+ to GPLv3+). However, I still cannot see a clear argument stating that a recipient of GPLv2+ cannot relicense to GPLv3+ upon redistributing, seeing your argument above seems to fail.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 29, 2016 12:17 UTC (Mon) by tao (subscriber, #17563) [Link] (3 responses)

Sorry, but no, the freedoms granted by the GPLv2 are NOT the same as those granted by the GPLv3.

With GPLv2 someone can "Tivoize" a device. With GPLv3 you cannot. It might not be a freedom that the FSF (or you) likes, but it's a freedom never the less.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 29, 2016 13:50 UTC (Mon) by anselm (subscriber, #2796) [Link] (2 responses)

Del- was talking about “freedoms” in terms of the FSF's Four Freedoms. In that context the GPLv2 and GPLv3 are equivalent in that they are both designed to preserve these. Being able to “tivoize” a device is not one of the Four Freedoms.

GPLV3 has been a boon for bait-and-switch licensing

Posted Mar 1, 2016 1:41 UTC (Tue) by Wol (subscriber, #4433) [Link] (1 responses)

Unfortunately, in this case, it's the black letter of the licence that rules, not the intent of the people who wrote it.

In any case, saying "I can distribute under v2, but you can't", is placing restrictions on what your downstream is allowed to do, which is a clear breach of the principles behind the GPL. (And, in telling other people what they can and cannot do with code that is not your own, you're exercising powers that copyright law does NOT give you.)

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Mar 1, 2016 8:58 UTC (Tue) by Del- (guest, #72641) [Link]

> Unfortunately, in this case, it's the black letter of the licence that rules, not the intent of the people who wrote it.

For legal documents like a license, it is very common (I would even say necessary) to supply specific definitions to concepts (like freedom) that are open to interpretation. The GPL does this right above the paragraph you quoted:

"The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change all versions of a program--to make sure it remains free software for all its users. We, the Free Software Foundation, use the GNU General Public License for most of our software; it applies also to any other work released this way by its authors. You can apply it to your programs, too.

When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for them if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs, and that you know you can do these things."

The black letter does not seem to support your interpretation.

> in telling other people what they can and cannot do with code that is not your own, you're exercising powers that copyright law does NOT give you.

Generally speaking this is clearly not a true statement. GPL is called copyleft exactly because it gives the recipient of the code powers through copyright law.

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 27, 2016 20:09 UTC (Sat) by paulj (subscriber, #341) [Link] (3 responses)

I still don't understand the difference. AFAICT you receive the "licence" (i.e. permission) from a copyright holder who "grants" you that licence (the "licence" that is granted, in the same sense as "to grant permission"). For the GPL, the normal process is that the GPL licence is granted by the copyright holder via 2-stages and using an indirection via referral to the FSF to achieve both consistency and (potential) "upgrade" ability.

Stage 1 is that the copyright holder of a work makes it be known that the work is licensed under the terms of the GPLvX, as published by the FSF, and as usually included alongside the work in a file called "COPYING", but otherwise obtainable directly from the FSF.

Stage 2 is the GPLvX licence document, which documents those terms fully.

Here's stage 1, as per the recommended GPLv2 text, to be attached to a work (e.g. included at the top of a file):

"This file is part of $WORK.

$WORK is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2, or (at your option) any
later version.

$WORK is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
General Public License for more details.

You should have received a copy of the GNU General Public License
along with GNU Zebra; see the file COPYING. If not, write to the Free
Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
02111-1307, USA. "

Then stage 2 is the FSF-published GPLv2 licence.

The initial licence is granted by the copyright holder, and the GPLv2 is part of that licence, by reference and reference to the FSF. Additionally, the licence granted may give the option to use any later version of the GPL than the referenced one, as published by the FSF.

"Licence" - a noun, the set of (conditional) permissions given/granted.

"Grant" - a verb, as in "to grant permission" or "to grant a licence" (licence being an old word for permission, deriving from latin), also as in "to give".

Was my understanding.

The copying file is _NOT_ "the grant", it is a legal document setting out conditions and permissions which may be given/granted as the licence to a work - except as an imprecise short-hand that is nounifying a verb. :)

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 28, 2016 0:02 UTC (Sun) by Wol (subscriber, #4433) [Link] (2 responses)

> The copying file is _NOT_ "the grant", it is a legal document setting out conditions and permissions which may be given/granted as the licence to a work - except as an imprecise short-hand that is nounifying a verb. :)

"except as an imprecise short-hand that is nounifying a verb"

Or is it a noun that has been verbified? "a grant" to me is perfectly normal (indeed, *ancient*, as in heraldry) usage.

At the end of the day, I think we see it pretty much the same - it's a two-stage thing (and if you don't understand that it is two stages, then you'll misunderstand what's going on - that is the crucial point).

But imho the GPLv2, or v3, which tells you what you can do, and what your obligations are, is the licence; and the COPYING file is what grants you permission to use the licence.

Going back to heraldry, I have in my possession a "grant of arms", which describes my grandfather's arms, and gives him permission to use them. The grant is very much separate from the arms themselves.

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 28, 2016 0:18 UTC (Sun) by Wol (subscriber, #4433) [Link]

And just to add, just because I enclose a copy of GPLv2 (or whatever) with my code, doesn't mean you can use the GPLv2 to copy my code.

You need a separate grant of permission to use the GPLv2. Your stage 1 GRANTS the end user permission to use the GPL (version as specified), which is why I call it a grant! :-) Because it is :-)

Cheers,
Wol

GPLV3 has been a boon for bait-and-switch licensing

Posted Mar 6, 2016 16:58 UTC (Sun) by paulj (subscriber, #341) [Link]

The "grant of arms" thing you're talking about is just showing that using the verb "grant" as a shorthand to refer to the documentation of the action (verb) of a grant of some right goes back a long way. The root meaning though is:

To grant a permission/right.

Each can be documented separately, and one might use "grant" as a shorthand for that documentation, but that remains a convenient, context-specific shorthand for what is the documentation of the act of the grant.

Don't eat money

Posted Feb 23, 2016 21:18 UTC (Tue) by bkuhn (subscriber, #58642) [Link] (1 responses)

> Well, they need money to eat.

I don't recommend eating money. I swallowed a penny once when I was a child. It's definitely a mistake.

Don't eat money

Posted Feb 23, 2016 21:42 UTC (Tue) by bronson (subscriber, #4806) [Link]

"I'd like to see some change in you!"

GPLV3 has been a boon for bait-and-switch licensing

Posted Feb 26, 2016 19:45 UTC (Fri) by Wol (subscriber, #4433) [Link]

> However, making your license toxic and offering a commercial option (via copyright assignment) is definitely not what I'd call "Free Software". It's at most a form of evaluation or shareware.

But, done right, it's perfectly in keeping with the goals of the FSF - ie maintaining the four freedoms. Like Qt code is poison-pilled to go BSD if there's a hostile takeover.

If I wanted to do that (assignments etc) I'd make it a condition of the assignment (that is, other people giving me their code) that ANY product containing their code is ALWAYS available under the GPL. In other words, I can sell closed licences, but I can't close the product because that would void my right to use their code.

Cheers,
Wol

Winning the copyleft fight

Posted Feb 14, 2016 12:14 UTC (Sun) by HIGHGuY (subscriber, #62277) [Link] (30 responses)

I think the smarter thing to do is actually the opposite: if you can't beat them, join them.

Relevant projects should start a non-profit organisation, employ their developers (even for contract-work) and start doing something with their project. Things like: paid support (bug bounty?), development services, etc.
For smaller projects there could even be some umbrella organisation taking care of multiple projects.

Companies will like this, because they:
- can contract developers that know the project (no study-phase)
- have someone to shout at
- basically get services at near-actual cost (because of the non-profit part)
- know that upstreaming isn't going to require a rewrite because it's written by or with close co-operation with upstream
- know that the maintenance of that code is externalised in the long run
- can share the cost of certain developments (especially smaller companies)
- can benefit of others' work too

Developers, especially those that aren't paid at this time, get something out of it (and can perhaps scale up their contributions). The major downside of GPL that remains is that GPL levels the playing field. Everyone gets the same benefits at the same time.
Therefore, personally, I wouldn't mind an initial offering where delayed-upstream is a part of. For example: the cost could be exponential, where 3m is doable, but 1yr is getting expensive even for larger enterprises. But even there, we'll need to teach them that having things upstream faster also means more real-world exposure and faster stabilisation of their code (i.e. an external QA dept).

Winning the copyleft fight

Posted Feb 15, 2016 9:22 UTC (Mon) by mjthayer (guest, #39183) [Link]

Semi-on-topic - I was interested to learn at FOSDEM that the MariaDB have agreements with the various enterprise Linux distributions, who subcontract the support work to them that they (the distributions) offer to their customers.

Winning the copyleft fight

Posted Feb 16, 2016 18:03 UTC (Tue) by fandingo (guest, #67019) [Link] (28 responses)

It's nowhere near that simple.

> - basically get services at near-actual cost (because of the non-profit part)

Nonprofit doesn't mean that overhead is the same or lower than for-profits. In fact, I'd expect a ton of operational inefficiency dramatically driving up costs. You'll need a whole cadre of business and support people to make this work. Who's going to manage them?

> - know that the maintenance of that code is externalised in the long run

I really, really hate this argument. You don't get to dump your code upstream and expect them to maintain it forever. This is probably the most nonsensical talking point that free software advocates get wrong. While I don't want to say that a contribution makes it your pet forever, projects are generally uninterested in contributions made for the sake of offloading all maintenance work. Maybe if you've got something really compelling.

But, yes, in general, I agree with your sentiment. We need to look at new options to draw in corporations to open source development, and beating them over the head with copyright lawsuits and C&D letters will, instead, drive them away as fast as they can run.

Winning the copyleft fight

Posted Feb 16, 2016 20:00 UTC (Tue) by raven667 (subscriber, #5198) [Link] (23 responses)

> Nonprofit doesn't mean that overhead is the same or lower than for-profits.

Sure it does, most non-profits pay less than for-profit companies for the same work and in the absence of a motivation to maximize profits will price their services much closer to their actual cost. Of course you can find fraudulent non-profits, churches too, which collect donations and spend lavishly on compensation for executives instead of a worthwhile cause, but you find that kind of fraud in any endeavor, private business as well.

> In fact, I'd expect a ton of operational inefficiency dramatically driving up costs.

Why? Other than some preconceived ideology which says profit is good and charity is bad I don't see any reason why this would be the case, or demonstration that this is the case in the real world.

Winning the copyleft fight

Posted Feb 16, 2016 22:38 UTC (Tue) by fandingo (guest, #67019) [Link] (14 responses)

> Sure it does, most non-profits pay less than for-profit companies for the same work and in the absence of a motivation to maximize profits will price their services much closer to their actual cost. Of course you can find fraudulent non-profits, churches too, which collect donations and spend lavishly on compensation for executives instead of a worthwhile cause, but you find that kind of fraud in any endeavor, private business as well.

Nonprofits also have far less incentive to maximize efficiency because increased efficiency doesn't correspond to any improvement in the business. There's really nothing to argue about: Look at our economies. They're dominated by for-profits, even when nonprofits are given lavish, unconscionable tax waivers. This has absolutely nothing to do with fraud, which doesn't make up a meaningful portion of either nonprofit or for-profit business. It's about the difference in incentives and how it leads each type of corporation down different roads.

> Sure it does, most non-profits pay less than for-profit companies for the same work

Huh? This doesn't even make sense, unless you're saying that nonprofits receive a discount from their suppliers. I think what you meant to say was "most goods and services sold by non-profits *cost* less than equivalent goods and services from for-profits." If so, obligatory citation needed. I'm not aware of any studies that even attempt to answer this question, probably because "equivalent good and services" is an almost impossible precondition.

>> In fact, I'd expect a ton of operational inefficiency dramatically driving up costs.
> Why?

Umm because a bunch of open source software *developers* would be terrible at this. What expertise do they have in running a support business? How do you organize a group of volunteers to be able to deliver what the paying customer wants? Who's going to keep the books, deal with multinational tax issues, write/review contracts, etc.? You must have some structure in place that just manages this business organization, and I can't fathom who would find that tedium interesting enough to do for free. Plus, it's not really something you can do on nights and weekends; your support business needs to have people available when the customer desires. Being good with computers doesn't have anything to do with being good at running a business -- for-profit or not.

Winning the copyleft fight

Posted Feb 16, 2016 22:57 UTC (Tue) by pizza (subscriber, #46) [Link] (8 responses)

> Nonprofits also have far less incentive to maximize efficiency because increased efficiency doesn't correspond to any improvement in the business.

Just because the organization's purpose is not to make a profit for its investors, that doesn't mean they have no incentive to improve efficiencies. After all, if your lower your expenses/overhead, that leaves more money for your actual mission.

Winning the copyleft fight

Posted Feb 17, 2016 1:54 UTC (Wed) by raven667 (subscriber, #5198) [Link] (7 responses)

Yeah, that premise seemed wrong to me too, after all most non-profits subsist on donations, which are harder to get than payment for services rendered, so have a very small amount of money to work with compared to the task they are trying to do, so efficiency is hugely important. Most non-profits I've had any contact with try to cut every possible corner to keep as many resources as possible focused on their primary mission.

Winning the copyleft fight

Posted Feb 17, 2016 2:25 UTC (Wed) by rahvin (guest, #16953) [Link] (3 responses)

You are missing his main point. There are base business expenses, things that you have to do if you operate as a business in this world, non-profit or for profit matters little in this. Payroll, taxes, unemployment insurance, OSHA compliance, etc. This list is very very long. These all cost money, lots of it. It's what comprises the vast majority of any business overhead expenses (the stuff other than salaries). They also take a lot of time, time that no one in their right mind would ever do for free. It's tedious boring and in the case of taxes (non-profits are not exempt from taxes) and other government compliance issues holds significant liability for messing up including potential prison time.

There is no evidence, none, that a non-profit can do this better than a for profit business. As these expenses often comprise more than the salaries being paid they are a significant cost of doing business and are essentially more important than the salaries being paid as far as cost management. Any business can cut corners to reduce cost (something that isn't necessarily a good thing).

The success or cost basis of a company has little to nothing to do with profit status of the corporate entity. The cost basis is almost always a function of management and luck. All things being equal the non-profit should be cheaper but they are never equal, as mentioned by the parent there is no proof whatsoever that non-profits are cheaper or function better. The type of company discussed would be a nightmare to manage, the more difficult managing the more overhead goes up. The most efficient businesses are highly specialized. What has been proposed here is the exact opposite.

Winning the copyleft fight

Posted Feb 17, 2016 3:10 UTC (Wed) by raven667 (subscriber, #5198) [Link]

> You are missing his main point.

Probably, I'm not the brightest tool in the shed. I thought the main point was that non-profits were by definition wasteful and disgusting organizations and the only kind of real organization to be proud of was a for-profit one. I may have

> There are base business expenses, things that you have to do if you operate as a business in this world, non-profit or for profit matters little in this.

Sure, but non-profits generally don't get to just throw money at these problems to make them go away, so they have be creative and have a lot of hustle in how they check off each of these overhead costs. Maybe they pay for a CPA but get a volunteer lawyer, or vice versa, maybe they outsource to a service which does these things, maybe they try to do it in-house and avoid writing a check, accepting the risk of doing it wrong figuring that enforcement is probably lax or that they have no real assets to seize anyway. Many non-profits are able to use the cost savings of having volunteers to be able to afford paying these fixed overhead costs. Non-profits which can't cover these costs fold, or downsize.

> There is no evidence, none, that a non-profit can do this better than a for profit business.

Maybe the better measurement axis is the size of the of the revenue, an organization flush with cash, profit or no, will often trade money for cognitive load, spend so as to not have to think, leading to hugely inefficient and stupid outcomes, whereas a leaner organization will spend much more time trying to be creative and save their limited resources. On that axis I think non-profits tend to be leaner organizations focused on their mission (obviously not always) whereas for-profits are more likely focusing on getting big and fat (although there are plenty of examples of charismatic leaders pushing organizations to stay lean as they grow). Money can hide a ton of sins in organizational structure.

Winning the copyleft fight

Posted Feb 17, 2016 8:50 UTC (Wed) by paulj (subscriber, #341) [Link] (1 responses)

All these tedious, boring things you list that are such core parts of running a business are hence also things for which there exist plenty of service providers who will manage nearly all of that for you. Those service providers have an incentive to do so reasonably efficiently - they at least provide a ceiling on the cost of carrying out these chores in an adequate manner.

Winning the copyleft fight

Posted Feb 17, 2016 11:57 UTC (Wed) by HIGHGuY (subscriber, #62277) [Link]

Sounds like CaaS: Company as a service ;)
But that was exactly what I was talking about with "some umbrella organisation taking care of multiple projects."
Basically pay for the services you need.

Winning the copyleft fight

Posted Feb 17, 2016 2:46 UTC (Wed) by fandingo (guest, #67019) [Link] (2 responses)

> Yeah, that premise seemed wrong to me too, after all most non-profits subsist on donations

Not the kind we're talking about. The subject is nonprofit corporations who charge for goods and services, not charities.

Winning the copyleft fight

Posted Feb 17, 2016 3:19 UTC (Wed) by raven667 (subscriber, #5198) [Link]

Yeah, I was clearly thinking of both at the same time, my mistake.

Winning the copyleft fight

Posted Feb 26, 2016 15:55 UTC (Fri) by Wol (subscriber, #4433) [Link]

Well, I worked for a trade association (a non-profit) and we had plenty of for-profit competition. So we had to watch costs or we'd get priced out of the market.

Where we scored was value-add - we had a lot of not-very-expensive things we'd do that provided great value to our members. My favourite example was the first Gulf War. When it all blew up we rang round our members with interests in the region and said "we know as little as you do, but our conference room is free. We're inviting everyone affected to come and share notes".

When you're an HR Exec wondering what on earth you do to look after your staff in such a disaster, to be presented with the opportunity to meet loads of other people in the same boat is not to be sniffed at ...

Cheers,
Wol

Winning the copyleft fight

Posted Feb 17, 2016 2:16 UTC (Wed) by raven667 (subscriber, #5198) [Link] (4 responses)

> Nonprofits also have far less incentive to maximize efficiency because increased efficiency doesn't correspond to any improvement in the business.

Non-profits have a much harder time generating revenue than businesses so with less money to work with they have to be highly efficient to get the most resources dedicated to their actual mission, there is a lot of pressure for efficiency, the easiest way to grow the mission is to be more efficient with the funds you already have.

> Look at our economies. They're dominated by for-profits, even when nonprofits are given lavish, unconscionable tax waivers.

WAT? I'm not sure what the point is you think you are making here, maybe you should be more explicit rather than assuming I know what you are thinking.

> This has absolutely nothing to do with fraud, which doesn't make up a meaningful portion of either nonprofit or for-profit business.

Sure it does, whole industries in for-profit land are dominated by fraud, we just had a financial system crash a few years ago because of rampant, unchecked fraud in the system, which is still rampant and unchecked. There are plenty of fraudulent non-profits as well, using donations to pay for lavish perks for executives, or just straight-up pocketing the money, which is why many non-profits publicly document how much is spent on overhead vs. mission to let donors make informed decisions as to what they support.

>> Sure it does, most non-profits pay less than for-profit companies for the same work
>Huh? This doesn't even make sense,

Sorry I wasn't clear, I meant in payroll, non-profits usually pay (far) below market rates for work to direct employees. Employees are making a trade off of income vs. doing important work and take part of their compensation in personal satisfaction instead of cash. This is one way non-profits are more efficient than for-profit businesses for the same head count, in fact many people work for free, satisfaction is sufficient such that no money changes hands. This would be insane (and abusive) for a for-profit company.

> Being good with computers doesn't have anything to do with being good at running a business -- for-profit or not.

Well sure, you don't measure a fish by how well it climbs trees, but any non-profit more than a handful of people hires at least some professional executive staff to do the business part of it, although compensating and retaining those people is a constant struggle. Non-profits tend to be very efficient in this way, accomplishing their missions with minimal management, rather than the army of do-nothings that most for-profit companies acquire into middle management. Dilbert isn't written about a non-profit.

> It's about the difference in incentives and how it leads each type of corporation down different roads.

I think you have some very incorrect assumptions about how those incentives actually work in practice with real people and real organizations, which leads you to bad predictions of their effects.

Winning the copyleft fight

Posted Feb 17, 2016 2:44 UTC (Wed) by fandingo (guest, #67019) [Link] (3 responses)

> Non-profits have a much harder time generating revenue than businesses so with less money to work with they have to be highly efficient to get the most resources dedicated to their actual mission, there is a lot of pressure for efficiency, the easiest way to grow the mission is to be more efficient with the funds you already have.

Source? I think you're conflating charities with nonprofits. A straight nonprofit corporation shouldn't have any more difficult than a similar for-profit generating revenue if their offerings are as enticing.

> Sure it does, whole industries in for-profit land are dominated by fraud

Umm, ok. This is not a political discussion forum, but your viewpoint is naive if you think the financial crisis was primarily caused by fraud. It was caused mostly by dealing in instruments too complex to understand, which resulted in far more actual risk than intended and risk mitigation that couldn't deal with systematic failures. It was tremendously complex problem with over leverage, and there were a bunch of other factors, but fraud was not a meaningful cause.

Winning the copyleft fight

Posted Feb 17, 2016 3:18 UTC (Wed) by raven667 (subscriber, #5198) [Link]

> I think you're conflating charities with nonprofits.

I think you are right, I was kind of conflating the two.

> This is not a political discussion forum

True enough, but if you make decisions based on beliefs which are not based on facts you are likely to be burned by surprise, finding the right answer is more important than "being right".

Winning the copyleft fight

Posted Feb 17, 2016 8:47 UTC (Wed) by spaetz (guest, #32870) [Link] (1 responses)

>> Non-profits have a much harder time generating revenue than businesses so with less money to work with they have to be highly efficient to get the most resources dedicated to their actual mission, there is a lot of pressure for efficiency, the easiest way to grow the mission is to be more efficient with the funds you already have.

> Source? I think you're conflating charities with nonprofits. A straight nonprofit corporation shouldn't have any more difficult than a similar for-profit generating revenue if their offerings are as enticing.

Source: "Anatomy of the Nonprofit Starvation Cycle - An Analysis of Falling Overhead Ratios in the Nonprofit Sector"
http://m.nvs.sagepub.com/content/44/3/539.abstract?sid=16...
or any other nonprofit journal, really. Donors care about administrative overhead, and non-profits dependent on donations run on the lower end of the possible overhead. If you mean non-profit corporations, that do not depend on donations, you might be right. But then, do you consider IKEA (which is officially owned by a non-profit foundation) REALLY as a good example of a non-profit corporation?

Winning the copyleft fight

Posted Feb 17, 2016 16:15 UTC (Wed) by fandingo (guest, #67019) [Link]

How about an article not immediately behind a paywall? I'm not sure how that's supposed to be useful.

> If you mean non-profit corporations, that do not depend on donations, you might be right.

Well, the idea presented here was not one of charity. I don't see any point in diluting the discussion.

> do you consider IKEA (which is officially owned by a non-profit foundation) REALLY as a good example of a non-profit corporation?

To answer your question, absolutely not. This is a common structure in many large corporations. Some subsidiaries, or even the parent, may be nonprofits, but that's just to exploit tax policies. They do capture profit somewhere within the hierarchy. For IKEA, the for-profit arm is IKEA Systems. I don't have any problem with such arrangements -- they are perfectly legal -- but I would never lump them together with nonprofits who are unaffiliated with a for-profit.

Winning the copyleft fight

Posted Feb 17, 2016 2:55 UTC (Wed) by flussence (guest, #85566) [Link] (7 responses)

> > In fact, I'd expect a ton of operational inefficiency dramatically driving up costs.
> Why? Other than some preconceived ideology which says profit is good and charity is bad I don't see any reason why this would be the case, or demonstration that this is the case in the real world.

I seem to recall GNOME making headlines a few months back for runaway spending. I'll let you draw your own conclusions whether or not this constitutes “dramatic operational inefficiency”, here's the raw facts:
https://www.gnome.org/wp-content/uploads/2015/10/GAR2014-...

Winning the copyleft fight

Posted Feb 17, 2016 3:37 UTC (Wed) by mjg59 (subscriber, #23239) [Link]

> I seem to recall GNOME making headlines a few months back for runaway spending.

You recall incorrectly.

Winning the copyleft fight

Posted Feb 17, 2016 3:40 UTC (Wed) by raven667 (subscriber, #5198) [Link] (5 responses)

I'm not sure what you are referring to as "runaway spending" (do you mean a positive feedback loop of waste?), the only thing I saw, which was mentioned on LWN some time ago, is that corporate sponsored outreach programs were not being reimbursed in a timely fashion which caused a cash flow problem until the invoices were paid. Was there something else you were referring to?

>The large increase in OPW income was the result of sponsors
>form 2012 and 2013 rounds reimbursing the
>Foundation for the payments that had been made
>to the interns in previous years. As a result, the
>board took a harder stance on requiring invoices
>to be paid before providing the services which
>corresponded to those sponsorships.

Actually that financial document is a great example of how an organization can run on a shoestring budget, they only had around $20k of admin expenses a year and as low as $130k of payroll (which may be including benefits, unless they are in the admin category) for two FTEs, an executive director and an administrative person. That' would be cheap for a single executive, let alone two people. Also note the payroll went up to $220k in 2013 as they contracted a sysadmin, the implication being that for the other years they weren't paying cash for the IT services they were receiving.

Winning the copyleft fight

Posted Feb 17, 2016 16:18 UTC (Wed) by flussence (guest, #85566) [Link] (4 responses)

I'm referring to spending large amounts (60-70% of the budget for two years) on something which can not be solved through any amount of money.

Now ostensibly, it was a project for fixing an under-represented demographic in FOSS (and I like that ideal), but it's been hijacked as leverage in Someone Else's Political War, and consequently attracted a bunch of insufferable holier-than-thou socialites on both sides. (To keep this vaguely on-topic: who stands to gain the most from manufacturing such a divide in Linux?)

It would be much more cost-effective to bluntly fix the root problem of those skewed demographics — FOSS has far too many nasty people, some in the public eye often enough that they set the example for “acceptable” behaviour, and some who do it on purpose as job security. Less damage control, more cutting out the damage!

Winning the copyleft fight

Posted Feb 17, 2016 16:51 UTC (Wed) by raven667 (subscriber, #5198) [Link] (1 responses)

I think it's worth pointing out that according to their own financial disclosure statement, the income which was donated specifically to fund OPW, when recorded separately, was the large majority of outreach donations, so OPW was one of their most successful funding efforts. People voted with their dollars on what kinds of outreach they wanted to fund and OPW was the winner.

Winning the copyleft fight

Posted Feb 17, 2016 23:08 UTC (Wed) by flussence (guest, #85566) [Link]

When you put it that way, it seems more sensible. Point taken, though I still think it was a naïve idea to begin with...

Winning the copyleft fight

Posted Feb 17, 2016 19:15 UTC (Wed) by ksandstr (guest, #60862) [Link]

>(To keep this vaguely on-topic: who stands to gain the most from manufacturing such a divide in Linux?)

Microsoft and Apple do. But that's insignificant in comparison to facts-on-the-ground. Suppose there was a Halloween document demonstrating such things to be Microsoft's official strategy -- what changes? Blaming $ELDERS_OF_ZION is a crutch for those who can already walk on their own.

>It would be much more cost-effective to bluntly fix the root problem of those skewed demographics — FOSS has far too many nasty people, some in the public eye often enough that they set the example for “acceptable” behaviour, and some who do it on purpose as job security. Less damage control, more cutting out the damage!

This is just another form of "exclude these people whom I call $OUTGROUP_DU_JOUR". There's no need to do so, and it's a bad idea in any case.

As for holier-than-thous (an amusingly recursive reason for advocating this kind of kickban rules: who's holier than s/h/it who wields that authority?), they and other such cancer will generally fuck off for greener pastures as soon as the current one dries up. This is what establishes the upper bound for their numbers in any given project right now. Notably, this "ignore them" method is the most cost-effective of all by its effectiveness and complete absence of false positives, consequent to forking being always possible -- which in itself drives away those who'd use rules and/or playground bullying to commandeer a community.

Winning the copyleft fight

Posted Feb 18, 2016 11:11 UTC (Thu) by zack (subscriber, #7062) [Link]

> I'm referring to spending large amounts (60-70% of the budget for two years) on something which can not be solved through any amount of money.

in their own words: https://www.gnome.org/press/2016/02/gnome-foundation-was-...

Winning the copyleft fight

Posted Feb 16, 2016 21:53 UTC (Tue) by pizza (subscriber, #46) [Link]

> But, yes, in general, I agree with your sentiment. We need to look at new options to draw in corporations to open source development, and beating them over the head with copyright lawsuits and C&D letters will, instead, drive them away as fast as they can run.

It is quite naive to think that corporations will follow rules that have no teeth.

I have no sympathy at all for corporations that do not comply with the licenses of the software they use, especially when the GPL is actually far simpler than typical commercial source code licenses.

Live by the IP sword, die by the IP sword.

Winning the copyleft fight

Posted Feb 17, 2016 9:25 UTC (Wed) by HIGHGuY (subscriber, #62277) [Link] (2 responses)

> Nonprofit doesn't mean that overhead is the same or lower than for-profits. In fact, I'd expect a ton of operational inefficiency dramatically driving up costs. You'll need a whole cadre of business and support people to make this work. Who's going to manage them?

I'm not talking about overhead. Overhead is part of the actual cost. I'm referring to the non-profit part. There's no reservation required for huge margins to pay your shareholders to stop shouting at you for not serving their interests. It means margins go down and the actual price point of your goods/services to go down.

> You don't get to dump your code upstream and expect them to maintain it forever.

I agree. But for actually useful code this is probably as true as it gets.
Also, having a company around your project doing contract work, doesn't dismiss you of being a good upstream for any contributions, including the ones you were contracted for. It's therefore still the developer's work to ensure that whatever is written plays well with upstream and that, where necessary (for one-off code), the proper abstractions are pushed upstream to allow for maintainability of everyone's code, even in the absence of the code itself in the upstream project.
Ultimately, you're competing with some developer inside the company building and maintaining a fork. In many cases it will be possible to find some middle ground in building/supporting crazy feature X for a company while still maintaining clean code in the upstream project. If not, perhaps the consultancy should comprise of education of why their fancy new idea is silly...

Winning the copyleft fight

Posted Feb 17, 2016 16:24 UTC (Wed) by fandingo (guest, #67019) [Link] (1 responses)

> There's no reservation required for huge margins to pay your shareholders to stop shouting at you for not serving their interests.

I don't think you have any concept of how little corporations pay in dividends (that's the term for this "pay"). It's not a huge amount anywhere -- certainly not enough to give nonprofits more advantage than they already receive from tax authorities. Plus, dividends are a lot rarer these days.

Winning the copyleft fight

Posted Feb 18, 2016 6:41 UTC (Thu) by HIGHGuY (subscriber, #62277) [Link]

There's 2 ways of making profit on stock. One is dividend, as you noted, the other is stock price growth. The latter being the far more common one.
It's also the more difficult one, since it depends only indirectly on real world metrics.
Therefore companies put a lot of effort in shareholder PR, ever-growing profits/margins, lowering costs, growing market value, short-term thinking and much more to keep their stock price growing.
Companies that don't have to revert to these tactics can think about other, more longterm factors instead like treating their employees right/better, doing long term work because it'll keep the project in better shape,...
I'll now let you decide on which model is more sustainable, satisfying for employees, etc.

Winning the copyleft fight

Posted Feb 14, 2016 12:59 UTC (Sun) by excors (subscriber, #95769) [Link] (11 responses)

> He also called for developers to go back to volunteer coding

My perspective on volunteer coding changed a bit when I got a job as a software developer. Sometimes there were little bits of code I'd written or contributed to as a personal project in the past, that turned out to be quite useful for what I was doing in my job. But using GPLv2 code at work was a big hassle (because the companies had policies and processes to make sure they were fulfilling their licence obligations properly, and the lawyers and managers and developers cared about doing it right, but it was a special case rather than the default mode for new software); and GPLv3 was essentially forbidden (because of patent stuff).

It was usually easier to find some alternative non-copyleft code or to just rewrite it from scratch, but that felt like a waste of time when I'd already written the same code years earlier. So nowadays I prefer non-copyleft licences for my personal projects, because it makes my life easier when I'm at work, as well as anyone else's who's in the same situation.

(The opposite is also a concern, in that it's a waste of time if I've written some proprietary code at work that I'd like to (but can't) use in a personal project years later. But then I'd have to put a lot of effort into open-sourcing that code at work, for no immediate benefit and a highly uncertain future benefit, so laziness wins.)

I think that's an unfortunate situation, because I'd like there to be more open source code for everyone to benefit from, but I'm not sure what I can do about it that won't take a lot more effort.

Winning the copyleft fight

Posted Feb 14, 2016 13:49 UTC (Sun) by halla (subscriber, #14185) [Link] (1 responses)

If _you_ wrote the code and licensed it GPLv2, then you could also relicense it for use at your work.

I wouldn't do that, I think that companies should pay for development instead of batten on the free software work I've done before, but that's just me.

Winning the copyleft fight

Posted Feb 14, 2016 15:38 UTC (Sun) by pizza (subscriber, #46) [Link]

> If _you_ wrote the code and licensed it GPLv2, then you could also relicense it for use at your work.

I've done this before, but it's only an option if it's code that is yours outright -- ie no third-party bits.

Winning the copyleft fight

Posted Feb 14, 2016 16:08 UTC (Sun) by pizza (subscriber, #46) [Link] (7 responses)

> and GPLv3 was essentially forbidden (because of patent stuff).

Anyone who says they object to the GPLv3 due to patents, but is okay with the Apache license, is full of crap.

In both cases, there is an _explicit_ patent grant. Apache's goes even further by adding a retaliation clause terminating the license if you initiate a patent suit. (Unfortunately that still won't protect you against trolls..)

(As an aside, even the classic permissive BSD licenses arguably include an implicit patent grant -- "Redistribution and use in source an binary forms, with or without modification, are permitted...")

Winning the copyleft fight

Posted Feb 14, 2016 17:17 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link] (6 responses)

> In both cases, there is an _explicit_ patent grant. Apache's goes even further by adding a retaliation clause terminating the license if you initiate a patent suit. (Unfortunately that still won't protect you against trolls..)
They differ in scope. Apache 2 is OK for companies - by releasing Apache 2 code you agree not to sue for the functionality in this code only.

GPLv3 also places requirements on patent licenses from third parties.

Winning the copyleft fight

Posted Feb 25, 2016 17:36 UTC (Thu) by Del- (guest, #72641) [Link] (5 responses)

Can you please document this. Is it your own opinion, or is there any legal authority (e.g., statement from anybody with formal background in law) you have reference to? I am genuinely interested. I thought GPLv3 and Apache 2.0 had the same patent protection for all practical purposes.

Winning the copyleft fight

Posted Feb 25, 2016 19:19 UTC (Thu) by pizza (subscriber, #46) [Link] (4 responses)

From the GPLv3 text:

Each contributor grants you a non-exclusive, worldwide, royalty-free patent license under the contributor's essential patent claims, to make, use, sell, offer for sale, import and otherwise run, modify and propagate the contents of its contributor version.

And:

A “contributor” is a copyright holder who authorizes use under this License of the Program or a work on which the Program is based. The work thus licensed is called the contributor's “contributor version”.

Apache's text:

"...where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted."

Winning the copyleft fight

Posted Feb 25, 2016 19:33 UTC (Thu) by Del- (guest, #72641) [Link] (3 responses)

Thanks for the quotes, but I have read the text. I fail to see that there is any practical difference though. The wording in GPL seems better crafted and more explicit, but a natural interpretation of the Apache text yields the same practical implications.

Winning the copyleft fight

Posted Feb 26, 2016 1:58 UTC (Fri) by zlynx (guest, #2285) [Link] (2 responses)

As I understand it:

- In the Apache license, if someone else merged code infringing your patent and you didn't notice it before pulling their changes into your copy, you can still sue them.

- In the GPL3 version, once you release that code, you have granted a patent license on the entire thing, including your competitor's infringing code.

Kind of a big difference.

Winning the copyleft fight

Posted Feb 26, 2016 16:04 UTC (Fri) by Wol (subscriber, #4433) [Link]

The tragedy, of course, is that all such patents are invalid as per SCOTUS precendent and the EU Patent Treaty.

The problem is proving it :-(

Cheers,
Wol

Winning the copyleft fight

Posted Feb 27, 2016 10:17 UTC (Sat) by Del- (guest, #72641) [Link]

> As I understand it:

So this is your personal interpretation as an arm chair lawyer?

I tried to read through the text again, and I agree with your interpretation of the patent clause in GPLv3. However, Apache's text is horribly ill-crafted. It is not at all clear to me whether that clause puts the same restrictions. It really is poorly crafted, makes me wonder what kind of legal council the Apache guys got.

As it stands, I far prefer the patent clause in GPLv3 to the one in Apache, and I believe most customers/users, corporate or not, agrees with me.

Winning the copyleft fight

Posted Feb 25, 2016 10:08 UTC (Thu) by Del- (guest, #72641) [Link]

> I think that's an unfortunate situation

Yes it is, and that is exactly why we need copyleft. My hope is that your company looses the competition to a more open company. We have seen this happen all over the place the last two decades. Companies (even Apple until they got a dominating market position) embracing open source and GPL in particular, and winning the market. Which companies made it big the last twenty years? How did they build their software stack?

If you want to be an open source developer, you will need to find a new employer.

Winning the copyleft fight

Posted Feb 15, 2016 8:44 UTC (Mon) by jezuch (subscriber, #52988) [Link] (14 responses)

> He also called for developers to go back to volunteer coding

He also called for me to lessen the amount of time I spend with my loved ones :) Or to switch to half-time work, and half-pay, which would not sit well with those loved ones. (OK, I don't have that many of them, but I plan to... some day.)

Seriously, though, my plate is already full of stuff I want to do in my free time and after the whole day (in the evening) and the whole week (on the weekend) of working on software at work, working on software *more* after that gets pretty low on my priorities list. Call it a burn-out, or something, if you will. Maybe in more developed societies people have enough time for leisure and hobby work, but over here... not so much. I'd love to, but Bradley Kuhn at al. would do well to consider realities of non-first-world countries.

Winning the copyleft fight

Posted Feb 15, 2016 16:57 UTC (Mon) by cstanhop (subscriber, #4740) [Link] (5 responses)

This is also a problem for those of us in the "first-world". At least I have found it to be extremely difficult to find free time (and also be free of the claims my employer makes on any IP generated on my own time with my own resources). Of course, maybe I'm just not working hard enough. :-/

Winning the copyleft fight

Posted Feb 16, 2016 16:16 UTC (Tue) by mathstuf (subscriber, #69389) [Link] (4 responses)

It's ridiculous that employers think they can claim ownership of work done outside of their purview on the employee's own time. Now, if you work at $financial_firm and work on FOSS market analytics software in your free time, there *might* be an argument there, but for working on Python software when the company is a C# shop? I turned down an internship when the best answer they could give was "depends on your manager"…and my manager wouldn't be known until I started.

Winning the copyleft fight

Posted Feb 18, 2016 10:44 UTC (Thu) by rcochran (guest, #105426) [Link] (3 responses)

> It's ridiculous that employers think they can claim ownership of work done outside of their purview on the employee's own time.

If you want to work on free software in your own free time,
then it is important to have this specifically covered by your
employment contract. Otherwise, you'll always have to beg
for permission, and the answer will often be "no."

Winning the copyleft fight

Posted Mar 9, 2016 17:52 UTC (Wed) by nye (guest, #51576) [Link] (2 responses)

>If you want to work on free software in your own free time,
>then it is important to have this specifically covered by your
>employment contract. Otherwise, you'll always have to beg
>for permission

In what possible situation could this be true? What jurisdiction are you talking about?

What if I want to talk about bees in my free time, will I have to beg my employer for that? What if I want to cook dinner? Write a novel? Play the banjo? Breathe the air?

If you *specifically* sign a contract promising not to, then *maybe* a court might be willing to uphold that requirement even though it's clearly unreasonable, but absent such a contract your employer can't simply require you to get their permission for arbitrary out-of-work activities, at least not in anywhere resembling civilisation.

Winning the copyleft fight

Posted Mar 9, 2016 18:47 UTC (Wed) by viro (subscriber, #7872) [Link] (1 responses)

You are forgetting one nasty detail - contract claiming ownership of everything written in hours you are billing them for might easily find sympathy from judge. And proving that none of your work on $OTHER_PROJECT had been done during the working hours and thus does not fall under the conditions spelled in contract is likely to be tricky.

Winning the copyleft fight

Posted Mar 10, 2016 10:04 UTC (Thu) by anselm (subscriber, #2796) [Link]

I do stuff for work during working hours on a computer provided by my employer, and work on personal projects in my personal time on my own computer. The work computer does not contain any material related to my personal projects and vice-versa. My contract stipulates that the stuff I do for work belongs to my employer, but my employer would find it very difficult to claim ownership of any of the stuff on my private computer (not that they want to).

Things might get muddled in such a setup when you want to use for work some of the stuff you wrote on your own time, or vice-versa. A blanket “all your code are belong to us” clause doesn't really help here.

Winning the copyleft fight

Posted Feb 17, 2016 9:14 UTC (Wed) by fb (guest, #53265) [Link] (6 responses)

> He also called for developers to go back to volunteer coding — off hours, it is not necessary to quit one's job to do this (though he did note that some companies will claim ownership of even an employee's off-hours work). Early free software was written this way. Bradley said that he works nights and weekends for the cause; he would like for the development community to do the same. Developers should also be willing to fork permissively licensed software under copyleft when necessary.

To me this sounds as unrealistic as it gets. It also sounds as Bradley is so deep into it that he forgets how the 'discretionary free time' of most people goes down as they acquire a job and/or build a family.

The basic fact of life is that any high maintenance project will require funding of some sort to work out without burnout in the long term.

Yeah, some folks keep having loads of time for hobbies (FOSS or not) even while working. In my experience, this later group is nearly 100% comprised of single people or married-but-childless. All colleagues I have with serious commitment to hobbies are either unmarried or childless.

[....]

Once I was on IRC and someone showed up and mentioned that he dreamed of doing volunteer contributions to FOSS projects, but he could never find the time and that he was amazed at people like us (the devs of this one FOSS project) who did manage to do it... My answer was "Look, *all* the regular FOSS contributors here are paid by Red Hat to work full time on it". (PS: I no longer work for Red Hat.)

Winning the copyleft fight

Posted Feb 17, 2016 13:32 UTC (Wed) by pizza (subscriber, #46) [Link] (3 responses)

> To me this sounds as unrealistic as it gets. It also sounds as Bradley is so deep into it that he forgets how the 'discretionary free time' of most people goes down as they acquire a job and/or build a family.

While the latter part of your statement is indisputably correct, you are sort of missing Bradley's point -- If you care about Software Freedom, you need to actually contribute in order to make a difference -- and that's not something most people can do except on a volunteer basis.

Winning the copyleft fight

Posted Feb 19, 2016 8:58 UTC (Fri) by fb (guest, #53265) [Link] (1 responses)

> If you care about Software Freedom, you need to actually contribute in order to make a difference -- and that's not something most people can do except on a volunteer basis.

Fair. My point is that IMHO (mostly) volunteer contributions will //not// deliver state-of-art or at least excellent (complex) applications. IMHO it is just not going to happen.

I can see that happening for very simple applications or applications aiming at the niche of software engineers / academic research staff, but not really on other cases.

Winning the copyleft fight

Posted Feb 19, 2016 12:32 UTC (Fri) by pizza (subscriber, #46) [Link]

FWIW I think you're probably correct.

But that's not a problem with Free Software per se; instead it's an issue of scaling. Witness what happened with smartphone apps, for example.

Winning the copyleft fight

Posted Feb 19, 2016 12:14 UTC (Fri) by ira (guest, #93573) [Link]

I think you ignored a potential route to getting copyrights:

Getting them from your employer. If you work on open source, they may not care.. or may not even WANT them.

Clearly, different employers will feel differently, and your status in the community may also influence this. If you are a well established committer, getting copyrights may be much simpler, etc.

Volunteer time was only one suggestion.

Posted Feb 23, 2016 21:26 UTC (Tue) by bkuhn (subscriber, #58642) [Link] (1 responses)

Note that I didn't say volunteer time on copylefted projects wasn't the only solution; it was part of various solutions I proposed, such as insisting to keep one's own copyrights in your negotiations with your employers and also insisting that paid work be copylefted rather than non-copylefted.

Volunteer time was only one suggestion.

Posted Feb 23, 2016 21:27 UTC (Tue) by bkuhn (subscriber, #58642) [Link]

Er, sorry for double negative, volunteer time is not the only solution, I meant.

Winning the copyleft fight

Posted Feb 23, 2016 21:24 UTC (Tue) by bkuhn (subscriber, #58642) [Link]

> I'd love to, but Bradley Kuhn at al. would do well to consider realities of non-first-world countries.

I was admittedly speaking to an audience of affluent people from wealthy countries. Your points about the lack of leisure time in less wealthy countries is salient indeed.

However, you can still refuse to write proprietary software for a living, no matter where you live. There are other jobs.

The bar has been raised

Posted Feb 17, 2016 20:42 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (3 responses)

Also it should be noted that the complexity level of today's software is far higher than it had been during 90-s and even early 2000-s.

Back then most of software was quite primitive. Look at GIMP circa 1999: http://www.mathematik.uni-ulm.de/help/gimp/images/screens... Writing something like this in spare time was quite feasible.

These days it's not really possible to write, for example, a modern image editor in one's spare time alone. You either need to work on it full time or have an efficient collaboration.

The bar has been raised

Posted Feb 18, 2016 15:14 UTC (Thu) by gioele (subscriber, #61675) [Link] (2 responses)

> Also it should be noted that the complexity level of today's software is far higher than it had been during 90-s and even early 2000-s.

> Back then most of software was quite primitive. Look at GIMP circa 1999: http://www.mathematik.uni-ulm.de/help/gimp/images/screens... Writing something like this in spare time was quite feasible.

> These days it's not really possible to write, for example, a modern image editor in one's spare time alone. You either need to work on it full time or have an efficient collaboration.

Maybe somebody wrote this mail in the late 90's:

Also it should be noted that the complexity level of today's software is far higher than it had been during the '70s and even early '80s.

Back then most of software was quite primitive. Look at SuperPaint circa 1973: <http://de.slideshare.net/RandyShoup/devopsdays-silicon-va...>... Writing something like this in spare time was quite feasible.

These days it's not really possible to write, for example, a modern image editor to rival Photoshop in one's spare time alone. You either need to work on it full time or have an efficient collaboration.

This "GIMP" application that they announced just some days ago will always remain a toy.

The bar has been raised

Posted Feb 18, 2016 21:59 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

> Also it should be noted that the complexity level of today's software is far higher than it had been during the '70s and even early '80s.
That's also true. A single good developer could realistically write the whole software stack from ground up in early 80-s. By 90-s that moved up to the level of most applications.

Now it's even higher. Niche applications are still somewhat accessible to lone weekend coders, but that's probably it.

The bar has been raised

Posted Feb 19, 2016 11:13 UTC (Fri) by jezuch (subscriber, #52988) [Link]

> Maybe somebody wrote this mail in the late 90's:
> Also it should be noted that the complexity level of today's software is far higher than it had been during the '70s and even early '80s. (...)

That's true, but the difference is that the software development techniques were even more primitive back then so it evens out. There was no equivalent progress in this regard from the '90s to the present as was from the '70s to the '90s. (There was some but not as much; we're *still* using C, for example.)

Winning the copyleft fight

Posted Feb 19, 2016 16:27 UTC (Fri) by leoc (guest, #39773) [Link]

Android N switches to OpenJDK, Google tells Oracle it is protected by the GPL.

Winning the copyleft fight

Posted Feb 25, 2016 14:57 UTC (Thu) by shlevy (guest, #87221) [Link]

> Copyleft licensing is still an effective strategy, he said; that can be seen because we've had the chance to run a real-world parallel experiment — an opportunity that doesn't come often. A lot of non-copyleft software has been written over the years; if proprietary forks of that software don't exist, then it seems clear that there is no need for copyleft; we just have to look to see whether proprietary versions of non-copyleft software exist. But, he said, he has yet to find a non-trivial non-copyleft program that lacks proprietary forks; without copyleft, companies will indeed take free software and make it proprietary.

Seriously? This is a terrible measure of the effectiveness of non-copyleft licenses. This reasoning basically boils down to "non-copyleft licenses would only ever be acceptable if no one ever took advantage of their non-copyleft aspects". Some better questions: Do non-copyleft projects stagnate while their proprietary forks flourish? Are there many cases where useful work would likely have been done and pushed back if the project had been copyleft but wasn't because the project was non-copyleft? Or, if you acknowledge that developers might just have different values than you do: Do non-copyleft projects accomplish what their authors want them to?

Winning the copyleft fight

Posted Feb 25, 2016 20:17 UTC (Thu) by prometheanfire (subscriber, #65683) [Link]

RPC (rackspace private cloud) is also fully upstream/open.

https://github.com/rcbops/rpc-openstack/ sources from https://github.com/openstack/openstack-ansible/