User: Password:
|
|
Subscribe / Log in / New account

Toward healthy paranoia

Toward healthy paranoia

Posted Sep 12, 2013 22:09 UTC (Thu) by khim (subscriber, #9252)
In reply to: Toward healthy paranoia by lsl
Parent article: Toward healthy paranoia

Well, unless you build your device to specifically deny full access to the owner.

Or when you include fixes for hardware flaws in software. You may be surprised, but in embedded world that's not abnormal case, it's typical situation that you include some software safeguards against hardware faults. This means that you need make it hard to use unauthorized software and if user will circumvent this restriction somehow (most likely be removing some kind of seal) - well, it's his/her choice, your can righteously say that if use circumvented your protection then s/he's on his/her own and warranty is null and void.

GPLv3 breaks this scheme apart: you are forced to offer some way to replace some components of your system (often central components of your system) and even if you'll disclaim the warranty for such a case (GPLv3 explicitly allows that) you still are faced with support calls burden.

But this change itself was not the biggest problem with GPLv3: GPLv3 requirements are not all that onerous per se. Biggest problem with GPLv3 was “I am altering the bargain, pray I don't alter it any further” perception. Which basically poisoned GPL well.


(Log in to post comments)

Toward healthy paranoia

Posted Sep 12, 2013 23:36 UTC (Thu) by pizza (subscriber, #46) [Link]

> But this change itself was not the biggest problem with GPLv3: GPLv3 requirements are not all that onerous per se. Biggest problem with GPLv3 was “I am altering the bargain, pray I don't alter it any further” perception. Which basically poisoned GPL well.

One other comment that I think bears mentioning.

All previously-distributed GPLv2 software is still GPLv2; it doesn't magically stop working or relicense itself.

So at the end of the day, the response to folks who take the "oh noes, the people who write software we get to use and modify for free may, at some point in the future, change the terms of new software they write after that date!" attitude come off as if they are complaining that their free pony doesn't come in the color they want.

Nevermind that the alternative licensing schemes (eg BSD, Apache) could just as easily end up in the same situation, with future releases licensed under different terms. And historically, that's precisely what tends to happen as the tragedy of the commons unfolds yet again.

Honestly, I just don't get it. :)

Toward healthy paranoia

Posted Sep 13, 2013 7:17 UTC (Fri) by khim (subscriber, #9252) [Link]

Nevermind that the alternative licensing schemes (eg BSD, Apache) could just as easily end up in the same situation, with future releases licensed under different terms.

Sure—but that's exactly the point.

Honestly, I just don't get it. :)

I'll try to explain. Just what is (or may be I should say “was”?) the advantage of the copyleft scheme? To the vendor, I mean, not to the end user?

The answer is simple: the sole advantage is unchangeable tit-for-tat bargain. It's as simple as that. From vendor's POV Apache or BSD license is clearly superior: you can use it without worrying about compliance (even if you forget to do anything remedy is to add couple of sentences to the manual), it's easy and simple, no added costs. But, as you've noted, Apache or BSD creation can always go extinct: it's creator is free to add additional restrictions at any time which, indeed, leaves you stranded.

Now, historically GPL offered an alternative—sure, you are saddled with some additional hassles if you use copylefted software, but in exchange you get the bargain which can not be altered: no matter what happens next version will also carry the same terms and can be used with the same rules as previous one. With GPL vendor is bound by some strict rules yet everyone else must follow them as well! That's the only selling point of a GPL, really (vendors don't care about user's freedom thus that part flies right over their heads). One can not do the same with GPL because of you may not impose any further restrictions on the recipients' exercise of the rights granted herein requirement. Well… one cannot alter the bargain if they are not FSF as it turned out. FSF can do that (and everyone knew they could do that), but as long as that ability was theoretical vendors were ready to ignore it (after all there are bazillion other things which can go wrong for them, what's one purely hypothetical scare compared to all that?).

This was powerful offer and vendors were ready to tolerate the problems the GPL creates in exchange for that powerful promise.

Except now this offer is null and void: FSF altered the bargain and, even worse, explicitly promised to alter this bargain further (how else can you interpret Change is unlikely to cease once GPLv3 is released. If new threats to users' freedom develop, we will have to develop GPL version 4.?). This effectively killed one and only advantage of the copyleft: it's no longer possible to guarantee that new versions will carry the same terms. In fact it's now known that terms will be changed if new threats to users' freedom develop. Just what is the advantage of GPL over BSD if it's terms can be changed at any point anyway?

So at the end of the day, the response to folks who take the "oh noes, the people who write software we get to use and modify for free may, at some point in the future, change the terms of new software they write after that date!" attitude come off as if they are complaining that their free pony doesn't come in the color they want.

Not “in the color they want” but “in the same beige color it used yesterday”—and they have chosen this particular stud-farm in the first place because farmer promised them that all the followup batches of free ponies will come in the exact same beige color! They could have chosen black ponies or white one or even rainbow ones, but they have chosen dull beige because it was not clear if black ponies or rainbow ponies will be available tomorrow or not. But now instead of beige ponies you are getting brown ponies! WTF???!!! What happened to the bargain which attracted them to beige ponies in the first place?

Note how Linux (which is also developed under GPL terms, remember?) is not affected by this strong copyleft backlash. Why? Because it didn't alter the bargain! When FSF unleashed it's “weapon of mass oponion” Linus explicitly refused to participate. I think this is what saved Linux. MINIX is in state similar to where LLVM was back when GPLv3 happened, but companies don't see the need to spend their money on it because Linux (with it's unchanged bargain) is “good enough”.

Toward healthy paranoia

Posted Sep 13, 2013 10:38 UTC (Fri) by hummassa (subscriber, #307) [Link]

Your whole argument seems based on the following premise:

the FSF/GNU can alter the GPL at their whim.

Which is, first, a false premise (they do have bylaws and promises and goals and they can't just *change* the licenses in a new version, they have to *justifiably* change them), and second, and more important, a moot premise (people can continue to develop GPLv2, LGPLv2, LGPLv3 software without any outside change to their licensing, for that you can only drop the "vX or later" on your license).

Toward healthy paranoia

Posted Sep 13, 2013 15:08 UTC (Fri) by wookey (subscriber, #5501) [Link]

Nobody makes people use GPLvX _or later_, which is required for your argument to make any sense. If you are happy with GPLv2 then use that, if you are OK with GPLv3 then use that. If you are happy with 'whatever the FSF says this decade' then choose 'or later'.

You are quite right that big corps don't like GPLv3, but in my experience that's almost entirely due to patent-paranoia, rather than anything to do with potential changes in new licences. Because GPLv3 is a lot clearer about the fact that they really do have to agree not to assert patent claims on 'anything derived from this' than v2 was. So this problem is actually a side-effect of the whole software-patent disaster rather than much do with changes trying to retain practical user freedoms.

The growing realisation that you can't trust _any_ code you (or someone) can't examine, rebuild and replace, especially if it came from the US, could make a real difference, and shift things back towards copyleft. We shall see. I'd like to think so, but then I already thought copyleft mattered.

Toward healthy paranoia

Posted Sep 13, 2013 15:46 UTC (Fri) by khim (subscriber, #9252) [Link]

Nobody makes people use GPLvX _or later_, which is required for your argument to make any sense.

Huh? The GPLv2 actually only discusses two possibilities:
1. If the Program specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation.
2. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation.

The idea that you can just omit “or any later version” and thus make sure new ideas of GPLv3 will not affect you was Linus' inventions, not FSF's one—and others were not so enlightened.

Anti-GPLv3 started at the end of distribution tree: vendors revolted and only after that happened developers decided that they will forget about it since GPL is no longer acceptable. But by now a lot of developers see the writing on the wall: if you want to create code which will be used by real people to solve real problems - you should not pick the GPLv3. Some old projects get away with relicensing because they are important enough (think Samba), but how many popular GPLv3-licensed projects can you name which gained popularity as GPLv3 project and not as project which first gained popularity as GPLv2 project and then changed the bargain.

Of course first case of "altered bargain" was AGPL, not GPLv3 but there situation was different: some projects adopted it, most ignored it, but there was no bargain alteration, few projects (if any) went from GPLv2 to AGPLv1.

If you are happy with GPLv2 then use that, if you are OK with GPLv3 then use that.

What if you relied on “no further restrictions” clause and hoped to receive similar terms for newer versions of software which is now only available under GPLv3 license (this is what happened with GCC and Samba)? Should you wait till GLibC will be relicensed under GPLv3 and then start to suddenly look for a replacement or should you just abandon it and go with something like Bionic?

You are quite right that big corps don't like GPLv3, but in my experience that's almost entirely due to patent-paranoia, rather than anything to do with potential changes in new licences.

Well, may be, but it's the same case in the end: bargain was altered and both people who don't like new bargain and people who don't want to deal with next unexpected change in the bargain are revolting.

Toward healthy paranoia

Posted Sep 13, 2013 18:36 UTC (Fri) by hummassa (subscriber, #307) [Link]

> Huh? The GPLv2 actually only discusses two possibilities:

Actually, the section 9 of the GPLv2 is discussing two alternatives to the (obvious, IMNSHO) option where you are accepting the terms of the license you are reading in the moment you make copies of a GPLd work. Section 14 of the GPLv3 discusses the same two alternatives.

The forementioned sections are in the line of:

* these are the terms of this license; but if someone licenses the work as "GPLvX or later", it means that you have the option (not the obligation) of taking it under the term of later GPL's; and if someone licenses some work as "GPL", it means that you have the option (again, not any obligation) of taking it under the term of any version of the GPL that you care for.

Toward healthy paranoia

Posted Sep 19, 2013 17:26 UTC (Thu) by Wol (guest, #4433) [Link]

The idea that you can just omit “or any later version” and thus make sure new ideas of GPLv3 will not affect you was Linus' inventions, not FSF's one—and others were not so enlightened.

Actually, the GPL (v2 or otherwise) ITSELF discusses NONE of this. There is a load of blurb - which is not part of the GPL itself - which discusses this.

There is a LOT of rubbish in the comments on this article, and it's pretty much all down to the fact that far too many people don't actually understand how copyright works.

Cheers,
Wol

Toward healthy paranoia

Posted Sep 19, 2013 20:22 UTC (Thu) by khim (subscriber, #9252) [Link]

Actually, the GPL (v2 or otherwise) ITSELF discusses NONE of this.

Sorry, but it does:

9. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns.

Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation.

Discussion about “later versions” is very much part of GPLv2—but Linus was apparently first who studied it enough to understand what it may mean in the future, found that GPLv2 is not like MPL (which explicitly says that You may also choose to use such Covered Code under the terms of any subsequent version of the License published by Netscape) or CC-BY-SA (which says You may distribute, publicly display, publicly perform, or publicly digitally perform a Derivative Work only under the terms of this License, a later version of this License with the same License Elements as this License, or a Creative Commons iCommons license that contains the same License Elements as this License (e.g. Attribution-ShareAlike 2.0 Japan)). and consciously decided to reject this possibility.

P.S. It'll be interesting to hear outcry from Creative Commons users when these licenses will alter it's bargain enough—this should be even bigger fiasco then GPLv2-to-GPLv3 transition because with GPLv2 you at least have Linus's option of rejecting said transition. No such luck with Creative Commons.

There is a LOT of rubbish in the comments on this article, and it's pretty much all down to the fact that far too many people don't actually understand how copyright works.

Well, sure. The fact that people actually distribute anything under Creative Commons and never even discuss the fact that by doing so they give complete power over their creations to some random guys over there but discuss GPLv2-to-GPLv3 transition to the death is large enough clue. Remember how FSF unilaterally shifted the whole Wikipedia without asking it's authors about anything to CC-BY-SA 3.0? Well, folks behind Creative Commons can easily do similar trick to the whole corpus of the Creative Commons-licensed art.

Toward healthy paranoia

Posted Sep 19, 2013 23:44 UTC (Thu) by lsl (subscriber, #86508) [Link]

> Remember how FSF unilaterally shifted the whole Wikipedia without asking it's authors about anything to CC-BY-SA 3.0?

The FSF? Unilaterally? Is it even possibly to make a more biased account of that story? As stated in your link 17.000 Wikimedia/Wikipedia authors voted on it with 75 % being in favour of the license change. The FSF was just an accomplice in that plot originating in the Wikimedia communities as it had the power to allow licensing changes through the 'or later' clause.

Toward healthy paranoia

Posted Sep 20, 2013 1:29 UTC (Fri) by khim (subscriber, #9252) [Link]

As stated in your link 17.000 Wikimedia/Wikipedia authors voted on it with 75 % being in favour of the license change.

Right. Small percentage of Wikipedia copyright holders held a vote and even among them about 10% (that's over thousand copyright-holders, remember?) voted against said change. Without FSF's power to unilaterally change the license they faced lengthy (probably multi-year) process with uncertain outcome. They convinced FSF to apply it's power and bam: opinion of over thousand people about how their work can be used went to the wolves.

The FSF was just an accomplice in that plot originating in the Wikimedia communities as it had the power to allow licensing changes through the 'or later' clause.

Indeed. Note that votes, opinions and all other stuff was only needed to convince FSF, FSF had no need for any votes to change the license. It could have made it even if votes showed that two guys are for the change and 1698 are against.

P.S. Note that I'm not saying that this change was bad. I'm saying that it was made against of express wishes of sizable chunk of copyright holders. Recall how long it took for the Dell to overcome Icahn's opposition - and this is in case where deeds are supposed to be done by vote while Icahn only had 6% of voting power. With copyright you are not supposed to vote. Indeed, when nine lines were included in Android against Sun (and Oracle) wishes they raised racked to the sky, won the argument and only failed to convince court that incorrectly appropriated nine lines are worth billions. Yet FSF or CC can do such changes relicensing twice per day (if that'll be their wish) and nobody can say anything at all.

Toward healthy paranoia

Posted Sep 20, 2013 15:15 UTC (Fri) by nybble41 (subscriber, #55106) [Link]

> They convinced FSF to apply it's power and bam: opinion of over thousand people about how their work can be used went to the wolves. ... I'm saying that it was made against of express wishes of sizable chunk of copyright holders. ... With copyright you are not supposed to vote. ... Yet FSF or CC can do such changes relicensing twice per day ... and nobody can say anything at all.

You can't claim that the contributors didn't have a say in the license. They agreed to any changes the FSF might make when they agreed to the "or later" clause in the original license. This is a bit like complaining that the code you published under the GPL is being used to guide nuclear missiles; it was licensed for use by anyone, for any purpose, and you can't take that back later. In the same way, if you choose to license your work with an "or later" clause you give up control over how the work may be licensed in the future to whichever organization publishes the license. If that isn't what you want, don't license the work under an "or later" clause in the first place.

Of course, there is no requirement for the project to include contributions without such a clause, and doing so may seriously complicate project management down the road. If no compromise can be reached they may simply have to do without your contribution.

Toward healthy paranoia

Posted Sep 20, 2013 16:31 UTC (Fri) by khim (subscriber, #9252) [Link]

You can't claim that the contributors didn't have a say in the license.

They didn't.

They agreed to any changes the FSF might make when they agreed to the "or later" clause in the original license.

Well, sure. They accepted that the Free Software Foundation may publish new, revised versions of the GNU Free Documentation License from time to time. And they were promised that such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns.

In the same way, if you choose to license your work with an "or later" clause you give up control over how the work may be licensed in the future to whichever organization publishes the license.

Sure. But this is not what transpired here. Instead of receiving new similar in spirit version of license they were transferred en-masse from one Suzerain to another. That's fundamental violation of principle The vassal of my vassal is not my vassal. This may or may not be legal, but I know that it's not something I would like.

If that isn't what you want, don't license the work under an "or later" clause in the first place.

That's what I try to do lately, yes. I was significantly more forgiving to these clauses in the past, but after GPLv3 and GFDLv1.3 abuses of power by FSF it becomes more and more clear to me that the ability “to bugfix the license” is not good enough reason to give this much power to third party. Instead it's more honest to use BSD (or maybe Apache) license and give equal powers to everyone. Situation is not all that dissimilar to problem of Canonical's copyright assignment.

Toward healthy paranoia

Posted Sep 20, 2013 18:00 UTC (Fri) by nybble41 (subscriber, #55106) [Link]

> Well, sure. They accepted that the Free Software Foundation may publish new, revised versions of the GNU Free Documentation License from time to time. And they were promised that such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns.

Good, we're in agreement then. The FSF obviously felt the proposed Creative Commons license was "similar in spirit" to the GFDL, while resolving a number of issues which arose because what it was being applied to was not, in fact, documentation. They weren't breaking any promises, just using their position as the authoritative publisher of new versions of the GFDL to resolve real problems and concerns relating the GFDL in the context of WikiMedia.

> That's fundamental violation of principle The vassal of my vassal is not my vassal.

There are no "vassels" here, only free individuals choosing licenses for their contributions without thinking through all the possible consequences.

Toward healthy paranoia

Posted Sep 20, 2013 21:05 UTC (Fri) by khim (subscriber, #9252) [Link]

The FSF obviously felt the proposed Creative Commons license was "similar in spirit" to the GFDL, while resolving a number of issues which arose because what it was being applied to was not, in fact, documentation.

Few questions:
1. How could FSF know if CC-BY-SA 10.0 will have any resemblance to GFDL at all? They allowed relicensing from GFDL 1.2 to CC-BY-SA 10.0, after all.
2. Their promise quite explicitly said any later version published by the Free Software Foundation - is it fair to abuse this permission to switch to some other license not published by Free Software Foundation?
2. If CC-BY-SA is “similar in spirit” and actually resembles GFDL then why FSF says (quite explicitly) that we do not want to grant people this permission for any and all works released under the FDL?

They weren't breaking any promises, just using their position as the authoritative publisher of new versions of the GFDL to resolve real problems and concerns relating the GFDL in the context of WikiMedia.

No. They were abusing their position as the authoritative publisher of new versions of the GFDL to loan certain GFDL-licensed works to other fiefdom. They most definitely don't feel that CC-BY-SA is “similar in spirit” enough to give free pass to all GFDL users, they only exchanged parts of their congregation, not all of them.

There are no "vassels" here, only free individuals choosing licenses for their contributions without thinking through all the possible consequences.

It's idle talk. We can agree that individuals have chosen licenses without thinking too much about consequences, but there are also the fact that FSF treated these “free individuals” as serfs who have no power over their own creations because they once signed them away by choosing to license their work under “GNU Free Documentation License, Version 1.[012] or any later version”. Their wishes were irrelevant, their intents were ignored, new license was created to fulfill the Wikimedia Foundation's request, not to satisfy unanimous resolution of Wikipedia authors. Indeed with unanimous resolution they could have switched to any license of their choosing without FSF's involvement.

Note that CC-BY-SA is even more devious then GFDL: it embeds ability to use “a later version of this License” in the text of license itself. Even if you distribute something under CC-BY-SA 2.5 or CC-BY-SA 3.0 one may use text of CC-BY-SA 10.0 (which can include anything Creative Commons Corporation will want to include in it) and you can not disagree with that hijacking by omitting “or later” text from the license grant.

Toward healthy paranoia

Posted Sep 20, 2013 16:42 UTC (Fri) by khim (subscriber, #9252) [Link]

Of course, there is no requirement for the project to include contributions without such a clause, and doing so may seriously complicate project management down the road. If no compromise can be reached they may simply have to do without your contribution.

If I decide to participate in some project then I should accept the license they are using. “Or later” clauses, CLAs and all that. Doing anything else is sheer insanity. If my contribution is large enough I may decide to create separate project (which may or may not be pulled as third-party component into other one), if it's not large enough then I should accept the offer I was given, but I should not submit pieces under different license. Heck, even FSF explicitly says We recommend you use this license for any Perl 4 or Perl 5 package you write, to promote coherence and uniformity in Perl programming even when they say We urge you to avoid using it about it's first half (Artistic 1.0 license).

Toward healthy paranoia

Posted Sep 20, 2013 18:06 UTC (Fri) by nybble41 (subscriber, #55106) [Link]

> If I decide to participate in some project then I should accept the license they are using. “Or later” clauses, CLAs and all that. Doing anything else is sheer insanity.

Obviously. The project can't accept contributions without "or later" from just one contributor; it's all or nothing. The question was whether or not to participate in the first place given a project which has adopted an "or later" clause. If enough people choose not to participate in projects with such a clause, the projects will most likely drop it in order to regain contributors.


Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds