Debian and the PHP license
lwn.net
lwn.net
A bad HN automobile analogy would be if Ford had a clickwrap license on all new cars that summarized to "This Focus model car is provided by Ford, Ford wrote this Focus model car UI, Ford distributes this specific Ford Focus model, and you can officially do the FOSS thing with it, including Ford will not provide any guarantees". (note I'm guessing this isn't the real license on the ford focus UI, LOL, just a made up example)
And then GM sells a car, lets say, the Buick LaCrosse, and for reasons of apparent insanity slaps the Ford license on their car.
Next people are all WTF because GM released something under a license claiming that Ford owns it, wrote it, distributes it, and Ford won't provide guarantees (implying, I guess, that GM will).
Can I legally resell a GM car that claims its owned by Ford? So who owns this "GM" car, and if the license slapped on it is literally nonsense, is it even legally licensed software? So you're trying to sell a LaCross under the Focus license, are you claiming this thing is actually a Focus or that GM is a division of Ford, or Ford won't provide warranty coverage so I guess its implied GM will. If the license claims this is owned by Ford, but I bought it from a GM dealer, who's permission do I need to ask to relicense if needed? If I intentionally violate the license, can GM sue me if in writing they claim Ford owns this software, not GM? Or if Ford wanted, just to be jerks, could they sue me for using this GM software, because after all the license says Ford owns it. If the software crashes and kills people is a license falsely claiming Ford owns it a legal liability for GM for defamation (Ford owns it, its their fault not ours)?
- Ford makes car, licenses under Ford license.
- 3rd parties make add on products for the Ford automobile, like an updated radio, and say that it is licensed under the same license as the original car.
However, that would violate the sanctity of the "automobile analogy", so for better or worse I stay within the automobile mfgr community. A sacrifice of technical excellence for style.
http://lwn.net/1998/0611/ircii.html
I cold-called one of those guys and he was kind of amused by the whole thing. It was not much fun though.
Actually true. You will find competent developers who will change the license across their web app codebase from MIT to AGPL on a whim, saying AGPL is just the same as GPL.
I find legal stuff seriously interesting but also seriously scary.
The way some people presented it as a pitched battle between the Firefox and Debian teams is silly. Both were protecting their standing in some way, both appreciated the other's position, and a solution was found that causes no problems for either of them (aside from a small amount of user training along the lines of "yeah, it is Firefox under a different name").
It would be nice if Mozilla would follow suit, and come up with an official name for the open source project and community builds which is distinct from their official branded builds. Right now Debian uses "IceWeasel", GNU uses "IceCat", and other distros would have to come up with their own name if they made changes that were incompatible with Mozilla's trademark policy.
On the other hand, in pretty much every repository, you find "Chromium", but if you instead download straight from Google you get "Chrome".
Iceweasel is literally just Firefox packaged for Debian, except with a different name and icon. I don't think there is any functional difference (beyond the normal distro-specific packaging that every distro does).
Chromium lacks a number of features that Chrome has, because Chrome comes bundled with a bunch of proprietary components that will never be distributed with Chromium.
A basic example is that Chromium cannot (without a patch) display PDFs in-browser, the way Chrome can. Chromium also cannot support Chromecast, or any of the many other Chrome-specific features, of which there are many[0]
[0] https://i.imgur.com/AIxYzl9.jpg (screenshot is from Firefox, but it illustrates how much Google focuses on Chrome-specific functionality.)
Iceweasel is literally just Firefox packaged for Debian,
except with a different name and icon.
It looks like there has been some effort to do what I've described, come up with a generic default name for Firefox builds that can be used as a default, unbranded version[2] rather than requiring every distro that doesn't want to or can't comply with Mozilla's trademark policy have to do it on their own, but the effort appears to have stalled.
Chromium lacks a number of features that Chrome has,
because Chrome comes bundled with a bunch of proprietary
components that will never be distributed with Chromium.
[1]: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=354622 [2]: https://bugzilla.mozilla.org/show_bug.cgi?id=682414
Mozilla lets you use Firefox if you don't change the source; I believe Fedora and other distros release it with the original name. The difference is that, unlike Debian, they don't apply patches to the upstream source in their builds.
And I find it hard to believe that Google allows you to use the Chromium mark if you release a patched build; they're probably no different than Mozilla in that regard.
That said, I couldn't find a real trademark policy page for Chromium. If it's the generic Google one: http://www.google.com/permissions/trademark/our-trademarks.h... then Debian is clearly in violation of it.
http://anonscm.debian.org/cgit/pkg-chromium/pkg-chromium.git...
Google also documents that Chromium may have distro-specific patches applied and how it differs from the Google Chrome binaries provided by Google:
https://code.google.com/p/chromium/wiki/LinuxChromiumPackage...
https://code.google.com/p/chromium/wiki/ChromiumBrowserVsGoo...
They already did, it's called "Firefox". The ones that chose another name did so because they're making a different browser.
Google needed a different name because Chrome isn't open source at all.
im on mobile now ...
I've seen, and argued against, people using the WTFPL before: http://www.wtfpl.net/about/
Some, even in projects that could be considered "serious" and reused in another project (I won't touch a WTFPL licenses project because to me, it's not a license and is way to legally ambiguous).
I've seen way to many developers think they can write up some text and call it a license and it's "good". Not the case.
Pick a real license that was written by lawyers and has well understood (and preferably previously challenged) legal ramifications.
That said, this would probably be better served with an explicit public-domain declaration or equivalent (like the CC0 license).
There's not much preventing me from re-assinging copyright to myself (in countries that allow it) and then suing the original creator.
There's nothing to stop me from taking your code, re-licensing it under say a proprietary license, then suing the original creator.
There's nothing that stops me from taking your project's branding/symbol/graphics/name and claiming it as my own and then suing you.
There's nothing that stops me from taking your work and then not providing attribution back to the original creator.
By default there is nothing that absolves the license holder from warranty nor implied fitness for use. (If it destroys my computer running your code, I can sue you, and worse, you accept liability by default). That is, unless you include the optional snippet of text that absolves any warranty issues... but since it's optional, and not on the main license page (it's in the FAQ page), many authors forget to include this.
There are other things too... like, legally, what does "DO WHAT THE FUCK YOU WANT" actually mean? The law is black and white (or at least we try to be), being in the grey opens the door for abuse. Has it been challenged before in a legal sense? Has a lawyer reviewed it?
While I certainly understand the intent and spirit of this license; this license appears to be nothing more than a bad (and somewhat distasteful) joke that serves more harm than good.
I agree, better off choosing a public domain license or one of the several vetted FOSS licensees.
If you can't back anything up, I don't think you should be spreading your own interpretations in this manner, presenting them as truth. We already have far too many people confused over licenses, and passing that confusion on doesn't help anyone.
But I don't see where you substantiate any of your claims.
You seem to just want to defend an obviously joke license.
This license, from my interpretations, forfeits most of the copyright protections, including allowing someone to re-license (per the FAQ of the license).
At the end of the day, you are free to use whatever license you think suits you best.
That does not mean it actually suits you best. (As demonstrated by Debian putting their foot down on other BS licenses such as this).
You keep saying copyright licenses are used to "forfeit" (by which I assume you mean surrender) rights, so as to (exclusively) transfer them from you to someone else. It doesn't work like that. They are used to grant other people rights. They normally do not remove any rights from the author.
And you still maintain your circular definition of copyright, in which authorhood or copyright itself is a right protected by copyright and that through one of the rights protected by copyright (and granted to you by an author's license), you can obtain that right which is copyright (or authorhood). Can you please back any of it up?
Like I said before -- use a more appropriate license that still provides the goals you intend.
As a side note -- don't accuse me of not backing up my claims when neither do you... at this point, we're just two guys bickering on a forum.
But let's not go about making unfounded claims as to what you are allowed to do with it.
> don't accuse me of not backing up my claims when neither do you... at this point, we're just two guys bickering on a forum
Since a lot of your claims assume an exclusive transfer of rights (the method of which is not stated in the berne convention), I have already linked to and cited a US Copyright circular to back up my claim that transfer doesn't happen the way you assume it does. I can also cite the same circular and the Berne Convention to find what rights of authors are protected. Your circular understanding of copyrigts and the rights it reserves for the author assumes rights which appear in neither of the two documents (but the text is online, so prove me wrong). I've also pointed out the illogicalness of your definition.
Even so, the onus is on you to back up the claims you made. You started the discussion by making these claims.
But let me cite one more passage. You probably can't tell where it is from since you haven't read the text. Go do so already.
Independently of the author's economic rights, and even after the transfer of the said rights, the author shall have the right to claim authorship of the work [..]
So unless you're making an argument about Somalia (in which case you can forget copyright altogether), please drop the bullshit about claiming others' things as your own and then suing the real author. I mean sure, you can try... just as you can try to take the car of someone who released his software under WTFPL. "Because the license said I can!" won't hold up.
You can however license someone to use something that is protected by copyright. This is what licenses do; they give you a license to use the content.
This is exactly what CC0 tries to do, it works around the fact that you cannot (easily) give away or stop copyright, but can state that you distance yourself from all the rights that come with having the copyright. See also: http://creativecommons.org/about/cc0
In what jurisdictions would that be interpreted as copyright reassignment? I can see how you might argue that "re-assign copyright to myself" is WTF you want to do, but the onus is on the creator to assign copyright to you.
> There's nothing to stop me from taking your code, re-licensing it under say a proprietary license, then suing the original creator.
Suing them for what? There's nothing to prevent you from suing pretty much anyone, for any reason, including any of the other licenses. Second of all, if you're saying there's nothing preventing your suit from being successful, I still don't understand on what grounds. It doesn't matter what licence you've relicensed under if the original author holds copyright.
> There's nothing that stops me from taking your project's branding/symbol/graphics/name and claiming it as my own and then suing you.
You could do this no matter what the license. What about the WTFPL would make your lawsuit any more successful than any other licence?
> There's nothing that stops me from taking your work and then not providing attribution back to the original creator.
This is fully within the intent of the WTFPL.
> By default there is nothing that absolves the license holder from warranty nor implied fitness for use.
You have a good point that this is something many people would want. Do you know in which jurisdictions this is actually needed? I wasn't able to find any sources for this.
> Has it been challenged before in a legal sense? Has a lawyer reviewed it?
Excellent points.
First, the license grant is clearly ambiguous. Most license grants are written in terms of exclusive rights of copyright (ie reproduction, derivative work creation, public performance). This one isn't. "Do whatever the fuck you want' is unclear as to exactly what rights you can exercise and can't, because "whatever you want" can be found to not literally be "whatever" (and in fact, when most people say "you can do whatever you want", they often mean "whatever you want within reason")
Maybe a judge will agree with "it means all of them", maybe they won't. It depends on the situation. I wouldn't want to bet my life on it. Damages would likely be minimal in such a case, except in places with high statutory damages awards. You might get enjoined from using it.
Second, As for reassigning copyright, probably not. In most jurisdictions, this requires a specific signed instrument.
Third, you can't put it under a new license and sue the creator, they also have rights to it.
Fourth, "You have a good point that this is something many people would want. Do you know in which jurisdictions this is actually needed? I wasn't able to find any sources for this."
At the very least, in the US it's a very bad idea not to disclaim warranties. Particularly since people are making warranties on mailing lists as to fitness, etc, all the time.
I can't stress enough how dangerous it is to not disclaim warranties in the US.
> Has it been challenged before in a legal sense? Has a lawyer reviewed it? Excellent points."
OSI declined to name it an open source license, considering it mostly duplicative of others.
I specifically try to push people to avoid it, and the company i work for won't distribute software that uses it.
Can you offer some specific examples where an open source software project has suffered harm as a result? I generally trust your judgement, but would like to see the details.
I also appreciate the lack of IAALBTDNCLA in your post, although it might seem at odds with the disclaimer of warranties you advise. Is this omission a considered decision?
Not yet, but only because WTFPL is the only license i'm aware of that does not disclaim warranties, and it isn't all that popular.
I can offer plenty of examples of regular old sales without them that have gone bad. I can also offer examples of cases where judgement went for the defendants because the warranties were disclaimed.
As an aside, the WTFPL has a suggested warranty disclaimer in the FAQ, which is probably not effective in some places, which is even worse :(
In that sort of case, the only question is one of "damages", not of "validity of claim"
How? Where is the mutual consideration? WTFPL looks like a gratuitous license, rather than a contract.
> If the question is "can there be implied warranties on stuff given away for free", the answer is definitely "yes"
The usual implied warranties of concern are implied warranty of merchantability and implied warranty of fitness for a particular purpose; the former only applies when a good is sold, and even then only when it is sold by a merchant in the class of goods to which it belongs. And the latter applies again only with a sale, and only when the buyer relies on the seller's selection of the good for a purpose identified by the buyer, or at least known by the seller before selling the product to the buyer.
So if i understand your argument, because you are required to do nothing under the license, it is not a contract for lack of consideration?
If so, this is an interesting argument.
Let me turn it around then: Why, in your view, do most public domain dedications then include disclaimers of warranty (including CC-0), if they are pointless?
As for the second, A. A vast number open source projects are sponsored or sold by merchants in the class of goods to which it belongs (consultants, companies people selling things on app stores, etc) B. You can sell something for zero dollars. C. For fitness, I specifically mentioned representations often made on mailing lists (IE as to whether the software would work for a given person)
The gratuitous license argument is interesting (though from what I can tell, legal minds greater than mind still feel the urge to disclaim warranties in that case), but the contract/sale one is just completely uninteresting to me, given the lengths courts have gone in the past.
Yeah, that's like the first week of Contracts.
> Let me turn it around then: Why, in your view, do most public domain dedications then include disclaimers of warranty (including CC-0), if they are pointless?
Standard-form gratuitous licenses (incl. public domain declarations) often are written under the assumption that the transfer of the covered work may occur as part of a sales transaction (e.g., sale of media containing the work) from the licensor to the licensee, which is a situation in which implied warranties may be applicable.
WTFPL is absolutely not suited for that use case, at least on its own, though nothing prevents you from including both the WTFPL and a warranty disclaimer.
> B. You can sell something for zero dollars.
You can't have a sale without a contract, and while you can have a contract without money changing hands, you can't have a contract with an intended two-way exchange of obligations. One party merely offering a limited waiver of some of its exclusive rights without asking anything in return that it wouldn't be entitled to anyway isn't a two-way exchange of obligations.
> The gratuitous license argument is interesting (though from what I can tell, legal minds greater than mind still feel the urge to disclaim warranties in that case), but the contract/sale one is just completely uninteresting to me
I don't think you understand -- they are exactly the same argument. A sale involves a contract. A license without a contract is gratuitous.
You can't be interested in the argument of whether p is true but not be interested in the argument of whether ~p is true.
Now you are just being a dick. I"m trying to be nice here and make sure i understand your argument. In most cases, courts won't look at consideration at all, and outside of the US, whether consideration is required at all varies.
" you can't have a contract with an intended two-way exchange of obligations. "
This is false, of course. You can have unilateral contracts accepted by performance of something that a contract requests, contracts implied in fact, contracts implied in law, and all sorts of interesting quasi-contracts, plenty of which still have warranties.
But that's also "first week of contracts".
"I don't think you understand -- they are exactly the same argument. A sale involves a contract. A license without a contract is gratuitous."
Not in all cases, actually, but if you are going to act like you have above, i'm not horribly interested in continuing the conversation, we'll just see what happens when someone gets sued over the WTFPL.
> This is false, of course.
Except, not.
> You can have unilateral contracts accepted by performance of something that a contract requests,
You can have contracts accepted by performance as described, but they aren't unilateral. Acceptance by performance is a mechanism for demonstrating acceptance (hence the name) of a contract offer with mutual obligation while simultaneously fulfilling some or all of the obligations on one side.
> contracts implied in fact,
Which are contracts where the acceptance of mutual obligation (and perhaps the actual content of the mutual obligation) is inferred from non-verbal communication, not an exception to the requirement for an intended two-way exchange of obligations.
> contracts implied in law, and all sorts of interesting quasi-contracts,
quasi-contracts are another name for implied-in-law contracts, which are not actual contracts, but equitable arrangements; given the nature of the equitable basis for these -- particularly the direction they run -- it would be odd to see implied warranties, or even some equitable analog, mattering to them except to reduce liability that might be due to the provider. But equity can be be weird, so I wouldn't rule it out entirely that there might be some relevance there.
> But that's also "first week of contracts".
Indeed.
Since the license wants to be enforceable, it goes out of its way to create a contract. Are you saying that the simple statement "Use this as you see fit" would also be held to be a contract, even in the absence of consideration? If so, why all the convolutions and verbiage in the current licenses?
And getting back to disclaimer of warranties, what about the absence of any stated license? Do you feel that a disclaimer of warranty is no more or less necessary with a license than without?
By that logic, public domain is "ambiguous".
> There's nothing to stop me from taking your code, re-licensing it under say a proprietary license, then suing the original creator.
By that logic, virtually all the "BSD-style" licenses and their ilk are "ambiguous".
> There's nothing that stops me from taking your project's branding/symbol/graphics/name and claiming it as my own and then suing you.
By that logic, virtually every FOSS license that doesn't extend to branding (e.g. pretty much everything other than the MPL) is "ambiguous".
> There's nothing that stops me from taking your work and then not providing attribution back to the original creator.
See the public domain remark above.
> By default there is nothing that absolves the license holder from warranty nor implied fitness for use.
That is true, and would be a worthwhile addition as a default rather than an afterthought.
> legally, what does "DO WHAT THE FUCK YOU WANT" actually mean?
It means "do what you want". Unless you're claiming that lawyers don't have a strong-enough grasp of the English language to be able to properly interpret extraneous expletives (which wouldn't surprise me to be quite honest...).
The notion of public domain can vary from country to country. The first time I wanted to release something inconsequential on Google Code I was surprised that it didn't let me choose "Public domain" and forced me to choose a license. Here's their FAQ about it: http://code.google.com/p/support/wiki/FAQ#Can_I_host_code_th...?
This is a very simplistic view of the phrasing.
For starters, when I ask someone "hey do you mind if i do x", and they say "do whatever you want", it doesn't usually mean "yes it is okay if i stab you with a knife", it means "yes it is okay to do x".
This phrase does not have a single, simple definition that is clear and robust.
Since this is in the scope of, say, a program's source code, the implicit meaning would be "do what you want [with this code]". Explicitly stating that would help, but it's not really that necessary unless you feel like being needlessly pedantic.
Virtually all license agreements are also implicitly bound by jurisdictional laws, so "yet it is okay if i stab you with a knife" would not be a valid interpretation of "do whatever you want" - in a legal sense - unless the agreement is taking place in a jurisdiction in which stabbing someone with a knife is legal.
Depending on how reasonable x is, a judge may or may not buy it.
Note also that the suggested license header is: "/* This program is free software. It comes without any warranty, to * the extent permitted by applicable law. You can redistribute it * and/or modify it under the terms of the Do What The Fuck You Want * To Public License, Version 2, as published by Sam Hocevar. See * http://sam.zoy.org/wtfpl/COPYING for more details. */"
This is also bad if it really lets you do whatever you want - The clause "use it or modify it under the terms of ..." implies there are limitations on how you may use/modify it - in particular, only under the terms of the WTFPL. If the WTFPL truly has no restrictions, this wording is meaningless. Under various canons of interpretation, a judge will usually find that people didn't mean for words to be meaningless. What limitations/meaning will be ascribed is also up in the air.
Realistically, the right wording to use there would have been "you may use it without restriction, as the WTFPL version 2.0 explains" (or something similar).
Again, you can argue this both ways. But it is just not the certainty it seems some want to ascribe to it.
As I stated above, the artwork is now free even by Debian guidelines, so the only thing holding back Debian there seems to be inertia.
So how exactly can Debian patch Iceweasel using Mozilla's patches between October and January?
https://lists.debian.org/debian-devel/2005/06/msg01182.html
In 2006, Mike Connor contacted Debian and stopped just short of saying outright that Gervase should never have done that:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=354622#59
(Also see the next few messages from Eric and Mike where they worked out some of the details)
So the short version is that there was a lot of confusion from Mozilla about what Debian should be doing here, and at each step Debian adjusted the contents of the package in accordance with the information given.
As far as the artwork license is concerned, that's actually a distraction; the conclusion of the discussion with Mike Connor was that Debian could only use the firefox trademark if the Mozilla Foundation was pre-approving all releases, and it would be quite tricky to make this work with Debian's release processes. Using the iceweasel branding is much less work for everybody and doesn't seem to cause any problems for users.
"I'd be for this [removing offending PHP modules] in a heartbeat if it would make people switch to a saner programming language, but that's wishful thinking."
I don't think most developers would miss the craziness, I don't think the features are tied to the craziness of the language and I'm sure a lot of us would change as soon as it became production ready (if not earlier.)
(Hoping for Hack to be a step in the rigth direction.)
It all comes down to semantics of what linking means. The PHP license is pretty much identical to the Apache license and you could indeed make a case for not allowing any GPL'ed software to be "linked to" from Apache either.
See http://www.apache.org/licenses/LICENSE-1.1.
The PHP license was chosen to match the Apache license because Apache and PHP are tied so closely to each other.
This hair splitting over linking, derivation and aggregation has been going on since the beginning of time. My stance is that you can indeed ship PHP licensed PEAR components on the same cd or in the same tarball as GPL'ed code because I see it as an aggregate work. This changes if you take PEAR code, modify it and copy-paste it directly into your own work. Then it moves from aggregate to derived. But the intent of the PEAR components is to be used in aggregate form. The PHP license allows you to use it in derived form as well, of course, but then you should be choosing a license other than the GPL for the derived work.
The FSF has a FAQ on aggregation here: http://www.fsf.org/licensing/licenses/gpl-faq.html#MereAggre...
That text is heavily biased towards compiled software and they talk about executables and memory spaces which don't really apply in this case. If you don't consider using a PEAR component as aggregation then it logically follows that you also cannot have Apache call your code so you will have to stipulate that nobody can use your code from Apache. I think this is an extreme interpretation that pretty much nobody out there shares.
In short, I don't see an issue here. Move along.
For PECL extensions that are linked into PHP, the license must be compatible with the PHP license. That means you can not GPL a PECL extension or you would be violating the GPL. Note also that if you write an extension that links against a GPL'ed library you will be violating the GPL. If you need to link against a GPL'ed library, get permission from the author of the library to use the library under a compatible license.
The license of any PEAR/PECL package can be found in the head of all source code files, inside the <license> tag of the package description file (package.xml) and also on the package homepage.
It's things like this that make me think "what did we do to deserve PHP?"
That's not so simple; as a work produced by someone, a legal document may be subject to copyright. So you might need a license to the license to modify and reuse it :)
http://www.adamsdrafting.com/the-contract-drafter-as-copyrig...
The problem is that lots of other projects need to apply some license, and just for simplicity say "I'll just apply the PHP license," even though it makes absolutely no sense for anything but the PHP project itself.
It makes complying with the license impossible.
I don't think it affected the interpretation much. If the author and the regents disclaim the warranty, and there are no regents, the straightforward interpretation is that the author disclaims warranty.
(I believe quite a few bits of Windows will still, thanks to advertising clause, remind you about the Regents of the University of California)
The reality is that the PHP license is a license which exists solely for PHP itself and makes no sense whatsoever for projects which are not PHP, leaving them essentially license-less and thus "all rights reserved".
You could replace "license" with a lot of other words and it would still a possibly accurate statement. "mysql api" comes to mind, as well as "classes" and "URL query variables like $_GET"...