It is also the reason why the OSI keeps a list of approved licenses.
Far more apropos in the context of this thread is the FSF's license list and explanation of copyleft. Debian's concerns map far closer to Free Software movement's concerns than the Open Source movement's concerns because Debian cares about software freedom. The Open Source movement dismisses software freedom and advocates licenses that constitute gifts to proprietors (including the X11 license and the new BSD license) because businesses are the main target of the Open Source message. The Free Software movement, on the other hand, is concerned with software freedom and insuring these freedoms for all computer users. This insurance is one of the reasons why the GNU project (or perhaps it was the FSF) invented the concept of copyleft.
If you didn't take time to understand the license before contributing to the project, you can't complain that you don't like the result.
Just because you contribute something to a project doesn't mean you relinquish your copyright power. This is why the Mozilla project is currently looking for some authors who contributed copyrighted work to that project. The Mozilla project can't change the license on these contributions without the copyright holder's permission.
Would you get one of, say, the media player apps if you couldn't get it for free online?
Yes, if it were Free Software and priced reasonably, but that's completely beside the point. This is a market issue, not a problem for the GNU GPL. Ultimately network distribution is something you don't have the budget to do in the way you want to do it. So instead of wallowing in what you can't afford to do, try distributing your program in a way you can afford.
I understand that the GNU GPL does not require this...
Then this thread was always completely outside the context of the GNU GPL. This thread is becoming less and less apropos for this discussion topic.
I have to pay upwards of $200 per year for an SSL certificate to protect my e-commerce from the man in the middle, don't I?
Well, if you can't afford that, you'll have to reconsider doing e-commerce that way (maybe you won't do e-commerce at all). This isn't brain surgery and you're not running a huge store. It's also not my responsibility to come up with a viable business plan for you.
I have considered this path [using GPG in combination with contacting Oberhumer].
Your posts convey no sense of this at all. In fact you tried to take me to task by pointing me to a webpage you thought would corroborate your assumption you could only contact this person via PGP encrypted e-mail. It turns out that wasn't the case; apparently you didn't read that page too closely.
However, when I looked into GPG, I found no way of getting my key signed by a prominent GPG user, or even by any GPG user. Isn't that a prerequisite for entering the "web of trust" or whatever they call it?
Briefly: No it is not, and if you read documentation on GPG you'll see you what you need to use encrypted and signed e-mail.
But again, there's a more important issue here. You need to ask yourself whether you need to contact Oberhumer at all. I don't think you do. You don't need to obtain another license for Oberhumer's work because you have no genuine problem complying with the GPL. This was the only reason for contact and since this turned out to be another non-problem, the need to contact Oberhumer vanishes.
You try selling a game that doesn't have a downloadable playable demo. Players like to know what they're getting into before they buy something.
Welcome to the free market. Nobody owes you bandwidth or storage.
The sheer size of the source code may prohibit me from distributing the software electronically at all.
So don't distribute your work electronically. Distribute source and binaries together simultaneously by burning both onto CD-Rs. Someone else might volunteer to redistribute the work electronically for you. So much of this issue centers on your erroneous belief that the GNU GPL requires you to distribute electronically.
I am not [the only copyright holder to this work]. Read what I had written...
I did read it, quite carefully. I figured at least one of the following things might have occured to you:
The last paragraph of section 3 of the GNU GPL says that "offering equivalent access to copy the source code from the same place" will meet your requirement to distribute complete source code to your work. That place can be a CD-R.
You could accede to Oberhumer's wishes and try GPG. Contrary to your claim "I cannot contact Mr. Oberhumer because he does not accept mail not encrypted with PGP", there is no requirement to use PGP. Oberhumer supplies information for contact with GPG on the page you pointed me to. GPG is a a Free Software PGP replacement.
You could see if Oberhumer supplies a postal address with your copy of the software then try sending a letter. Low-tech, yes, but it could work to make initial contact.
This is a non-problem if and only if reversibly concatenating an executable with asset files counts as either "mere aggregation" and/or "just data" under the GPL.
I very much doubt that distributing a single file would count the way you want it to be counted. That's a question for the FSF, albeit a rather academic one considering that this whole issue is a non-problem if you drop the phony electronic distribution requirement (that never really existed) and distribute your program via inexpensive CD-Rs (which have approximately 600MB more capacity than you said you need to distribute your program).
...[H]ow should I distribute my software electronically in order to comply with the letter and spirit of the GNU GPL?
Exactly what part of the GNU GPL are you referring to? I'm not aware of the part of the GNU GPL that requires "electronic" distribution. Also, if the work is not a derivative of anything GPL'd, the GNU GPL would not apply to you (assuming you are the copyright holder on this work). It would only apply to those you distribute your work to.
But didn't it make a difference that he was putting a URL to *his* website on the CD?
Put briefly, no. That makes no difference. Nowhere in the GNU GPL does it say Fred can distribute a URL pointing to the source code to satisfy his obligation to make source code available. Therefore choice 1 is the correct answer to that question.
I had thought that this issue was precisely what the Lesser GPL was supposed to solve though, all this seemed to do for me was raise the point that maybe Microsoft has a valid arguement.
No, the LGPL was never meant to be a gift to business. That concept is the motivation of another movement, not the Free Software movement. Read the FSF's essay on the LGPL to learn what the LGPL is meant to do.
Also, Microsoft has no valid point regarding the GNU GPL and Microsoft doesn't act in accordance with their lies. Nobody owes them the ability to make derivative works. It is not the Free Software developer's responsibility to make sure software proprietors have code to extend. This is one of the points that separate the Open Source movement (who strongly advocates the X11 and new BSD licenses, both of which allow for proprietary derivatives) and the Free Software movement (which chiefly advocates the GNU GPL, a license that does not allow proprietary derivatives).
Under those circumstances there is no way I am going to be able to offer my application under Linux without an awful lot of reinventing the wheel.
You could release your software under the GNU GPL and write derivative code from any GPL-compatible licensed work you wish. You could sell the software and support for the software, even sell warranties for your work. Your competitors will not be able to distribute derivatives of your software without using the GPL for their releases, thus allowing you to share their distributed improvements.
Does someone want to tell me if I've lost the plot, and why?
You have misunderstood what copyright law allows for. This issue has little to do with the GNU GPL and LGPL per se, more to do with how copyright law deals with output of processes and copyright power regarding derivative works. Like any other software license, the GNU GPL and LGPL inherit many things from copyright law, one of them being that they cannot control the licensing of the output of a process (there are exceptions but I'm trying to keep this concise and simple). So, even though gcc is GPL'd, you can compile non-GPL'd software with gcc.
Linking, however, is a different issue. Linking constitutes a derivative work which does fall under the aegis of copyright law. Therefore the license for your derivative work is not completely under your control. Different licenses offer differing levels of permissiveness for creating and distributing derivative works. For example, if you were to distribute your work linked against the LGPL'd GNU C library you would have to follow what the LGPL says for licensing your derivative work. The LGPL does not make you use the LGPL as the license for your derivative work, but it has some say in what the license for your derivative work must allow.
I suggest you read a lot of licenses very carefully and weigh your options. Keep in mind that nobody owes you the ability to make or distribute derivative works. If you are unwilling to agree to the terms of someone's license, write your own code.
Looking at the overall gist of your post, it's clear you don't understand a good deal about Free Software or why the GNU GPL says what it says. You should read the essays at the GNU project's philosophy page and learn about the Free Software movement.
Ensuring freedom for who, exactly?
All computer users. The Free Software movement's message of software freedom is aimed at all computer users.
When I develop software "for free", I want people to use it.
The Free Software movement is not just interested in popularity. The primary motivation of the GNU project is to give users freedom. Also, talking about developing software "for free" suggests you are talking about software in terms of price. The Free Software movement focuses on freedom, not price.
But most importantly, the GPL fails in its efforts to encourage code sharing, because it locks out the largest potential resource of high-quality, useful code: business.
The GNU GPL is a very (if not the most) widely used Free Software license. Lots of people find it works for code sharing very well. However, it is not the job of the Free Software movement to look out for business interests in the way you want them to (treating businesses as charities is the work of another movement). If businesses want software developed, they can work hard and develop it themselves or hire someone to write it for them. If they want to share and share alike, they are welcome to get a head start and develop a derivative of a GNU GPL-covered work. Many businesses have chosen the latter route; the GNU GPL does not "lock out" business users, unless you define "business" to exclude IBM, Microsoft, Apple, and a host of commercial GNU/Linux distributors (just to name a few) all of whom make and/or distribute GNU GPL-covered software.
BSD licenses OTOH allow everyone to share. Sure, some people abuse the situation. But the community loses nothing in real terms.
The second sentence of that quote denies the validity of what you said in the first and third sentences of that same quote. Anytime someone distributes new BSD-covered code they are distributing Free Software, and that's great. Unfortunately they've chosen a license that allows for proprietary derivatives (even incompatible proprietary derivatives—a practice commonly referred to as "embrace and extend"). The community cannot freely use the changes made to these derivatives and the developers of the new BSD-licensed work have essentially treated some organization as a charity, not insisted on developing software as equals.
What really worries me is that gcc (and glibc) may be too restrictive, licence wise, to develop closed software at all - in which case I kind of wonder why FSF have not gone after those who do so.
Copyright law doesn't allow the FSF to stop you from using gcc to compile non-free software. For years, NeXT (now Apple) used gcc to compile NeXTSTEP and OPENSTEP (which, contrary to whatever you may read into the name, is a proprietary operating system). It was a completely seperate issue that NeXT tried to extend gcc without abiding by the GNU GPL by proposing to distribute their Objective C front end as.o files.
This doesn't respond to the point made earlier and it doesn't acknowledge the power of patents. The point made earlier was that with the BSD licenses, you can always benefit from someone else's contribution. So still being able to use the code without the patented contribution in no way addresses that point because you are talking about the program without discussing the patented contribution. Also, what you gain with copyright (i.e., "still hav[ing] the original code") you can lose with a patent. Copyright law and patent law are completely separate areas of law that can conflict. A patent holder (whether it is the one that "contributed" to your BSD-covered program or someone else) can write their patent so broadly it encompasses your extant work, thus preventing you from distributing your work, forcing you into a costly licensing fee, compelling a court appearance, or cross-licensing your patents with them (the real value of software patents is in cross-licensing).
2) it is unclear how the patented code was 'contributed' to the project in the first place.
The code could have been sent to you to make it known to you, informing you of its existance in a way where they can win a copyright infringement suit (see Lawrence Rosen's column in the December 2001 "Linux Journal" magazine for details on this) as well as a patent infringement suit. Whatever the reason, they have a patent on that improvement and you probably don't. Patents are written broadly enough to encompass an idea (not just a particular expression of that idea), so it's entirely possible (even likely) you won't be able to reimplement the improvement without infringing on the patent.
Rather, it seems that the people who developed the patented code did exactly what they are allowed, which is they used the BSD code freely.
There's no doubt they acted within the limits of the BSD licenses, that was never the question. The question was whether you want that to happen to your program: Do you want to maintain the ability to share improvements to your program even in the face of software patents? Do you want the other users of your software to be able to share and modify the patented improvements? If so, the BSD licenses are not an appropriate license choice for your program.
What, exactly _is_ the point of LGPL if you can't link an LGPL lib with proprietary code without opening that code?
The FSF has an essay directly answering your question: Why you should not use the Library GPL for your next library?. The GNU LGPL was renamed after this essay was written to stand for the Lesser General Public License to reflect that it does less to ensure software freedom.
Seems there is room to define a license like "This is GPL, with the exception that you may freely dynamically (ie. at runtime) link this software with software bearing any license without any obligations of the GPL having any effect on that software".
Such a license would go against the whole point in using the GNU GPL. Free Software developers have to make advantages for one another and produce software that is Free Software. Linking with non-free software is antithetical to the Free Software movement because this movement maintains everyone should have the four fundamental freedoms of Free Software for all of their software. Distributing software linked with non-free software prevents people from having complete software freedom (more specifically, studying, modifying, and distributing source code).
...[W]hen legal considerations become so pervasive and convoluted, so that one must be a specialist,i.e. a lawyer, to properly understand them, the law is no longer helpful, but a detriment and distraction to the task at hand, in this case computer programming.
Some might say the same thing about computer programming or using a computer recreationally. Fortunately we don't need to be experts to understand a great deal about copyright, we need to study it, ask questions, have discussions, and learn. There are so many interesting intersections between computing, economics, law, and politics, I sincerely doubt a reasonable person could find only "smallish" interests after reading some good books.
Understanding a system is the first step to changing it. Try to understand what's going on with copyright law by reading non-technical books written for a general audience like Lawrence Lessig's "The Future of Ideas" or Siva Vaidhyanathan's "Copywrongs and Copyrights". Both of them contain plenty of commonly accessible examples on where the law and computing intersect. Then try reading the references they point to, such as US Supreme Court cases and important appellate court decisions. I've found that the more I persist by reading and learning, the more I come to understand the ins and outs of copyright law and the more I appreciate the effort.
Yes, I am distributing the GPL-covered program to these people. They can request source code from me. If I am unprepared to satisfy that request it is my responsibility to not distribute the program to them in the first place.
Nobody said Fred couldn't do (I), the question is what fulfills his obligation to distribute source to the modified browser.
To whittle down the answers on that particular question, keep in mind what constitutes distributing complete source code to Fred's binary. Obviously Joan's source code doesn't have Fred's modifications, so I is out. II, III, and IV might look good until you notice that III only distributes patches to Joan's original source. That's not good enough in the event that you want complete source code to Fred's derivative and the source code Fred generated his patches against is inaccessible, so III is out. That leaves II and IV, which are equivalent and good (putting aside the difference betwen "alongside" and "along with"). [I know the real standard here is what the GPL says, not what parallel scenario I can dream up to justify the GPL's clauses, but I'm trying to write something accessible to those who aren't so familiar with the GPL.]
So once you modify a GPLed program, you're required to only distribute the modified version and never the original??????
No, that's not what it says at all. Whether you distribute the original source code is up to you. The GPL only requires you distribute the source code to the binaries you're distributing (if you're distributing binaries). Recipients of the binary should be able to modify it, and that requires the source code that corresponds to their binary (not some other version, no matter how improved it may be).
#3 seems hypocritical--RMS can create all the restrictions he wants, but any code that links with his license cannot have any?
Question #3 is aimed at making sure people don't try and sneak around the GPL by linking in code that is not available under the same permissions as the GPL. It's important to make sure people can share and modify the whole GPL'd program, not just part of it. The GPL allows Fred to charge for distributing Joan's browser or any derivative of Joan's browser. So linking in code that can't be distributed for a fee is linking in GPL-incompatible code.
Therefore the acceptable answers are quickly narrowed to the "Yes" answers (1 and 2): the sample license violates the GPL. Of those two, answer 2 is correct because the section 6 of the GNU GPL states that Fred cannot apply additional restrictions to any part of the code. Answer 1 is incorrect because the GPL only requires distributing binaries linked against GPL-compatible code and the GPL is not the only GPL-compatible license.
Maybe I'll have to come up with my own, however I hear that it comes with legal tanglings and etc.
You heard correctly, writing your own license is very difficult. It is not to be taken lightly. The GNU GPL has seen real court time and been confirmed as an enforceable and binding license. I think most people would be hard pressed to create a license that can survive court examination for the entire license.
Although I was thinking of coming up with one that disallows a bunch of people and companies that've pissed me off from using my program. Maybe I should include RMS on the list!
Before you rush off in a huff to write a license you'll most likely regret, consider that your aggravation stems from your misunderstanding of what the GNU GPL and GNU LGPL say. Since you got 3 questions wrong, you shouldn't be so quick to judge RMS, the GNU GPL & LGPL licenses, or the FSF harshly. I suggest jotting down your objections, learning what the Free Software movement is after, how the movement RMS started goes about achieving their wishes, learning what obstacles stand in the way, and then see if your objections have merit.
You sound like you're talking about Microsoft, the chief objector to the GNU GPL, and their FUD speeches attacking the GNU GPL (in a similar name-calling, no-specifics-required, mud-slinging way). But for all their flaming, they're not doing the one thing you'd think they would do given their position: avoid distributing GNU GPL-covered works.
No. The FAQ explicitly states that you must offer the source through mail for a small fee if you do not ship source with the binaries. It doesn't matter how you distribute the thing.
Actually, when it comes to distributing binaries without source code and you want to go with the written offer option, section 3b of the GNU GPL says the applicable cost is "no more than your cost of physically performing source distribution". So you don't have to charge anything if you don't want to.
But on a larger issue, your rejection is simultaneously unclear and doesn't invalidate anything you quoted from me. I'm fully aware of the requirement of offering source code through the mail when someone orders it (I never contested this requirement). You read something into my words that was not there. I was not trying to present a complete list of alternatives (the GNU GPL itself is fine for that).
Now, a lot of people would think that if you distribute the binary via FTP, having a source package right next to it would be 'distributing source with the binary', just with the option for the user of actually taking it or refraining as they see fit. I strongly suspect that FSF:s interpretation rather is that the source must accompany the binary in the same, single, package.
Consider your suspicions allayed. One question speaks to your suspicions. Distributing two archives as you describe is fine. Distributing source code in the same archive as the binaries is also fine. Network distribution makes it possible for the binary archive to be accessible where the source archive is not (this can happen via FTP, not just via web sites). This presents a problem and is easily solved by distributing source code and binaries together in the same archive.
The argument for mail is that people may not have access to a network (which they evidently had thirty seconds earlier, when downloading the binary).
When the GNU GPL was written not everyone had network access that allowed them to download all the source code they want to GPL'd programs. People still don't have that level of access. Networks do go down, some people pay per byte and have to ration their usage, and a variety of other limitations. The GPL has to account for people passing copies of GPL-covered software via other means too. I don't know about you but I have manually passed out copies of GPL-covered software to people (I always put the complete machine-readable source code on the same medium as the binaries if I'm distributing binaries). These people deserve source code too.
At the same time, there seems to be no specification as to what physical medium to use; I could mail a QIC40 tape with the source and not having to worry if you can dredge up a tape unit able to read it. Yes, this is stupid.
If you're only distributing source code, you would want to distribute it in a medium people could use otherwise people would not want your source code. If you're distributing binaries, section 3a specifies "a medium customarily used for software interchange", section 3b specifies a written offer (valid for at least 3 years) that offers source distribution on "a medium customarily used for software interchange", and section 3c talks about passing on a copy of the information you received in accordance with section 3b.
Oh, BTW, if question #8 is about
static linking, please make this clear; as it stands, you can answer either way, depending on how you interpret the question.
It appears you are trying to send a message to the editors of the quiz. You should know that I did not write the quiz.
Question 4 sets the scene here, "FooCorp distributes a modified version LibIdo library linked to their proprietary program Frobber.". The GNU LGPL notes in the preamble, "When a program is linked with a library, whether statically or using a shared library, the combination of the two is legally speaking a combined work, a derivative of the original library." and thus subject to the terms of the license of the work being linked in (in that quiz scenario, the LGPL).
Section 6 of the GNU LGPL requires the license under which you distribute a "work that uses the Library" to allow for "modification of the work for the customer's own use and reverse engineering for debugging such modifications". So it would not matter whether Frobber links in the modified LibIdo statically or dynamically, either way Frobber is a derivative of the modified LibIdo and that LibIdo is a derivative of Peter's library. Therefore, in the end, the proper answer to the question #8 is "yes".
Since when did legal studies become the main concern for computer afficianados[sic]?
I was thinking just the opposite: When was it reasonable to be ignorant of the laws that control what you can do with a computer? When was it a good idea to not know about the terms of widely-used Free Software licenses? Maybe the computer afficionados you know would be better off knowing what copyright law says and what penalties lie in wait for them if they violate someone's copyright. Being ignorant of the DMCA is particularly unwise in the USA, just ask Dmitry Sklyarov about the pleasures of being held in jail away from home.
I think Microsoft hacked into FSF's servers to scare people away from open-source.
Funny. I think the FSF put that up with the mindset of helping people to better understand the GNU GPL and GNU LGPL, and to increase the number of volunteers helping them with license issues (as they ask for if you answer all the questions correctly). Also, the FSF doesn't do "open-source" they do Free Software. There's a big difference.
These legalisms are paralyzing to a fertile, inquisitive mind.
These legalisms are just another part of life to learn about, organize protests around, and work within. A mind will not stay "fertile" if it is ignorant of the rules of society. It is shameful, not helpful, to remain ignorant of the law.
So long as Fred and Joan aren't a part of the same organization, distribution has occurred. The GNU GPL calls such an event "distribution" and that is the term I should have used.
...and location is hardly the issue.
Distributing source code and binaries from the same place is apropos because the GNU GPL says it is. See section 3 of the GNU GPL which includes:
If distribution of executable or object code is made by offering access to copy from a designated place, then offering equivalent access to copy the source code from the same place counts as distribution of the source code, even though third parties are not compelled to copy the source along with the object code.
In the quiz scenario, Fred distributes a CD with binaries of Joan's web browser to his unnamed friend. Giving out source code on the same CD as the binaries would have counted as distribution of the source code. Hence choice 2 ("He [Fred] can give out source on the same CD as the binaries.") is not the correct answer to the quiz question.
[All spelling in context.] Atleast with [the BSD licenses] I'm guarenteed that contributions back to software I wrote a majority of won't prevent me from using the damn stuff myself.
Someone could patent the code they contribute to your project and not license that patent to you. The BSD licenses do nothing to help you with software patents. The GNU GPL covers this situation.
I understand your desire for a simple permissive software license, but ensuring software freedom requires more complex licensing terms than what the BSD licenses provide.
While the FSF is technically right here, the example is silly. Are any software authors, even the FSF, really going to sue someone for giving binaries as a birthday present to a friend without source?
The example is not silly at all. People do distribute CDs of binaries without source code. The FSF won't sue someone who uses the GNU GPL to license their software in cases where the FSF is not the copyright holder (which is the vast majority of GNU GPL-covered software). Even in cases where the FSF holds the copyright, usually these cases are not settled by lawsuit. Usually infringements are settled by someone pointing out the GPL's terms and asking for compliance (sometimes with the assistance of Eben Moglen, general counsel for the FSF, who said, "[a] quiet initial contact is usually sufficient to resolve the problem.").
Also, the problem doesn't come up merely by not distributing source code with binaries, but by not making it available at all. One can comply with the GPL in ways other than distributing source code with binaries. Distributing both source code and binaries simultaneously from the same location is probably the most convenient way to comply.
For that matter, what cheapskate would give free software as a gift? This makes no sense even keeping the
gratis vs. libre issues in mind.
It makes sense when you consider that not everyone gives gifts to demonstrate how rich they are. Some people give gifts to show that they care about their friend's wishes. Sometimes this involves buying something, sometimes it involves spending time making something that has no price attached to it. Giving someone a copy of Debian GNU/Linux who can not get it on their own is one way to help them have software freedom.
I dare say that this example may highlight a situation where the GPL needs to be updated. Binary-only distribution to personal aquaintances of unmodified versions
should be perfectly fine.
No, the GPL does not need amending here. People need to be able to excercise all of their software freedoms including the ones that require source code: studying how the program works, adapting it to their needs, improving the program, and releasing their improvements to the community (known as freedoms #1 and #3 in the definition of Free Software). In order to come as close as one can to guaranteeing source code availability and offer reasonable practical alternatives, the GNU GPL includes the terms discussed in the first quiz question.
If I share software with you I can't simply tell you where the source is, I have to give it to you or make a written offer? Why can't I just share and get on with my life?
Why can't you ship source code and the binaries from the same location and get on with your life? One way to comply with the GNU GPL is to make source code available from the same location as the binaries. Websites become unavailable for various reasons and if people have only the binary and a pointer to the source code on a website, it's possible they would not be able to obtain the source code that compiles to the binary they have. This is not unheard of when people distribute binaries via CD. But my experience is most people don't find it difficult to ship source code and binaries from the same location.
However, if you read the bill (and the rest of the letter for that matter), you'll see that it describes perfectly OPEN SOURCE SOFTWARE.
No, it doesn't (and your point is not well served by pointing out how the letter starts off by correcting the use of that term). The Open Source movement is primarily targetting businesses. The Free Software movement speaks to all computer users. The Peruvian bill speaks to all Peruvian computer users, not just businesses. This is one way in which the wording in the letter is correct.
The bill is concerned with user's freedoms. The Open Source movement takes pains to avoid talking about software freedom because they believe their development methodology message will go over easier with businesses if they stop the freedom talk (in their FAQ they refer to freedom talk as "ideological tub-thumping" with no further analysis of the Free Software movement's motives). This is another way in which the wording in the letter is correct.
See how there's no mention of any obligation to redistribute the modified source.
I see now where your misunderstanding lies: There is no obligation to redistribute programs in the Free Software movement. The Free Software movement is geared around freedoms you need in "the system of free expression in a technological society" (from the FSF's Amicus brief for Eldred v. Ashcroft). Nowhere in the Free Software definition is there a requirement to distribute Free Software. Quite to the contrary:
You should also have the freedom to make modifications and use them privately in your own work or play, without even mentioning that they exist. If you do publish your changes, you should not be required to notify anyone in particular, or in any particular way.
They must have the right to do so but there is no requirement to do so.
Exactly—users must have the right to do so. That's freedom talk which is totally synonymous with the Free Software movement, not the Open Source movement.
I think any reasonable person has to conclude that Dr. Villanueva knew what he was talking about and he meant what he said throughout the letter, including his repeated support for Free Software.
The problem with "free" is a problem in English, not other languages where the difference between price and liberty are clear. The FSF offers a list of translations to use for Free Software when speaking other languages so people understand you're not talking about price (yes, I know the link points to a term on "Freeware", that's where they put it because the list answers another question relating to the term "Freeware").
The fuss regarding the phrase "open source" has to do with using the name of the wrong movement to talk about what Dr. Villanueva refers to in his letter. Early on in the letter he corrects the Microsoft rep's use of the term, telling Microsoft he is speaking of Free Software, not Open Source. There is a big difference between the two movements.
Well, this, and by this I mean the whole imbroglio where Dr. Villanueva is now the cause celebre of the open source movement...
Please make the effort to use the term Free Software (not "open source") when referring to the movement referred to in the letter. The bill and the letter all refer to only Free Software. The Open Source movement stands for something quite different.
Slashdot Mirror
Far more apropos in the context of this thread is the FSF's license list and explanation of copyleft. Debian's concerns map far closer to Free Software movement's concerns than the Open Source movement's concerns because Debian cares about software freedom. The Open Source movement dismisses software freedom and advocates licenses that constitute gifts to proprietors (including the X11 license and the new BSD license) because businesses are the main target of the Open Source message. The Free Software movement, on the other hand, is concerned with software freedom and insuring these freedoms for all computer users. This insurance is one of the reasons why the GNU project (or perhaps it was the FSF) invented the concept of copyleft.
Just because you contribute something to a project doesn't mean you relinquish your copyright power. This is why the Mozilla project is currently looking for some authors who contributed copyrighted work to that project. The Mozilla project can't change the license on these contributions without the copyright holder's permission.
Yes, if it were Free Software and priced reasonably, but that's completely beside the point. This is a market issue, not a problem for the GNU GPL. Ultimately network distribution is something you don't have the budget to do in the way you want to do it. So instead of wallowing in what you can't afford to do, try distributing your program in a way you can afford.
Then this thread was always completely outside the context of the GNU GPL. This thread is becoming less and less apropos for this discussion topic.
Well, if you can't afford that, you'll have to reconsider doing e-commerce that way (maybe you won't do e-commerce at all). This isn't brain surgery and you're not running a huge store. It's also not my responsibility to come up with a viable business plan for you.
Your posts convey no sense of this at all. In fact you tried to take me to task by pointing me to a webpage you thought would corroborate your assumption you could only contact this person via PGP encrypted e-mail. It turns out that wasn't the case; apparently you didn't read that page too closely.
Briefly: No it is not, and if you read documentation on GPG you'll see you what you need to use encrypted and signed e-mail.
But again, there's a more important issue here. You need to ask yourself whether you need to contact Oberhumer at all. I don't think you do. You don't need to obtain another license for Oberhumer's work because you have no genuine problem complying with the GPL. This was the only reason for contact and since this turned out to be another non-problem, the need to contact Oberhumer vanishes.
Welcome to the free market. Nobody owes you bandwidth or storage.
So don't distribute your work electronically. Distribute source and binaries together simultaneously by burning both onto CD-Rs. Someone else might volunteer to redistribute the work electronically for you. So much of this issue centers on your erroneous belief that the GNU GPL requires you to distribute electronically.
I did read it, quite carefully. I figured at least one of the following things might have occured to you:
I very much doubt that distributing a single file would count the way you want it to be counted. That's a question for the FSF, albeit a rather academic one considering that this whole issue is a non-problem if you drop the phony electronic distribution requirement (that never really existed) and distribute your program via inexpensive CD-Rs (which have approximately 600MB more capacity than you said you need to distribute your program).
Exactly what part of the GNU GPL are you referring to? I'm not aware of the part of the GNU GPL that requires "electronic" distribution. Also, if the work is not a derivative of anything GPL'd, the GNU GPL would not apply to you (assuming you are the copyright holder on this work). It would only apply to those you distribute your work to.
Put briefly, no. That makes no difference. Nowhere in the GNU GPL does it say Fred can distribute a URL pointing to the source code to satisfy his obligation to make source code available. Therefore choice 1 is the correct answer to that question.
No, the LGPL was never meant to be a gift to business. That concept is the motivation of another movement, not the Free Software movement. Read the FSF's essay on the LGPL to learn what the LGPL is meant to do.
Also, Microsoft has no valid point regarding the GNU GPL and Microsoft doesn't act in accordance with their lies. Nobody owes them the ability to make derivative works. It is not the Free Software developer's responsibility to make sure software proprietors have code to extend. This is one of the points that separate the Open Source movement (who strongly advocates the X11 and new BSD licenses, both of which allow for proprietary derivatives) and the Free Software movement (which chiefly advocates the GNU GPL, a license that does not allow proprietary derivatives).
You could release your software under the GNU GPL and write derivative code from any GPL-compatible licensed work you wish. You could sell the software and support for the software, even sell warranties for your work. Your competitors will not be able to distribute derivatives of your software without using the GPL for their releases, thus allowing you to share their distributed improvements.
You have misunderstood what copyright law allows for. This issue has little to do with the GNU GPL and LGPL per se, more to do with how copyright law deals with output of processes and copyright power regarding derivative works. Like any other software license, the GNU GPL and LGPL inherit many things from copyright law, one of them being that they cannot control the licensing of the output of a process (there are exceptions but I'm trying to keep this concise and simple). So, even though gcc is GPL'd, you can compile non-GPL'd software with gcc.
Linking, however, is a different issue. Linking constitutes a derivative work which does fall under the aegis of copyright law. Therefore the license for your derivative work is not completely under your control. Different licenses offer differing levels of permissiveness for creating and distributing derivative works. For example, if you were to distribute your work linked against the LGPL'd GNU C library you would have to follow what the LGPL says for licensing your derivative work. The LGPL does not make you use the LGPL as the license for your derivative work, but it has some say in what the license for your derivative work must allow.
I suggest you read a lot of licenses very carefully and weigh your options. Keep in mind that nobody owes you the ability to make or distribute derivative works. If you are unwilling to agree to the terms of someone's license, write your own code.
Looking at the overall gist of your post, it's clear you don't understand a good deal about Free Software or why the GNU GPL says what it says. You should read the essays at the GNU project's philosophy page and learn about the Free Software movement.
All computer users. The Free Software movement's message of software freedom is aimed at all computer users.
The Free Software movement is not just interested in popularity. The primary motivation of the GNU project is to give users freedom. Also, talking about developing software "for free" suggests you are talking about software in terms of price. The Free Software movement focuses on freedom, not price.
The GNU GPL is a very (if not the most) widely used Free Software license. Lots of people find it works for code sharing very well. However, it is not the job of the Free Software movement to look out for business interests in the way you want them to (treating businesses as charities is the work of another movement). If businesses want software developed, they can work hard and develop it themselves or hire someone to write it for them. If they want to share and share alike, they are welcome to get a head start and develop a derivative of a GNU GPL-covered work. Many businesses have chosen the latter route; the GNU GPL does not "lock out" business users, unless you define "business" to exclude IBM, Microsoft, Apple, and a host of commercial GNU/Linux distributors (just to name a few) all of whom make and/or distribute GNU GPL-covered software.
The second sentence of that quote denies the validity of what you said in the first and third sentences of that same quote. Anytime someone distributes new BSD-covered code they are distributing Free Software, and that's great. Unfortunately they've chosen a license that allows for proprietary derivatives (even incompatible proprietary derivatives—a practice commonly referred to as "embrace and extend"). The community cannot freely use the changes made to these derivatives and the developers of the new BSD-licensed work have essentially treated some organization as a charity, not insisted on developing software as equals.
Copyright law doesn't allow the FSF to stop you from using gcc to compile non-free software. For years, NeXT (now Apple) used gcc to compile NeXTSTEP and OPENSTEP (which, contrary to whatever you may read into the name, is a proprietary operating system). It was a completely seperate issue that NeXT tried to extend gcc without abiding by the GNU GPL by proposing to distribute their Objective C front end as .o files.
This doesn't respond to the point made earlier and it doesn't acknowledge the power of patents. The point made earlier was that with the BSD licenses, you can always benefit from someone else's contribution. So still being able to use the code without the patented contribution in no way addresses that point because you are talking about the program without discussing the patented contribution. Also, what you gain with copyright (i.e., "still hav[ing] the original code") you can lose with a patent. Copyright law and patent law are completely separate areas of law that can conflict. A patent holder (whether it is the one that "contributed" to your BSD-covered program or someone else) can write their patent so broadly it encompasses your extant work, thus preventing you from distributing your work, forcing you into a costly licensing fee, compelling a court appearance, or cross-licensing your patents with them (the real value of software patents is in cross-licensing).
The code could have been sent to you to make it known to you, informing you of its existance in a way where they can win a copyright infringement suit (see Lawrence Rosen's column in the December 2001 "Linux Journal" magazine for details on this) as well as a patent infringement suit. Whatever the reason, they have a patent on that improvement and you probably don't. Patents are written broadly enough to encompass an idea (not just a particular expression of that idea), so it's entirely possible (even likely) you won't be able to reimplement the improvement without infringing on the patent.
There's no doubt they acted within the limits of the BSD licenses, that was never the question. The question was whether you want that to happen to your program: Do you want to maintain the ability to share improvements to your program even in the face of software patents? Do you want the other users of your software to be able to share and modify the patented improvements? If so, the BSD licenses are not an appropriate license choice for your program.
The FSF has an essay directly answering your question: Why you should not use the Library GPL for your next library?. The GNU LGPL was renamed after this essay was written to stand for the Lesser General Public License to reflect that it does less to ensure software freedom.
Such a license would go against the whole point in using the GNU GPL. Free Software developers have to make advantages for one another and produce software that is Free Software. Linking with non-free software is antithetical to the Free Software movement because this movement maintains everyone should have the four fundamental freedoms of Free Software for all of their software. Distributing software linked with non-free software prevents people from having complete software freedom (more specifically, studying, modifying, and distributing source code).
Some might say the same thing about computer programming or using a computer recreationally. Fortunately we don't need to be experts to understand a great deal about copyright, we need to study it, ask questions, have discussions, and learn. There are so many interesting intersections between computing, economics, law, and politics, I sincerely doubt a reasonable person could find only "smallish" interests after reading some good books.
Understanding a system is the first step to changing it. Try to understand what's going on with copyright law by reading non-technical books written for a general audience like Lawrence Lessig's "The Future of Ideas" or Siva Vaidhyanathan's "Copywrongs and Copyrights". Both of them contain plenty of commonly accessible examples on where the law and computing intersect. Then try reading the references they point to, such as US Supreme Court cases and important appellate court decisions. I've found that the more I persist by reading and learning, the more I come to understand the ins and outs of copyright law and the more I appreciate the effort.
Yes, I am distributing the GPL-covered program to these people. They can request source code from me. If I am unprepared to satisfy that request it is my responsibility to not distribute the program to them in the first place.
Nobody said Fred couldn't do (I), the question is what fulfills his obligation to distribute source to the modified browser.
To whittle down the answers on that particular question, keep in mind what constitutes distributing complete source code to Fred's binary. Obviously Joan's source code doesn't have Fred's modifications, so I is out. II, III, and IV might look good until you notice that III only distributes patches to Joan's original source. That's not good enough in the event that you want complete source code to Fred's derivative and the source code Fred generated his patches against is inaccessible, so III is out. That leaves II and IV, which are equivalent and good (putting aside the difference betwen "alongside" and "along with"). [I know the real standard here is what the GPL says, not what parallel scenario I can dream up to justify the GPL's clauses, but I'm trying to write something accessible to those who aren't so familiar with the GPL.]
No, that's not what it says at all. Whether you distribute the original source code is up to you. The GPL only requires you distribute the source code to the binaries you're distributing (if you're distributing binaries). Recipients of the binary should be able to modify it, and that requires the source code that corresponds to their binary (not some other version, no matter how improved it may be).
Question #3 is aimed at making sure people don't try and sneak around the GPL by linking in code that is not available under the same permissions as the GPL. It's important to make sure people can share and modify the whole GPL'd program, not just part of it. The GPL allows Fred to charge for distributing Joan's browser or any derivative of Joan's browser. So linking in code that can't be distributed for a fee is linking in GPL-incompatible code.
Therefore the acceptable answers are quickly narrowed to the "Yes" answers (1 and 2): the sample license violates the GPL. Of those two, answer 2 is correct because the section 6 of the GNU GPL states that Fred cannot apply additional restrictions to any part of the code. Answer 1 is incorrect because the GPL only requires distributing binaries linked against GPL-compatible code and the GPL is not the only GPL-compatible license.
You heard correctly, writing your own license is very difficult. It is not to be taken lightly. The GNU GPL has seen real court time and been confirmed as an enforceable and binding license. I think most people would be hard pressed to create a license that can survive court examination for the entire license.
Before you rush off in a huff to write a license you'll most likely regret, consider that your aggravation stems from your misunderstanding of what the GNU GPL and GNU LGPL say. Since you got 3 questions wrong, you shouldn't be so quick to judge RMS, the GNU GPL & LGPL licenses, or the FSF harshly. I suggest jotting down your objections, learning what the Free Software movement is after, how the movement RMS started goes about achieving their wishes, learning what obstacles stand in the way, and then see if your objections have merit.
You sound like you're talking about Microsoft, the chief objector to the GNU GPL, and their FUD speeches attacking the GNU GPL (in a similar name-calling, no-specifics-required, mud-slinging way). But for all their flaming, they're not doing the one thing you'd think they would do given their position: avoid distributing GNU GPL-covered works.
Actually, when it comes to distributing binaries without source code and you want to go with the written offer option, section 3b of the GNU GPL says the applicable cost is "no more than your cost of physically performing source distribution". So you don't have to charge anything if you don't want to.
But on a larger issue, your rejection is simultaneously unclear and doesn't invalidate anything you quoted from me. I'm fully aware of the requirement of offering source code through the mail when someone orders it (I never contested this requirement). You read something into my words that was not there. I was not trying to present a complete list of alternatives (the GNU GPL itself is fine for that).
Consider your suspicions allayed. One question speaks to your suspicions. Distributing two archives as you describe is fine. Distributing source code in the same archive as the binaries is also fine. Network distribution makes it possible for the binary archive to be accessible where the source archive is not (this can happen via FTP, not just via web sites). This presents a problem and is easily solved by distributing source code and binaries together in the same archive.
When the GNU GPL was written not everyone had network access that allowed them to download all the source code they want to GPL'd programs. People still don't have that level of access. Networks do go down, some people pay per byte and have to ration their usage, and a variety of other limitations. The GPL has to account for people passing copies of GPL-covered software via other means too. I don't know about you but I have manually passed out copies of GPL-covered software to people (I always put the complete machine-readable source code on the same medium as the binaries if I'm distributing binaries). These people deserve source code too.
If you're only distributing source code, you would want to distribute it in a medium people could use otherwise people would not want your source code. If you're distributing binaries, section 3a specifies "a medium customarily used for software interchange", section 3b specifies a written offer (valid for at least 3 years) that offers source distribution on "a medium customarily used for software interchange", and section 3c talks about passing on a copy of the information you received in accordance with section 3b.
It appears you are trying to send a message to the editors of the quiz. You should know that I did not write the quiz.
Question 4 sets the scene here, "FooCorp distributes a modified version LibIdo library linked to their proprietary program Frobber.". The GNU LGPL notes in the preamble, "When a program is linked with a library, whether statically or using a shared library, the combination of the two is legally speaking a combined work, a derivative of the original library." and thus subject to the terms of the license of the work being linked in (in that quiz scenario, the LGPL).
Section 6 of the GNU LGPL requires the license under which you distribute a "work that uses the Library" to allow for "modification of the work for the customer's own use and reverse engineering for debugging such modifications". So it would not matter whether Frobber links in the modified LibIdo statically or dynamically, either way Frobber is a derivative of the modified LibIdo and that LibIdo is a derivative of Peter's library. Therefore, in the end, the proper answer to the question #8 is "yes".
I was thinking just the opposite: When was it reasonable to be ignorant of the laws that control what you can do with a computer? When was it a good idea to not know about the terms of widely-used Free Software licenses? Maybe the computer afficionados you know would be better off knowing what copyright law says and what penalties lie in wait for them if they violate someone's copyright. Being ignorant of the DMCA is particularly unwise in the USA, just ask Dmitry Sklyarov about the pleasures of being held in jail away from home.
Funny. I think the FSF put that up with the mindset of helping people to better understand the GNU GPL and GNU LGPL, and to increase the number of volunteers helping them with license issues (as they ask for if you answer all the questions correctly). Also, the FSF doesn't do "open-source" they do Free Software. There's a big difference.
These legalisms are just another part of life to learn about, organize protests around, and work within. A mind will not stay "fertile" if it is ignorant of the rules of society. It is shameful, not helpful, to remain ignorant of the law.
So long as Fred and Joan aren't a part of the same organization, distribution has occurred. The GNU GPL calls such an event "distribution" and that is the term I should have used.
Distributing source code and binaries from the same place is apropos because the GNU GPL says it is. See section 3 of the GNU GPL which includes:
In the quiz scenario, Fred distributes a CD with binaries of Joan's web browser to his unnamed friend. Giving out source code on the same CD as the binaries would have counted as distribution of the source code. Hence choice 2 ("He [Fred] can give out source on the same CD as the binaries.") is not the correct answer to the quiz question.
Someone could patent the code they contribute to your project and not license that patent to you. The BSD licenses do nothing to help you with software patents. The GNU GPL covers this situation.
I understand your desire for a simple permissive software license, but ensuring software freedom requires more complex licensing terms than what the BSD licenses provide.
The example is not silly at all. People do distribute CDs of binaries without source code. The FSF won't sue someone who uses the GNU GPL to license their software in cases where the FSF is not the copyright holder (which is the vast majority of GNU GPL-covered software). Even in cases where the FSF holds the copyright, usually these cases are not settled by lawsuit. Usually infringements are settled by someone pointing out the GPL's terms and asking for compliance (sometimes with the assistance of Eben Moglen, general counsel for the FSF, who said, "[a] quiet initial contact is usually sufficient to resolve the problem.").
Also, the problem doesn't come up merely by not distributing source code with binaries, but by not making it available at all. One can comply with the GPL in ways other than distributing source code with binaries. Distributing both source code and binaries simultaneously from the same location is probably the most convenient way to comply.
It makes sense when you consider that not everyone gives gifts to demonstrate how rich they are. Some people give gifts to show that they care about their friend's wishes. Sometimes this involves buying something, sometimes it involves spending time making something that has no price attached to it. Giving someone a copy of Debian GNU/Linux who can not get it on their own is one way to help them have software freedom.
No, the GPL does not need amending here. People need to be able to excercise all of their software freedoms including the ones that require source code: studying how the program works, adapting it to their needs, improving the program, and releasing their improvements to the community (known as freedoms #1 and #3 in the definition of Free Software). In order to come as close as one can to guaranteeing source code availability and offer reasonable practical alternatives, the GNU GPL includes the terms discussed in the first quiz question.
Why can't you ship source code and the binaries from the same location and get on with your life? One way to comply with the GNU GPL is to make source code available from the same location as the binaries. Websites become unavailable for various reasons and if people have only the binary and a pointer to the source code on a website, it's possible they would not be able to obtain the source code that compiles to the binary they have. This is not unheard of when people distribute binaries via CD. But my experience is most people don't find it difficult to ship source code and binaries from the same location.
No, it doesn't (and your point is not well served by pointing out how the letter starts off by correcting the use of that term). The Open Source movement is primarily targetting businesses. The Free Software movement speaks to all computer users. The Peruvian bill speaks to all Peruvian computer users, not just businesses. This is one way in which the wording in the letter is correct.
The bill is concerned with user's freedoms. The Open Source movement takes pains to avoid talking about software freedom because they believe their development methodology message will go over easier with businesses if they stop the freedom talk (in their FAQ they refer to freedom talk as "ideological tub-thumping" with no further analysis of the Free Software movement's motives). This is another way in which the wording in the letter is correct.
I see now where your misunderstanding lies: There is no obligation to redistribute programs in the Free Software movement. The Free Software movement is geared around freedoms you need in "the system of free expression in a technological society" (from the FSF's Amicus brief for Eldred v. Ashcroft). Nowhere in the Free Software definition is there a requirement to distribute Free Software. Quite to the contrary:
Software licenses requiring publication or notification are non-free (such as the APSL, the Apple Public Source License, which requires any modified version of APSL-covered software "deployed" in an organization must be published).
Exactly—users must have the right to do so. That's freedom talk which is totally synonymous with the Free Software movement, not the Open Source movement.
I think any reasonable person has to conclude that Dr. Villanueva knew what he was talking about and he meant what he said throughout the letter, including his repeated support for Free Software.
The problem with "free" is a problem in English, not other languages where the difference between price and liberty are clear. The FSF offers a list of translations to use for Free Software when speaking other languages so people understand you're not talking about price (yes, I know the link points to a term on "Freeware", that's where they put it because the list answers another question relating to the term "Freeware").
The fuss regarding the phrase "open source" has to do with using the name of the wrong movement to talk about what Dr. Villanueva refers to in his letter. Early on in the letter he corrects the Microsoft rep's use of the term, telling Microsoft he is speaking of Free Software, not Open Source. There is a big difference between the two movements.
Please make the effort to use the term Free Software (not "open source") when referring to the movement referred to in the letter. The bill and the letter all refer to only Free Software. The Open Source movement stands for something quite different.