Free Software Licensing Quiz

mpawlo writes: "How much do you know about free software licensing? Time to find out! In a quiz presented by the Free Software Foundation you can test your abilities. How should Joan license her web browser?"

Comments

Re:Being a "cheapskate" and software freedom.

[jbn-o]

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".

Re:Ugh.

[jbn-o]

What about question 2, why can't he do number I?

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.

Re:Lawyers

[jbn-o]

...[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.