Free Software Foundation files suit against Cisco for GPL violations [LWN.net]

The GPL is designed to encourage sharing

Posted Dec 11, 2008 20:56 UTC (Thu) by bignose (subscriber, #40) [Link]

> when the subject matter is esoteric enough, the GPLed library is sometimes the only good implementation, making the use restriction rather frustrating.

That's pretty much the point. If you want to benefit from the freedom you receive in a GPLed implementation, then you don't get to restrict the freedom of recipients of your derived work.

Frustrating the efforts of those who want to restrict software recipients's freedoms is a big part of the design of the GPL. It sounds like the GPL is working as designed.

History of LGPL

Posted Dec 11, 2008 21:47 UTC (Thu) by dmarti (subscriber, #11625) [Link]

LGPL was always for standard or clone libraries that made it possible to build UN*X applications on GNU. For libraries that gave you something new, the GPL Even the GNU Readline library was released under GPL (which resulted in BSD Editline as a clone under a non-reciprocal license).

Free Software Foundation files suit against Cisco for GPL violations

Posted Dec 11, 2008 20:57 UTC (Thu) by daney (subscriber, #24551) [Link]

No one is forcing you to use GPLed code. If you don't like to license, too bad. There are plenty of sources for non-GPL code, use one of them instead or write it yourself.

Free Software Foundation files suit against Cisco for GPL violations

Posted Dec 11, 2008 21:06 UTC (Thu) by atai (subscriber, #10977) [Link]

There is always the BSD out there, right?

The ultimate goal

Posted Dec 11, 2008 21:28 UTC (Thu) by whacker (guest, #55546) [Link]

I think it is important to get people to understand that knowledge is not a material object, it is something that is most useful only when it is shared.

Thats the ultimate goal. The GPL is a tool that helps us get there.

The ultimate goal

Posted Dec 12, 2008 0:16 UTC (Fri) by felixfix (subscriber, #242) [Link]

Knowledge is only useful when shared?

OK, share with us your bank account numbers, id numbers, name, address, mother's maiden name, passwords ...

But I suppose you meant knowledge that others have generated for themselves, not dry facts.

So, share with us your business's strategic plan, upcoming products, market analysis ...

Not what you meant? You meant just software maybe?

I think you probably have a lot of limitations on just what knowledge you meant that is most useful only when shared. Please amplify.

Most useful

Posted Dec 12, 2008 1:39 UTC (Fri) by ncm (subscriber, #165) [Link]

Criticizing based on a misquote is pretty cheap argumentation.

We distinguish a fact as knowledge, vs. raw data or ephemeral plans, by how repeatedly and generally useful it is. Artificially restricting knowledge harms the body public by interfering with this repeated use. Causing such harm is legal, generally, so must be specifically discouraged when we want to prevent it. The GPL is one way to do that. The patent right grant is another way. Tenure is a third.

The ultimate goal

Posted Dec 12, 2008 8:29 UTC (Fri) by job (subscriber, #670) [Link]

That's very different types of information. Passwords are supposed to be a secret. They are not supposed to be "useful" to others in any meaningful way.

The other things you mention are publicly available. Plus every time I do consulting work I put them on paper in order to get paid. That's pretty useful to me.

The ultimate goal

Posted Dec 13, 2008 1:01 UTC (Sat) by man_ls (subscriber, #15091) [Link]

Indeed, passwords are only useful when shared: the owner has to share them with the authenticator to make them work. If whacker shared with you his/her passwords then they would be useful for you to steal his identity (or whatever they protect).

The quote "knowledge is most useful only when it is shared" is surprisingly accurate, even beyond software where it was supposed to be applied.

Free Software Foundation files suit against Cisco for GPL violations

Posted Dec 11, 2008 21:47 UTC (Thu) by gmaxwell (subscriber, #30048) [Link]

"when the subject matter is esoteric enough, the GPLed library is sometimes the only good implementation" thats exactly when the GPL makes sense for a library.

For those who are in the business of promoting free software (such as the FSF) the lesser GPL is primarly useful when GPLing a library would only encourage gratuitous incompatibility and not encourage the GPLing of more software. ... Such as is the case when the functionality the library provides is widely duplicated.

(If you don't agree with the goal, you won't like this reasoning but this is not an unreasonable position)

FSF "Why you shouldn't use the Library GPL for your next library"

Posted Dec 11, 2008 23:10 UTC (Thu) by dwheeler (guest, #1216) [Link]

It's not a "de-emphasis" on the LGPL. The FSF specifically recommends that Free software developers use the GPL and not the LGPL, for libraries, as described in "Why you shouldn't use the Library GPL for your next library". If you think about their goals, it makes sense why they'd recommend that. Not everyone agrees with their goals, of course.

Having said that, the FSF is more pragmatic than some give them credit for. In many cases, even if you completely support the FSF's goal, the GPL is not the best license for a library, and the FSF admits this. They developed and maintain the LGPL specifically because they'd rather get half a loaf than none. Heck, Stallman encouraged the Ogg Vorbis library's move to permissive (BSD-style) licensing, because he perceives the codec patent threats to be a bigger threat than any copyright threat. Projects like GNOME _specifically_ require that all their "core" libraries to be LGPL-compatible.

Free Software Foundation files suit against Cisco for GPL violations

Posted Dec 12, 2008 0:11 UTC (Fri) by dlang (✭ supporter ✭, #313) [Link]

that's assuming that the GPL really does apply across the linking boundary. this is something that the FSF preaches is the case, but has never been tested.

I don't know if it's people not testing this case, or if the people going after violators concentrate on the blatant cases where no code is released at all.

Yes it's the case

Posted Dec 12, 2008 1:24 UTC (Fri) by gmaxwell (subscriber, #30048) [Link]

If I offer you software under a license that states "This license is only valid so long as no software you distribute contains the word 'bob'" and you package up my software with yours, and your software contains the word 'bob' then you are distributing the software I wrote without a license. The manner in which your and my software is or isn't combined isn't relevant, you still are obligated to conform to the license of my software or not distribute it.

So there really is no reason to doubt that the GPL can define the linking boundary however it likes. GPLv3 switched away from loaded terminology like 'derivative' specifically to avoid encouraging this class of confusion among laymen.

Yes it's the case

Posted Dec 12, 2008 1:40 UTC (Fri) by dlang (✭ supporter ✭, #313) [Link]

there are limits to what is considered reasonable for such licenses, and I don't believe that your example would be considered valid.

Yes it's the case

Posted Dec 12, 2008 6:55 UTC (Fri) by jamesh (subscriber, #1159) [Link]

If the license is not valid, what license would you have to distribute the code?

Sometimes court can declare PART of the license "null and void"

Posted Dec 12, 2008 17:04 UTC (Fri) by khim (subscriber, #9252) [Link]

For example if your license demands that you'll chop your hand before you'll start using the program court will usually say it's "unreasonable demand" and THIS PART of the license will be declared null and void (but the license as whole will still be valid). But in reality this rarely happen: usually some part of the license can be voided only when license contains requirements which are specifically forbidden by some law.

in which case you still don't have a license

Posted Dec 14, 2008 1:04 UTC (Sun) by pjm (subscriber, #2080) [Link]

Even if the GPL said you can only distribute this software if you do something illegal, you still wouldn't be able to distribute the software:

If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Program at all. (§7, GPLv2)

in which case you still don't have a license

Posted Dec 16, 2008 6:54 UTC (Tue) by k8to (subscriber, #15413) [Link]

Is that legal advice? You are a lawyer?

rhetorical questions

Posted Jan 9, 2009 0:50 UTC (Fri) by pjm (subscriber, #2080) [Link]

Sorry I forgot to add "IANAL" until after I'd already posted that comment (LWN doesn't have a comment edit facility).

I'm not sure what k8to means by the question "is that legal advice", but I believe the answer is fairly obvious for most meanings even without the IANAL disclaimer. I think points can be made more clearly without rhetorical questions such as the above; it actually comes across a bit sarcastic, which is not conducive to helpful discussion.

But that's not the point

Posted Dec 13, 2008 13:37 UTC (Sat) by dark (✭ supporter ✭, #8483) [Link]

The "or not distribute it" choice is the one available here. A program that, when run, links against a shared library that is already present on the user's system does not have to be distributed with that shared library. But the FSF claims that it is still covered by that library's license. That's exactly the point of dispute.

This tends not to matter for Linux distributions because they need to provide the library anyway, but it matters for independent software vendors.

Free Software Foundation files suit against Cisco for GPL violations

Posted Dec 12, 2008 20:50 UTC (Fri) by salimma (subscriber, #34460) [Link]

Even more bizarre would be Trolltech's use of GPL+exceptions. One of the exceptions for Qt is if it's linked to a BSD-licensed code -- if this BSD-licensed code is just a wrapper that is then linked to from proprietary code, they might as well have used LGPL from the start and avoid the exception mess.

That's pretty much my problem with FSF's over-promotion of GPL. The percentage of software out there that is effectively GPLed might be lower than one might think.