The GPL is a Joke

[Cogman]

http://www.jfplayhouse.com/2009/08/gpl-ideology-is-joke.html

Kind of interesting. I do agree with some of his points, the whole "If we don't have the source and can't distribute your source, then your company is evil." attitude is really the bane of the GPL community.

 

[degibson]

Extremes have their value. Often as bad examples, I will admit.

 

[Red Squirrel]

While I prefer open source / free software to proprietary, I also believe it's the software author's decision and I respect that. If the author wants to make it closed source, then so be it.

 

[Markbnj]

Stallman has been a nutbag for quite awhile now. I don't think he started that way, but I am not completely sure.

There is no particular reason for software to be free, or source code openly available. The idea that it had some inherent property that made it more like "the pursuit of happiness" than "a screwdriver" was always whimsical at best. Since people, including programmers, have to eat, I think over time you can expect software to be charged for, somehow.

 

[DaveSimmons]

The GPL makes perfect sense if you want to create something, make it available for modification, and make sure that your work isn't then closed off for profit by some other company.

This can be a barrier to use by people who are trying to make a living off their extensions, but the base is your work so it's fair for you to set the rules.

Other than that, yes Stallman seems like a nutbar zealot who thinks he knows the One True Way software must be created and shared.

 

[postmortemIA]

GPL mission is not to be everything for everybody, just to provide developers excellent means to make open and free software, and of course on benefits of users who get to use it

truth is, most companies use lots of stuff from GPL efforts. Take for instance GNU compiler collection, it is being used from microwaves to rockets.

 

[Cogman]

GPL mission is not to be everything for everybody, just to provide developers excellent means to make open and free software, and of course on benefits of users who get to use it

truth is, most companies use lots of stuff from GPL efforts. Take for instance GNU compiler collection, it is being used from microwaves to rockets.

I'm reminded of this http://xkcd.com/198/

You'd have to be an idiot to say that the GPL never works, or has never worked. There are cases where it is the best solution for the problem.

The problem is, the GPL and GNU movement has the attitude that it is ALWAYS the best solution for ANY problem. Further, anything that doesn't take that stance is not free software, it is evil.

That is the main issue I take with the GPL.

 

[Nothinman]

http://www.jfplayhouse.com/2009/08/gpl-ideology-is-joke.html

Kind of interesting. I do agree with some of his points, the whole "If we don't have the source and can't distribute your source, then your company is evil." attitude is really the bane of the GPL community.

I didn't get to read the whole thing but if it's yet another article bitching about Stallman, we don't need those either. Everyone knows he's batshit insane, there's nothing new to discuss about it.

The GPLv2 is a simple license that does what it was designed to do. Any zealotry around it is irrelevant.

 

[Modelworks]

My main complaint with GPL is that it doesn't always lead to getting useful information. Right now I am working with a lot of embedded hardware in commercially sold devices that just happen to be running linux. They have released their source as required under the GPL. The problem is it doesn't include any information on how to use that code with the hardware because the hardware is under NDA . So hobby programmers get the source , but there isn't much you can do with it.

 

[degibson]

My main complaint with GPL is that it doesn't always lead to getting useful information. Right now I am working with a lot of embedded hardware in commercially sold devices that just happen to be running linux. They have released their source as required under the GPL. The problem is it doesn't include any information on how to use that code with the hardware because the hardware is under NDA . So hobby programmers get the source , but there isn't much you can do with it.

This is a pretty classic way around the GPL. "Open" Solaris does something similar -- everything that is GPL is GPL, but there are a bunch of closed-source 'encumbered' binaries, too.

 

[Colt45]

Companies are entirely free to not use GPL'd code on their microwave, if having to provide the source is too difficult for them.

 

[Fallen Kell]

Also remember it is perfectly fine if you are using modified version of GPL'ed code for internal use only. It is only when you start selling a product which uses the code that you need to also make the source code available to anyone. And again, you can still have proprietary code as well as GPL'ed code on the same system, as long as you do not hard link the proprietary code into the GPL'ed code.

 

[Sureshot324]

My main complaint with GPL is that it doesn't always lead to getting useful information. Right now I am working with a lot of embedded hardware in commercially sold devices that just happen to be running linux. They have released their source as required under the GPL. The problem is it doesn't include any information on how to use that code with the hardware because the hardware is under NDA . So hobby programmers get the source , but there isn't much you can do with it.

This is a pretty classic way around the GPL. "Open" Solaris does something similar -- everything that is GPL is GPL, but there are a bunch of closed-source 'encumbered' binaries, too.

What are the rules of the GPL regarding this? If you make an extension to a GPL project but don't modify the source, just access the GPL project through it's API, does your project not have to be open source? What constitutes a derivative work?

 

[degibson]

What are the rules of the GPL regarding this? If you make an extension to a GPL project but don't modify the source, just access the GPL project through it's API, does your project not have to be open source? What constitutes a derivative work?

The subtleties of the GPL are beyond me, sorry.

 

[Cogman]

The subtleties of the GPL are beyond me, sorry.

It depends on the GPL version.

LGPL v2 - You can use static libraries with your code.
LGPL v3 - You can use dynamic libraries
GPL v2 - You can't use static libraries, Dynamic are fine
GPL v3 - Dynamic links are off limits.

Now, the dynamic library thing is still a bit of a gray area. So even though GPLv3 says that your not supposed to use them, I think there is currently some debate on whether that is legal to say or not.

Fair use would say that using API calls to another library would be legal as you aren't including the entire copy righted material, and your aren't copying major portions of the project into your own.

Stallman would say that you should burn in hell if you don't release your projects source (Actually, He'd probably say that even if you weren't linking to a GPLed library).

 

[chronodekar]

It depends on the GPL version.

LGPL v2 - You can use static libraries with your code.
LGPL v3 - You can use dynamic libraries
GPL v2 - You can't use static libraries, Dynamic are fine
GPL v3 - Dynamic links are off limits.

Is this accurate ? I've always wondered what the difference between LGPL and GPL was. The way I understood things, the v3 makes it difficult/impossible for TiVo (or the like) to use that code. (Perhaps, I'm simplifying it too much, so forgive me if I got things wrong)

-chronodekar

 

[Cogman]

Is this accurate ? I've always wondered what the difference between LGPL and GPL was. The way I understood things, the v3 makes it difficult/impossible for TiVo (or the like) to use that code. (Perhaps, I'm simplifying it too much, so forgive me if I got things wrong)

-chronodekar

AFAIK this it is accurate. There are some other subtilities, but I believe that these are the main differences.

GPL v3 is very restrictive (on its description of what gets GPLed), and is why a good number of open source projects haven't moved to it.

 

[postmortemIA]

I don't think it is. For example, you can't run any binary that uses Qt SDK without Qt dynamic libraries. And lot of proprietary stuff is being developed with Qt framework. Which is LGPL v2.

 

[Nothinman]

The subtleties of the GPL are beyond me, sorry.

It depends on the definition of a derivative work which I don't believe has been defined in court. I believe the general thought by technical people is that any kind of linking of a non-GPL'd project to GPL'd code is forbidden, hence the creation of the LGPL.

 

[degibson]

It depends on the definition of a derivative work which I don't believe has been defined in court. I believe the general thought by technical people is that any kind of linking of a non-GPL'd project to GPL'd code is forbidden, hence the creation of the LGPL.

FWIW, my general impression folks in industry (and by that, I mean the microprocessor industry) is that they tend to abhor the GPL, won't touch anything GPL'd. I'm sure thats due to a mix of true legal terror and simple paranoia. BSD is much more industry-friendly.

GPL seems like its better adopted in the systems community, probably because of the widespread reliance on Linux. Of course, these are only my impressions -- it would be interesting if there were some statistics available out there on industrial use of (L)GPL.

 

[Nothinman]

FWIW, my general impression folks in industry (and by that, I mean the microprocessor industry) is that they tend to abhor the GPL, won't touch anything GPL'd. I'm sure thats due to a mix of true legal terror and simple paranoia. BSD is much more industry-friendly.

GPL seems like its better adopted in the systems community, probably because of the widespread reliance on Linux. Of course, these are only my impressions -- it would be interesting if there were some statistics available out there on industrial use of (L)GPL.

It's probably just paranoia because I can't see how the GPL would affect a microprocessor at all, except maybe simulators. And then the libraries they link against are still their choice. Most of the core system libraries on Linux are LGPL and the GPL-2 itself has an exception for core system components, unless you redistribute them with your product.

BSD is more "friendly" in that it has no restrictions, unless you use the BSD license that comes with the advertising clause.

The GPL and the BSD aren't really comparable because the BSD license is really just 1 step above public domain. While the GPL's whole intention is to keep your code in the open if someone decides to use your code in their project.

 

[degibson]

It's probably just paranoia because I can't see how the GPL would affect a microprocessor at all, except maybe simulators. And then the libraries they link against are still their choice. Most of the core system libraries on Linux are LGPL and the GPL-2 itself has an exception for core system components, unless you redistribute them with your product.

I was indeed referring to simulators (though HDL can be GPL'd too). Industry sometimes wants to share portions of their simulators with certain people, but GPL tends to make that fairly hairy (or does it downright forbid limited sharing? I forget).

 

[Nothinman]

I was indeed referring to simulators (though HDL can be GPL'd too). Industry sometimes wants to share portions of their simulators with certain people, but GPL tends to make that fairly hairy (or does it downright forbid limited sharing? I forget).

The GPL only requires sharing the code with whomever you share the binaries with, but you can't stop them from sharing with everyone else which is the problem for some people/companies.

But there's nothing inherent with Linux that makes you GPL your code, it depends on which libraries you use and chances are those libraries are the same ones you'd use on a BSD licensed system.

 

[n0cmonkey]

I didn't get to read the whole thing but if it's yet another article bitching about Stallman, we don't need those either. Everyone knows he's batshit insane, there's nothing new to discuss about it.

I always stop reading when these people bring up HURD. There's no point in continuing. First, there are plenty of open source/free software projects that are succeeding, and second I can't think of a single OS/kernel that _is_ finished.

 

[Nothinman]

I always stop reading when these people bring up HURD. There's no point in continuing. First, there are plenty of open source/free software projects that are succeeding, and second I can't think of a single OS/kernel that _is_ finished.

No piece of software is ever finished in the typical sense of the word. But Hurd is special in that it's been in development for ~20 years now and still isn't anywhere near as usable as Linux or any BSD.