Python 1.6 Incompatible w/ GPL

WillSmith sent in a bit running over at LinuxToday addressing the fact that RMS thinks that Python 1.6 is incompatible w/ the GPL. The article is basically how to resolve this within Debian.

Comments

Re:The beginning of the end?

[jimhill]

RMS is most emphatically NOT an advocate of "open source". His cause is "free software" and there are many examples of software which meets the Open Source definition without being Free (the latest, apparently, being Python 1.6). Whether you agree with some of the stands he has taken in the past, honesty requires that you acknowledge he has never misled people about what his goals are and he has never cared what others thought of his methods.

The recent high-profile licensing flamewars have come about in part because of the success of RMS's cause. After many years trapped in universities and corporate basements, free software has become quite popular and its quality is recognized. ESR's "The Cathedral and the Bazaar" provided a stimulus in the form of demonstrating how users will improve software faster when they can provide more than just bug reports. As a result, the world has become fixed on the idea that releasing source means better software. There are two big problems with this success, though:

The first problem comes about through licensing schemes that provide source but little else to the user, the most egregious example I can think of being the SCSL abomination delivered by Sun. Licenses like these are little more than an attempt to get users to do your work for you while reserving all rights to that work to yourself. Stallman's attack on this particular license generated little flame because the Open Source crowd was also condemning it.

The second problem is a little more subtle and is best exemplified by the unending GPL vs. BSD bickering. A BSD-style license provides source and nearly unlimited rights to the user, but is little better than an SCSL in Stallman's view because it permits unscrupulous users to "take the code private", that is, to benefit from others' work without sharing their own. If you truly understand RMS's definition of free software as "I'll show you mine if you show me yours" then you will see how such a license is also unacceptable to Free Software advocates.

These second category licenses have enjoyed increasing popularity as an opportunity by developers to get their product in wider use. Many people misunderstand the requirements of a GPL-compatible license and think that anything they do using a GPL-comptaible product must be released -- don't laugh, I know one guy who refuses to write anything in Python because "then I'd have to give my code away." The question of people asking if they can release binaries compiled by GCC without having to release source demonstrates the prevalence of this miunderstanding. A quick way around that is to allow people to take the code private. Some like that, some don't. It's unfair to attack RMS for being just as insistent on free software now as he was before the great Open Source Revolution of the 1990s.

Why not dual license?

[tilly]

Perl has been doing this for many years with GPL and Artistic, easy to do, no possible problems.

You offer the code under a choice of licenses, refuse all patches that don't fit both licenses, and call it a day. If you wish to use Perl with some drivers or some packages, you may need to use the Artistic license. If you need to use it in others you may need to accept the GPL. Most people don't ever bother deciding...

Cheers,
Ben

GNU/RMS

[richnut]

GNU/I GNU/just GNU/Dont GNU/see GNU/what GNU/everyone's GNU/problem GNU/is GNU/with GNU/RMS. GNU/Maybe GNU/I'm GNU/just GNU/biased.

-GNU/Rich

We need to fix the "Opensource Definition"

[Mr T]

This topic has come up a number of times over the last few weeks and I think it has almost been treated comically. I don't see how we can devote the amount of attention to the theft (err, "sharing") of music and the various legal activities surrounding it and then treat our own copyrights with contempt. I've read so much RMS bashing that it makes me sick. RMS is doing us a favor! The GPL will be tested in court, that is eminent. Dotting the i's and crossing the t's is very important. It may sound like RMS and the FSF are on a power trip or something, which is probably true to some small extent, but it is also critical that the GPL be followed, it not be watered down by being misused, the misuses are known and publicized, and that we take care of it. Spiritually, we're (the royal we, as in GNOME, GNU, KDE, Linux, free software) are all on the same page but those nasty details need to be worked out. The fact that there haven't been any lawsuits within the community show that we're together in spirit.

Something else which I'll bring up that hasn't been is that this is about leverage. Don't confuse it with anything else. It is about positive leverage for change. As the GPLed code base grows, being able to use it becomes more and more desirable. To use it you need to contribute to it, that's the leverage. It's reaching a critical point now, there is a lot of good software that is GPLed. That is how and why RMS designed it. It's also a good thing, the leverage is controlled by us, there isn't an IBM or Microsoft on the other end, it's us. The community, the programmers and the users. It's not an RMS thing either, ESR brings it up at the end of Magic Cauldron and in his eyes it is much much more than that because he believes that opensource has already won and that in 3 to 5 years when just about everything is opensourced that leverage can be applied outside of software; applied to laws? social causes? all sorts of things, we're talking about wholesale social change! This isn't a communism or socialism thing, it's about restoring power to the people, the people own GPLed software and the people need to police it and protect it.

KDE not complying is damn important, it undermines everything that does. Python doesn't? Well it should and it will. There will be countless others, as the GPL's importance grows so will the amount of attention given to products which go against it and I'm betting that there is a pretty fair number of them.

I don't know if Bruce Perens or ESR are reading but I suggest that an ammendment be made to the opensource definition. There needs to be a level of license interop that is supported. We've already seen the NPL fail to comply with GPL, the QPL fail to comply with the GPL, the CNRI license fail to comply with the GPL, interoperability is critical. I want to use KDE code in GPLed GNOME code! I want to use GPLed GNOME code within KDE! I want to use mozilla code with both! If the interop can't be worked out then what is the point really? It seems like a lot of the ideas and theories about opensource go out the window if interop isn't part of the deal.

Grow up people!

[FascDot Killed My Pr]

First of all, licenses matter. To programmers they matter directly--which code can you reuse and who can reuse your code. To non-programmers they matter indirectly--which products get which features, when and how compatible.

Second of all, I seriously doubt RMS has a list of software vendors that he consults and then goes out to badger them into using the GPL (or a compatible license). More likely, Python went to him to see if they were GPL-compat. He said "no" which is entirely his choice. Furthermore, Python acknowledged that it was his choice by the very act of asking.

Third, this causes no problems to the FSF, the Python people or the general public. It's only a problem for groups like Debian that want to distribute A) only GPL and GPL-compatible software and B) Python.
--

The Straight Dope...

[Boulder Geek]

There is a FAQ on the new license. Excerpting (w/out permission, but I believe this to be fair use ;-):

7.Is it compatible with the GNU Public License (GPL)?

Legal counsel for both CNRI and BeOpen.com believe that it is fully compatible with the GPL. However, the Free Software Foundation attorney and Richard Stallman believe there may be one incompatibility, i.e., the CNRI License specifies a legal venue to interpret its License while the GPL is silent on the issue of jurisdiction. Resolution of this issue is being pursued.

To me it seems that RMS is really splitting hairs. Most contracts that I've seen specify a jurisdiction under which the contract is to be interpreted.