> there is no way around this truth

Of course there is because it's simply not true.

For example, the conditions in GPLv2 make it impossible to combine GPLv2-only software with Apache or with AGPL software. Thus, GPLv2 has a restriction that GPLv3 does not. As another example, GPLv2 has fewer methods for conveying source code than GPLv3, and thus it is more restrictive.

But beyond this nitpicking on the meaning of "restrictive", let us look at the real thing most people object to about GPLv3: the anti-tivoisation clause. Let us remember what it actually is about: along with the source code, you must convey any cryptographic keys necessary to allow the modified version to run. This is very similar in spirit to the same clause that requires you to convey installation scripts. In other words, it has always been the responsibility of downstream users to make sure that everyone can use the software.

It really is not a crazy restriction to ensure that people convey the necessary data in order to ensure the software can be installed. In fact, that's why GPLv3 calls all of this (install scripts, authorization keys and so forth) "Installation Information". GPLv2 already had this condition. GPLv3 merely added clarification of what "installation information" means. This clause also only applies if you are distributing the software together with the hardware ("User Product").

GPLv3 is more widely feared than understood. So, understand it, and help others understand it too.

Let's not forget the other contender for replacing Bitkeeper: Mercurial. We will also be celebrating its 10th year anniversary next week during the Pycon sprints.

Yes, because it's way too beta, almost alpha.

The last time that the NSA weakened an algorithm they recommended was by shortening the key for DES. Snowden confirms that properly implemented crypto still works, and Rijndael (AES) still seems strong. The problem aren't the algorithms, because the mathematics still check out. The thing to fear are the implementations. Any implementation for which we are not free to inspect its source is suspect.

There was recently a bit of a kerfuffle over RdRand.

Matt Mackall, kernel hacker and Mercurial lead dev, quit Linux development two years ago because Linus insulted him repeatedly. Linus called Matt a paranoid idiot because Matt would not allow RdRand into the kernel, because it was an Intel CPU instruction for random numbers that could not be audited. Linus thought Matt's paranoia was unwarranted and wanted RdRand due to improved performance. Recently Theodore T'so has undone most of the damage, but call RdRand still exist in Linux. I do not understand exactly if there are lingering issues or not.

That is an entirely different meaning of the term, but arguably, is the meaning in the original summary. Is there evidence of "open source" ocurring without the "intelligence" part?

The first OED citation for "open source" is from 1998:

open source adj. [first published, on the Internet on 8 February 1998, by E. S. Raymond in a revised version of his paper ‘The Cathedral and the Bazaar’; ‘[the term] was invented by Christine Peterson of the Foresight institute at a private meeting I ran a few days earlier’ (E. S. Raymond, private communication)] Computing (chiefly attrib.) designating software for which the original program files used to compile the applications are available to users to be modified and redistributed as they wish. 1998 InfoWorld 2 Mar. 75/4 ‘The popularity and success of Apache, the Linux operating system, the BSD version of Unix, and many other software applications prove the value and impact of open source development’, Linux creator Linus Torvalds said in Netscape's statement. 2003 Wired May 125/2 The chip king is hardly a stern taskmaster, giving only the most general directions to faculty. Much of the work is released open source, and wild tangents are encouraged.

If you have a citation that goes a centuries earlier than this, you should notify the OED about how this term has been in use for that long.

I didn't say that gcc is "closed", or non-free, as it were. There is no imperative to make sure free software is easy and convenient for all possible uses. For instance, I do not have a moral obligation to make sure the free software I create is easy to compile on Windows. Modularity is nice, but not an essential feature of free software. All that you really need is the source code, the preferred form of modification, and with that you can exercise the required freedoms. You can fork off gcc if you want. In fact, gcc is a fork, a fork called egcs that got branded back into gcc, as you well know.

Also, you seem to be suggesting that rms is a tyrant here for not allowing gcc to be modularised. Unlike a real tyrant, nobody has to actually listen to rms. If enough people wanted to do so, they could all abandon GNU, fork off gcc, do whatever they wanted with it, modularise it, and call it something else, yet again.

In this case, the danger of enabling non-free software on top of gcc overweighs the convenience of allowing modularity in gcc.

It is consistent with providing free software to make it difficult to create non-free software. The GPL's copyleft isn't non-free anymore than laws against me punching you in the face take away any significant freedom from me.

Oh, prepare to be lectured for saying "intellectual property"...

Aha! So this is why rms doesn't want to hear anyone talk about open sores!

GNU is supposed to be a free operating system as well as a group of people working towards building this OS. To a casual observer, however, GNU does not appear very active. Some of the most prominent and supposedly GNU packages, such as Gimp, Gnome, GTK+, and R are mostly GNU in name only. The hackers working on these projects have very little interaction with other hackers working on GNU projects and they very frequently espouse views contrary to GNU's philosophical aims. Thus to an outside observer, GNU does not appear to be a cohesive group of people working towards a common goal. Many GNU mailing lists being private further the public perception that GNU is not even actively producing software anymore.

What can be done to remedy this situation? How can we strengthen GNU, make it reach out again to the people it's supposed to be freeing?

Although GNU and the FSF's views are often thought to be exactly the same as yours, they are not. GNU and the FSF are many other people and although they overall have the same aims, individuals associated to each organisation may deviate slightly from your views.

The FSF right now is pretty indepenent from you. John Sullivan is actively leading it, but there are other very public members of the FSF. It has become independent from you, even if you're still the president of the FSF. Unlike its beginnings, the FSF is also no longer primarily concerned with creating free software, but rather it is now involved in campaigning for free software. Social activists mostly aligned with your views have replaced the hacker majority in the FSF.

GNU has no such clear independence. You have the final say on aything that happens in GNU, such as for example usinng bzr as a DVCS for Emacs, a choice of dubious tactical advantage that has generated much discontent. You have nevertheless vetoed any dissent on this topic. Your health is apparently deteriorating, and I hesitate to think what will become of GNU when you die.

Is there any clear path for the future governance of GNU you in the same way that the FSF has done this?

In addition to the terms of the GPL, You violate the AGPL when you (1) copy the source code and (2) modify it (3) host the modified version over the network and (4) don't provide source for your modifications. Since by default (1) is a copyright violation if there isn't an explicit permission to do so, if you say you don't agree to the AGPL, then you don't have permission to do (1) either. So if you do the above, you can't claim that you didn't agree to the terms of the AGPL without acknowledging copyright infringement.