I trust the FSF, but I have no idea who the FSF will be in another 20 years.
I trust the FSF, but I have no idea who the FSF will be in another 20 years.
> The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns.
> Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation.
(The GPLv3 goes on to support some stuff about naming a "proxy" to approve new versions in addition to the FSF.)
I'm no lawyer, but I imagine that would prevent glaringly-evil GPLv4 from being upgradable-to.
The part about being able to link AGPLv3 code, I'm not sure what is supposed to emphasis. Care to expand?
I used to think the AGPL was an EULA based on an interpretation biased by how people think the AGPL works. Upon looking into this more closely, I've realized this is most likely not the correct interpretation.
Instead, my current opinion is that the AGPL is just a terrible copyright license that is almost impossible to comply with (most AGPLed projects without copyright assignment violate their own license - if you think this doesn't make sense, what I mean is every contributor violates every other contributor's copyright), while not actually protecting against proprietary SaaSification in any legally enforceable way. I could go into more details, but honestly I've told this story one time too many, so I'll just encourage people to read clause 13, word for word, carefully, and think about what it actually means, not what the FSF has convinced everyone it means.
If a copyright license is the right or wrong place to address restrictions outside copyright law is an interesting discussion to have, but "similar in spirit" seems to match pretty well. I for one do not care if the legal restrictions sit under regular copyright law, software patent law, or Digital Millennium Copyright Act subsection DRM. They are pretty much similar in spirit.
I don't have a big problem with the and newer term. Let's imagine three possible scenarios. Note: Copyright assignment is different from an "and newer" clause and it's not the subject of this comment.
1. The new license is more restrictive, e.g. the FSF went full evil, GPL became an "All Rights Reserved" proprietary license. Outcome: everyone will just keep using the current version without an upgrade.
2. The new license is more pessimistic, e.g. the FSF went full pessimistic, GPL became a BSD or a public domain license. Commentary: Doing so will not create direct harms to the existing users. Since the software remains free, I'll simply keep using it. And speaking as a developer, in the very unlikely scenario that the FSF decided to end copyleft, chances are, I'd likely to agree. However, a dogmatic copyleft advocate may point out that it's a potential vulnerability.
3. The new license is similar, e.g. GPLv2 to GPLv3. This is the intended outcome.
There are some reasonable objections, but I don't see it as a big threat, so I have no problem with it.
But if an evil version of the GPL comes out and someone forks your code and relicenses to it, you can no longer merge their changes.
The spirit of GPLv2 is "I give you this source code, if you give it, or a modified version of it, to anyone, you have to (perhaps also) give it to them in a form they can actually use.".
The history of GCC includes a dark period in which version control wasn't public, and the project was dumping release tarballs on the GNU ftp server. Then, Cygnus Solutions started the Experimental GNU Compiler Suite (EGCS), a fork with a much more open development model, capable of attracting contributions from the whole Linux and embedded industry. Within a few years, many Linux distros had switched to EGCS, including Red Hat.
If I remember correctly, talks of reconciliation started after Red Hat acquired Cygnus: other distros were worried that their competitor had gained too much control over a critical component of the OS. Eventually, the GNU developers and the EGCS contributors agreed to merge back under the GNU flag, but keeping EGCS's open development model.
Nowadays, the GCC project is still hosted on the Red Hat-owned Sourceware servers, and its governance body is a technical board of contributors with various industry affiliations: https://gcc.gnu.org/steering.html
TL;DR: the FSF and the GNU project no longer have much influence on GCC's direction.
There wasn't a working Linux libstdc++ from the FSF that anyone could build until EGCS fixed it; no Linux distros that supported C++ ever used an FSF version. They had to use weird hacks built by HJ Lu.
please no license fights.
Think of what would have happened if some distros took alternate tooling paths? What about popular tools? We have to have both toolchains on our systems.
It is clear from how things turned out that EGCS wasn't about ownership or splitting the ecosystem. If one were to do this now, you would constantly rebase against upstream while you sort out the politics.
Clang/LLVM was both about implementation architecture and licensing, but had GCC been more modular, it might not have existed as well.
I am quite certain that when the generations that made this happen are gone, computer world will be back at business as usual like it was in the shareware days.
Also, there are diehard Libre GNU/Linux out there beside Stallman behind several projects.
They aren't. Google's play store and Apple's app store are examples of thriving shareware platforms.
Also, there is almost no shareware.
According to the definition [0] of shareware:
Shareware is a type of proprietary software which is initially shared by the owner for trial use at little or no cost with usually limited functionality or incomplete documentation but which can be upgraded upon payment.
This includes subscriptions and in-app purchases.
Nothing new really, unless you want to be pendantic on the real meaning of "open source" as defined by Open Source Initiative.
At the end of the day it won't matter, because those developers have bills to pay so eventually idealism gets replaced with real life.
And "who the FSF will be in another 20 years" is down to its statutes. Presumably the board has a duty to follow those, including (s)electing successors for themselves who also will do so.
The FSF has proven itself to be a cult incapable of moving on from or existing without RMS, so the simple fact is that, like all cults, it will die when he does. Luckily, the FSF does not embody Free Software, and the greater movement doesn't need either to survive.