I can't see how he can defend crippling software, and blocking usecases which clearly a large set of users want (all kinds of editor intelligent colouring, competition), out of fear of a closed-source boogyman which doesn't exist any more, thanks to clang (if it ever did).
It most certainly did. Read here[1] about the history of Objective-C. NeXT tried to keep as much of it as possible proprietary.
Sadly, because LLVM is so good but doesn't have copyleft protection, we live in a world where something new and interesting like Swift may never be non-proprietary.
[1] https://en.wikipedia.org/wiki/Objective-C#Popularization_thr...
Should Apple not have a right to make their own proprietary language?
Let's not forget that Apple is a big part of why LLVM is as well known as it is.
Languages can't be proprietary, anybody can re-implement some language if there is a spec. Whether you can name it the same or not is another issue but non-proprietary languages usually don't survive for very long so it would be in Apples own interest that swift ends up in as many hands as possible.
Oracle still doesn't understand this completely about Java, maybe Apple will never understand it about Swift, meanwhile all those non-proprietary languages will flourish.
This is wrong thinking and it misunderstands Apple's goals and motives for Swift.
First and foremost, it's a replacement for Objective-C, which is itself effectively proprietary - there are open source compilers and some legacy openstep implementations, but the driving force behind the language has been exclusively Apple for at least a decade now. In that context, Swift has the same expected level of success as Objective-C, owing to the community that exists around its predecessor. And, for Apple, the important metrics for success with Swift are entirely focused around the effectiveness of the language in support their platform. Given the relative popularity of the platform, they can reasonably expect that people will use the language to gain access to Apple's userbase, because it's a valuable commodity.
Second, Apple doesn't care if Swift has a long life, at least, not per se. They want it to be used and to be successful because it's where they're probably moving their platform in the medium to long term. But they're not invested in the longevity of the language so much as they're invested in their platform. If, in another decade, they decide that Swift has outlived its usefulness, they'll replace it and let it die. Any other scenario for the language failing entails Apple faltering as a company, in which case, they don't really care about the Swift's success.
(I snipped a "non-" off proprietary, because it didn't seem like what you were arguing from the rest of the post.)
Swift is currently in the same position that C# was in before Mono. It will probably be popular in certain circles, but it's not going to ever dominate the language field like Java, Javascript, and C do. Microsoft's recent freeing of the .NET infrastructure makes it a real, bona fied choice now. Apple should follow suit early, for the good of the language.
I'm not sure the cross-platform framework problem is one that helps them or makes much money compared to what is status quo. But at least their proprietary world got better for those of us who write for it.
In RMS's world, hackers close ranks to make sure the power stays in the hands of the people, not the corporations. Sure, corporations might develop a bit of proprietary code now and then, but they are outpaced by armies of hackers writing GPLed code and replicating any innovations they develop. All the best compiler toolchains are GPLed in this world, and corporations are basically forced to participate in the Free Software movement, or suckle at its teat. It would not be possible in practice to sell software, because any software your computer could run would have to be GPLed!
A lot of other stuff happened instead. Now we have hackers working at corporations writing open-source code and sharing it with other corporations and the world. Companies don't sell software so much as provide various human and computer services. Corporations aren't evil, they can provide funding and organizational structure that is actually pretty valuable. And so on.
In the world RMS wants, people would shun proprietary software due to it not giving them the rights which are afforded to them through copyleft.
This is what he is advocating, that people should not accept proprietary software due to it not respecting their rights, not that proprietary software should be illegal.
Proprietary software terms should be illegal.
But before that happened something else happened (which is more important, and you forgot about it, it seems) and that is GCC, Linux and other free software. Linux runs on tiniest embedded devices, Android phones, military hardware, servers, and so on. It was possible in large by GCC and GNU.
Have you been around in the early 90s and 80s. Having high grade OSes or compilers that were free as in beer or speech would have sounded ridiculous. You'd sound like you were from Mars if you talked about such things.
Yet here we are. We have reliable, high grade operating systems based on ideas of fruits of free software movement -- CentOS, Ubuntu, Deb etc. Because of GPL in large anyone can take advantage and reap the benefits of shared work (such as by using CentOS instead of RHEL) while letting Redhat employees still put bread on the table.
Show me Apple's OS X code that was modified for NetBSD and FreeBSD projects? Where are those improvements? I haven't seen them.
So yes now this "service" and "cloud" stuff is happening. But let not forget that for this to happen the even the idea of "open source" and "free software" as viable alternative had to take root. I believe if it were not for GPL and Stallman we wouldn't have had this culture of sharing between large corportations.
"In the world RMS wants, Apple would not have the right to develop Swift on top of a GPL-ed compiler platform and make it proprietary."
That's not incredible at all, and makes perfect sense.
I've never seen the the Free Software movement as "hackers" against "corporations". Places like Cygnus (bought by Red Hat) made a mint selling Free Software consulting services. Nobody ever had a problem with that. They thrived because they were good and they solved real problems for people. None of that is antithetical to either the Free Software movement or corporations.
I am glad that you are more clear about this position of his than he is himself. Personally, I would never have guessed it.
The FSF position is and always has been that the distribution of non-free/proprietary software works against the "freedom of computer users". That last word is important, because the FSF concerns itself with the rights of end users, not programmers or software companies. That's the big philosophical divide between "free software" and "open source": the 'free' side makes an ethical/consumer protection argument concerning end users, while the 'open' side makes a pragmatic argument mostly concerning developers (cathedral vs. bazaar, etc.).
EDIT: My interpretation of Stallman's politics is that he would happily support legislation to forbid nonfree software, but I think the FSF intentionally avoids the subject for the sake of their libertarian supporters.
[1]: https://gnu.org/philosophy/government-free-software.html [2]: https://gnu.org/philosophy/free-software-even-more-important...
He was also, a while after the start (like 1985), very harsh with "Software Hoarders" like myself who then worked on non-copyleft software.
Except, of course, the process would be very political, and he doesn't play that game at all well.
Should Apple's users not have the right to use, share, study, improve, extend and share improvements and extensions to the languages and tools they use?
Or, in answer to your question: no.
Eh, what ? Apple is already shipping their proprietary fork of Clang which supports their proprietary language Swift, this is exactly what Free Software which Stallman advocates is working against.
The whole idea of Free Software is that the end user will have the full rights to examine, modify, copy, run said software, proprietary plugins in GCC flies right in the face of that, so obviously Stallman is opposed to them.
Thats exactly the point. Anybody wanting to make a proprietary language implementation, or proprietary IDE is already using llvm anyway, and would remain doing so. New projects of that nature would almost certainly still use llvm. The only people really being affected by this are people who want to use GNU tooling, but are left with deliberately crippled tools.
The ability to develop plugins for GCC has existed for years now, but nobody has used it to do the evil things RMS insisted would happen (remember, he blocked the creation of the plugin interface for many years). The main reason is that Clang/LLVM is just much easier and nicer to work with (e.g. you don't have to write and maintain your own plugin to export an AST instead of having it be part of the core project). Hence all the brightest young minds in compiler research and language development are using LLVM in their work instead of GCC and an entire generation that could have been using and contributing to GNU software are working to improve 'the enemy'.
Its hard to think of a more valuable group of people to drive away. Or a more thorough way to drive them away than to try to put intellectual shackles on them. Stallman's suggestion of only exporting a gimped AST that only allows a few specific transformations gives people involved in languages and tooling the same reaction as RMS feels when told he can't see the code for the software on his printer.
>The whole idea of Free Software is that the end user will have the full rights to examine, modify, copy, run said software, proprietary plugins in GCC flies right in the face of that, so obviously Stallman is opposed to them.
Nobody is proposing creating proprietary plugins for GCC. People are proposing a GPL licensed plugin to export the AST for use by emacs. No proprietary code involved at all. RMS has said he would block the use of such a plugin in emacs, thus deliberately crippling emacs, which is why emacs is probably going to get forked (again).
>The ability to develop plugins for GCC has existed for years now, but nobody has used it to do the evil things RMS insisted would happen (remember, he blocked the creation of the plugin interface for many years).
First off the plugin api is unstable which discourages out-of-tree patches, secondly there's the gcc runtime library exception which only allows distribution of proprietary binaries created with gcc if said binaries was not generated using proprietary plugin(s).
>Nobody is proposing creating proprietary plugins for GCC. People are proposing a GPL licensed plugin to export the AST for use by emacs.
But allowing for that will also make it possible to use a proprietary backend with GCC, thus allowing the circumventing of rights which copyleft licensed software is there to ensure. It doesn't matter if you think it's 'unlikely to happen', again copyleft exists to ensure that this does not happen.
This is one of those, frankly, pretty rare cases where the users actually feel a limitation of the GPL (mostly because people that use Emacs has a significant overlap with programmers where the line between developer and end user is very blurry).
So adding proprietary license on top of software which one has not authored is a right that belongs to programmers?
Actually it is pretty irrelevant how many non-programmers are capable of compiling or meaningfully editing the C program. The point is that a programmer cannot (reasonably) edit to restrict you freedom.
However, after rethinking what I wrote, I should clarify that while the underlying principles are the same regarding what RMS is trying to do here (limit programmers to protect users), the GPL has nothing really to with this particular fiasco.