The guy is one of the few Legends in IT
Love or hate him, he has been a consistent driving force. Without RMS, the software industry would be a lot poorer.
GNU fans have celebrated their software movement's thirtieth birthday - a movement that started as rebellious bits and bytes of tools, and is now a worldwide phenomenon. Today, servers, PCs, mobile phones, tablets, and all manner of devices run operating systems and applications that owe their genesis to the idea of software …
Love or hate him, he has been a consistent driving force. Without RMS, the software industry would be a lot poorer.
Seriously, have you seen what he looks like?
My perspective is that I will never, ever take computing guidance from a right wing nutter who slags off dead people and who cannot even work out how to wash himself, cut his hair or make himself look presentable.
Yeaah. 'ause what RMS looks like is THE important bit in all of this.
Sure. Glad we got that sorted.
yeah screw that hippie. Look at that cloth wearing Gandhi guy, look how silly he looks non violent resistance must be stupid because he looks silly.
You are going to have a pretty limited life if you judge people purely by looks alone. Good luck with that.
Meanwhile the freshly shaven guy in the suit is repossessing your grandma's house and pouring money into lobbying to ensure she doesn't get free healthcare. But it's ok, at least he washed properly.
Stallman's certainly been very important and influential in the field, for a long time. He's certainly well-known (which isn't always true of the influential). Sure, call him a "legend" if you like. But "one of the few"? I dare say with a bit of work I could come up with a list of a hundred figures who are at least as important in the history of - take your pick - the IT industry, the place of computing in modern culture, computer science and other technical aspects of computing, etc.
And the article is somewhere between rather and wildly overstated. There was open-source software before the FSF and the GNU project. There's a great deal of open-source software outside it. Clarke (in usual form) makes no effort to substantiate his claims that "all manner of devices run operating systems and applications that owe their genesis to the idea of software freedom", much less to Stallman's particular version of it. It's entirely possible that we'd have the Linux kernel, say, had GNU never happened, and the economics of cheap hardware (driven mostly by gaming and other entertainment demand) would have ensured user-land components followed.
Before I continue with my somewhat more critical approach I do want to express my admiration for the whole project. Because there is no denying -what so ever- that the project as is is utterly impressive.
Heck, I've lived the day where you had to pay Sun a lot of money in order to get a hold of a C compiler in order to build software on Solaris. An operating system you also had to license before you could use it. In all fairness: the same applies of course for Windows, Visual Studio also wasn't as easily obtained as it is now.
And here came GCC along the road... Even usable on Solaris!
Still, I also think that the GPL isn't providing the amount of freedom which it could have. Here's not saying it's not providing any kind of freedom, surely not, but it still tells the user what they can and cannot do with the code. For example using bits and pieces of a program and then using that in another program but using a different license, that's a no no.
Personally I'm more in favour of the CDDL which also sees to it that the original author retains his rights to his work, but also provides others with the right to use that work to make something of their own, even re-license it if they want to. As long as the original work remains licensed under the CDDL.
When I give something away for free then all I care about is being credited for my work, but I don't feel comfortable at all by telling the users what they can or cannot do with my software. Because I simply don't see the real freedom in that.
"Heck, I've lived the day where you had to pay Sun a lot of money in order to get a hold of a C compiler in order to build software on Solaris."
I remember when Sun unbundled the ancient K&R C compiler. It was crap, but served to bootstrap GCC (leave GCC building overnight, come back next day & fix errors, repeat until GCC working).
I suspect that event was critical in the rise of Linux. Sun had been the epicentre of innovation in the 1980s, but when Solaris donned a suit and tried to ape Microsoft, hordes of geeks got p***ed off and found a new platform.
I too remember the unbundling of C. I seem to recall, but age rots the brain, however, I seem to recall that the unbundling was forced upon Sun and others by AT and T or one of those wonderful, American pseudo-legal anti-monopoly bodies, to give buyers more choice in the same way as, later MS was required to become more open to non-IE web browsers, possibly in response to the nagging of such as Stallman. It was a real pain for those of us used to what was a perfectly good compiler (good enough to build the operating system, X etc. at any rate) who now had to pay or get GNU CC and persuade management that "free" software was reliable and safe and spend unhappy hours getting hold of it, adjusting make files, making it compile, checking compatibility with system libraries and so on.
One forgets too that the likes of Sun, Pyramid and others were seen as a breath of fresh air and portability. BSD all ready existed, being the basis for most releases of UNIX and being pretty open, actually. So put Stallmann in perspective (still find Emacs to be a baroque mess that is just too heavyweight and requiring too much customising to be immediately useful, unlike vi). Even his GCC comes with a thousand options that can vary with the platform and, as for "long options" with "--". I see the point and avoid them at all costs, most software providing long and short versions. Eh? Clever chap; bit head up his own arse though.
The good thing is, GNU released the makers of the Linux kernel and, to a large extent, OSX, from the costly work of implementing the shell and its tools. Yes, OSX shell is mainly BSD and GNU.
Zealotry not withstanding, at one time, most proprietary systems came with source code and one could spend endless hours in internals courses for customers. It was said that Chinese students learnt UNIX, in the absence of sufficient computers, by studying the source code, before GNU and long before Linux or even Minix (probably the inspiration for Linux).
and 5 years later implement it completely.
I fully expect Replicant to fail - but that's because I expect to be running Linux on my phone and tablet rather than waiting for someone to come up with a cut down version of an app for Android, when the full app can often be modified for touch screen usage in a thousandth of the time it takes convert part of it to android.
And I can always run up Android for leisure time.
Whether the device runs an ASOP or GNU/Linux based operating system (both of which require the Linux kernel) you have the same problem that Replicant aims to solve - that these devices rely upon numerous proprietary binary blob drivers that aren't updated or supported by their creators and cannot be legally distributed without the accompanying device.
> that these devices rely upon numerous proprietary binary blob drivers that aren't updated or supported by their creators
A lot of this is graphics, and that nut is being cracked open right now. Right now it pretty much only the community, but slowly the phone manufacturers will get that they should push out the docs, or just do a full Intel and do open drivers themselves. It just makes things easier, for them as much as anyone.
Also, sad though it is, getting GNU/Linux working on Android blobs has been getting a lot of love too. Though I'm really not sure that is a good thing, though I guess its better than being stuck with Android.
I think the only stupidity here is calling his efforts stupid.
Sure, you may not agree with his opinion or his preferences (for example see my ideas about the GPL vs. for example the CDDL) but that doesn't make them stupid. At the very end he and his associates can say that they truly know how an Android environment works and operates, and that's something not many people can say.
Who cares if they re-invented the wheel by doing so? That's pure geekdom for you (IMO): you do it because you can. Who cares about the rest?
Heck; even if this project does fail it's still not stupid. At the very end it's a valiant effort.
Nothing personal here; but you do realize that if people took your approach when Stallman started sharing his visions about a free Unix environment we'd never have come this far?
Wrong -- http://redmine.replicant.us/projects/replicant/wiki/GalaxyS3:
Known freedom issues:
* Proprietary and likely signed bootloader
* Modem, WiFi, Bluetooth and NFC firmwares are non-free
* Modem could access GPS
And a huge omission:
* Image signal processing subsystem.
This last one being a really large lump of IP that differentiates the good phone vendors from the rest.
I can tell you, having had access to the full documentation (under NDA) of a particular SoC used in phones, you still don't get any documentation on the ISP or GPU.
I hit what I thought was the read comments button expecting to be treated to a plethora of sage words, not realising that there weren't any. Whilst I'm here I might as well make a comment.
I'm with Stallman most of the way. Open source software is a good thing. It helps drive innovation. However, I am happy with companies such as Apple and MS not releasing all the source code to their operating systems. I tend to look on devices (Apple in particular) as hardware and OS combined. As long as I have got the APIs then I can produce code for these systems.
I shall now retire a safe distance and whatch the fireworks begin.
No fireworks - I agree with you. From a pragmatic perspective you'll always have a mix of components, and Stallman is a bit too hardcore. Having said that, it is exactly because Stallman was funded to remain hardcore that the concept hasn't been compromised into oblivion - where he is right is that compromises can represent the thin edge of the wedge (see the current issues with W3C allowing DRM into the specs).
Like him or just respect him, we certainly need a few who don't compromise.
> However, I am happy with companies such as Apple and MS not releasing all the source code to their operating systems.
I'm not happy with that since there's no reasonable way to check for backdoors without the source and who knows what the NSA may have demanded be put in there. The whole tool chain, including compiler has to be auditable and so does the hardware. Sadly, we're far from it.
The whole tool chain, including compiler has to be auditable and so does the hardware.
Sure. Let us know when you finish auditing your CPU.
> Sure. Let us know when you finish auditing your CPU.
Do you really not get it though? If such design was out in the open and GPLed or under BSD license we'd automatically have all sorts of activity on this from people with various backgrounds, companies, universities, private individuals, including well-intentioned security researchers, students etc. who'd have a chance to e.g. check that Intel's random number generator hasn't been fiddled with (http://www.theregister.co.uk/2013/09/10/torvalds_on_rrrand_nsa_gchq/).
While that wouldn't prove that the processors we have in our PCs is exactly as per said design we might have the chance of purchasing from a non-American entity that we feel is more likely not to be after our data.
You cannot build a commercial project without access to vast libraries of open standard (if not source) tools and toolsets like programming languages. Not cheaply or easily anyway.
And you cannot justify developing expensive complicated applications if no one pays for them.
But once developed they are ripe to be ripped off. Frankly 15 years ago we would have paid a fortune for a site license for - say - libre office, or a current linux distro.
Once the nature of the problem and the solution are well understood - well its simply me-too engineering.
And that's why patents have a limited life. So that once they run out, the world gets to benefit.
We should be more relaxed. The fact is the two are not in competition - free and closed source. They are different parts of an ongoing cycle.
Now for that open source shim that allows windows programs to access an API/libray set that means they can run natively in linux.. :-)
"You cannot build a commercial project without access to vast libraries of open standard (if not source) tools and toolsets like programming languages."
Hmm. For many, many years people did. For that matter, MS and others still do, depending upon what you class as "cheap" or "expensive". It depends upon what you count as cost. Money? Time? Effort? If the free stuff costs you time that should be going on your actual business, that can be a high cost. For most people, it may be an absolute barrier. Even time spent searching online for bug fixes, installation hints, usage advice, updates is all lost, expensive time, before even implementing what one found, if you found it.
Actually, IOS app writers are writing "commercial" projects, though it can be argued that the underlying system is very cheap, being derived from BSD. It depends upon how one defines "commercial". Even Stallmann can be thought of as "commercial", in that he is making sufficient profit to pay him enough to travel, to buy computers, live somewhere .... He must be making a profit, bit of an industry actually. GNU has distribution and control systems, documenters, programmers, hosts, probably lawyers and copyright experts. These may be a mixture of paid, sponsored, volunteer etc.; but nevertheless, it is a form of organisation that does not run at a loss. Is he any less commercial than a self employed accountant or plumber or builder or one-man PC support business?
Actually, so does GNU, though not "commercial". But most of its programmes do assume a working, underlying system with system libraries, even if, later, it replaces some of those with its new versions. It has to bootstrap from somewhere.
I think we will end up with lots of different Linux distros for phones. Once the piece start falling into place, they will be reused and respan by other distros. ARM is slowly getting more standardized, open ARM graphics drivers is happening, ophono is still alive and well. The phone will become the pocket PC (x86 if Intel have their way). It will happen, it might just not be quick.
Ubuntu phone might be the first, but later down the line, I expect a lot of distros will have a phone spin. The crazies can have their Slackware/Gentoo/Arch phone and hopefully I'll finally get a Debian phone (chroot on Android isn't enough!).
I was with you until this comment, "The crazies can have their Slackware/Gentoo/Arch phone...". That's when you got the DV.
Mines the one with the Slackware install media in the pocket...
your link doesn't work, and the video is worth watching!
There's my khaki jacket; I'm taking the pith helmet too.
Yep, you are definitely taking the pith.
Really, in RHODESIA ? how quaint, how's the empire going, old chap ?
I joined the Replicant project this summer as a result of the publicity. Prior to publicity I wasn't aware it existed, despite having been reverse-engineering and hacking code of Android devices since 2010.
It quickly became apparent - to me at least - that Replicant is not going to achieve anything until it changes its focus onto up-to-date devices and engages developer attention as Cyanogenmod does. With the rate of change in the mobile arena an 18-month old (and getting rapidly older) device isn't going to attract developer attention and certainly not users - who wants a 2 year old device which lacks key functionality including, sometimes, the ability to make and receive calls? May as well carry a brick around.
The problem for Replicant is that most if not all AOSP-based devices depend on binary blob proprietary drivers for key hardware interface functionality - video output is the obvious one, but also things like sound, radio interface, and GNNS/GPS. Cyanogenmod does this and suffers much pain figuring out how to get older binary blobs to work with newer, revised, AOSP ABI/APIs.
I decided to focus on GPS blobs by beginning to reverse-engineer the Nokia MEIF binary GNSS/GPS protocol used in Broadcom chip-sets that more devices are using so that support can be added to gpsd and other location-awareness daemons and stacks. This work is useful to the general F/OSS community rather than the handful - literally - of Replicant users.
My belief is that we'd do better overall to have developer teams focus on the particular functionality in binary blobs that prevents users from being the masters of their own devices, rather than try to maintain a fork of an entire OS.
"Ubuntu, while under GPL, has been politicised by its maker Canonical’s decision to follow its own path at the risk of community peace, putting in increasingly commercial features and a roadmap that’s alienated sections of the open-source faithful and those concerned about their web privacy."
There seems to be a question here, does a company take open source and make commercial extensions that are then kept proprietary eg. Apple with it's OS/X extensions to BSD, or do they release the commercial extensions under GPL and let the community decide whether they wish develop to them further or not.
Personally, as long as Canonical release the source to the "more commercial features" they are building into Ubuntu under GPL, I don't really see a problem, as Stallman has been very clear open source is free but not as in beer, ie. it can be a commercial product.
I think there are a few things to remember. You can make Linux as commercial as you like as long as you stick to the rules of the GPL, that is, you reveal the alterations (if any) to the kernel code. The second and more important thing to remember, is the difference between kernel space (the Linux kernel) and user space (anything outside the kernel). So if you want to hide your code or use something else but GPL you do it in user space. And that is quite common in the embedded "space". Third there are many different licensees to chose from, and there is also GPL2 and GPL3. Nothing wrong with you comment.
The world has changed since 1983. Today, the vast majority of users aren't computing experts. Everything is connected to the wild internet. We manage our bank accounts and our credit cards online, while bad guys try to steal our login credentials. Dodgy apps can be installed with couple of clicks ("yes, I grant key-logging permissions to FreeCandyCrush4U"). It's impossible for end users to know who to trust; so unsurprisingly they default to trusting only apps approved by Apple / Google / Amazon.
Solve that trust issue, and you might make inroads into freer software.
The only effective solution to this problem is education. Using a complex and potentially dangerous system such as the internet will always be troublesome for those who remain ignorant of its workings. Apple products are a perfect example of how the "walled garden" serves only the company that controls the walls. Dumbing things down to the Apple level just helps a corporation create dependant customers, people who sadly miss out on some amazing things that can be done by anyone given a bit of guidance.
'how the "walled garden" serves only the company that controls the walls.'
1. OSX is, in no way, a "walled garden". If you think so, you have not used it and do not understand UNIX.
2. IOS is controlled. It is a system for consumer devices, much like the programme in your car's or washing machine's electronics or a Nokia mobile. But, you can choose from any of many thousands of programmes to install. You can write one yourself; as with Linux kernel patches, you just need to adhere to certain standards to preserve the integrity of the device and other users.
3. For the vast majority of users, the approach serves them brilliantly: it provides a level of consistency, simplicity, easy maintainability and reasonable security. App developers can earn money too, using the implicit guarantees provided by the system. Indeed, stock Android devices do the same, just not so successfully because they lack the muscle with the ISPs and Google failed to implement "app stores" properly, in good time from the start. In fact, one could argue that Android or Windows mobiles are more restricted: each supplier inflicts their own idea of a good shell interface on the buyer (including Google) and, unless you want to go down the "jail breaking" route, you live with that, with all the differences and inconsistencies resulting (pro: "freedom of 'choice'" by buying a different model, or even IOS). Worse, for most people, no matter how bad the security or bugs (is broken security not a bug?) there is little prospect of an upgrade until they buy a new mobile. How is that for a walled garden? Buy another? Great.
4. Those who fancy themselves as technical and knowing what they are doing can do what they like by jail-breaking IOS or Android (and probably Windows).
5. Have you tried asking Samsung or HTC for all of their source code and the build tools to compile and install it yourself?
Anon because of all the automatic brickbats I shall receive.
>>> But Firefox’s Linux kernel, called Gonk, is licensed under an Apache and the Mozilla Public Licence (MPL) <<<
Surely the kernel itself still has to be under GPL.
Are you talking about some specific modules or user space extensions?
The article isn't differentiating between the GNU Hurd kernel and the GNU Toolkit (the Unix-compatible command-line programs and gcc compiler). Together they comprise GNU. The distinction between the two is important. Hurd itself - which promised much, and delivered little and far too late - is now more-or-less rendered irrelevant by the upstart Linux, but the GNU Toolkit runs on Linux and many other systems.
And a little historical point: the article says: "By 1991, the GNU team had written enough useful stuff to encourage volunteer developers to port the GNU sources to Linux". This is not incorrect, but, in fairness to Stallman and the FSF, the GNU Toolkit was mature enough as far back as 1986 that it was a complete and usable product. At Acorn, I did the world's first ARM port of the GNU Toolkit, onto the BSD Unix kernel. It wasn't exactly trivial, not helped by some truly dumb coding techniques* in parts of the Toolkit, but the result was eminently useful. Fortunately, the code quality in the GNU Toolkit has improved a lot since then!
(* For C fans: using chars as negative indices from the ends of buffers in EMACS didn't work out so well when ANSI allowed the signedness of chars to be a compiler writer's choice.)
Joining two sentences together:
"Stallman wants us using Replicant, the 2010 fork of the Android source isn't exactly going anywhere"
Replicant isn't even akin to the <1% market share "Linux Desktop". And Stallman wants us to use it. He doesn't want us to use a usable system, just free (as in beer) software. Meanwhile in the real world where time is money, I prefer things to work.
The open source movement fails frequently to understand that software has non tangibles, such as design and usability. It's rare that the OS even stops to think that their users might not be particularly like them, or even care. Instead there's a lot of naval gazing.
We note that the company that has probably sold 100 times more UNIX systems than anybody else - and to regular consumers - is really big on things such as UXD and design - and leverages vast amounts of Open Source software licensed under non-GPL licenses. They've had a large part in rendering thoroughly obsolete, the FSF's first software release: GCC.
Once upon a time, we had Maemo. Well, OK, we've had a number of freer-than-Android candidates, but Maemo is the one that was backed by a major vendor, and the one I bought into. I think it was Jan.2010, Maemo was absolutely the biggest buzz at FOSDEM, just before Nokia abandoned its dev community and embarked on the path to oblivion.
The rise of Linux was different: its circumstances were that a former dev-friendly platform (SunOS) had donned a suit and turned corporate (Solaris). In a world where Android remains dev-friendly, and software portability is the norm (so many of us don't target any particular platform), RMS's vision may face a higher hurdle to gaining critical mass.
Have an upvote from a happy ex Maemo/now Meego user... Shame indeed it was axed, but as you say, Android presents a very different competitor than commercial Unix used to.
Maemo served a a great distraction to get Micro$oft scared and deploy their trojan horse Elop. While they were busy killing Maemo in the but, Android had room to become the ultimate threat to the empire.
Still using it here! Although I need to upgrade to the CSSU edition..
"label libertarian - a label that, perhaps for most Brits, conjures a right-leaning American stereotype who owns lots of guns, lives in Nebraska and inhabits the Tea Party fringes of the Republican Party. Basically, someone who likes to look others square in the eye and say: "You aren’t the boss of me.""
An American Libertarian is someone who wants to enjoy the benefits of services like national defence, firehouses and police, but doesn't want to pay the necessary taxes to provide them...
They also have difficulty looking you square in the eye, since theirs are swivelling too much.
No, libertarians are freedom obsessed zanies -- people like Jefferson, John Adams, Franklin, Madison, Mason, Herbert Spencer, and H.K. Mencken provide good examples of libertarianism before it was called that.
The basic idea is that some government is necessary but government's tendency to grow for its own benefit at the expense of others means that government needs to be kept in check.
Take a good hard look at government anywhere and sensible people some something in it.
Big deal, an American Republican is one who only wants a national defense and bible beatings when the NSA catches you having sex wrong and an American Democrat is one who wants to enjoy the benefits produced and paid for by everyone else while the NSA quells rumblings of dissent.
while the NSA quells rumblings of dissent
That's not fair. The NSA are not a bunch of jack-booted thugs. They're spies who inform on you to other agencies, and the latter provide the jack-booted thuggery.
GNU was never turned into an actual OS due to delays on the GNU Hurd kernel. enter a young student called Linus who just built a Minix-inspired kernel and glued all up using the GNU tools stack. Ta-da! Of course, this meant Linux took the limelight originally intended for GNU, thus the dick-wavering eternal fights over how the OS should be called. GNU/Linux, as "GNU Linux" implies Linux is part of GNU which it isn't. RMS insists that you should even actually say "gnoo slash leenuks" when referring to the full name.
That said, I agree with other commenters; RMS is needed as the radical element he is, to keep balance on the force and avoid awful stuff like the DRM'd HTML5 we've just been announced. Yeech!