Slashdot Mirror
RMS Replies to "The Stallman Factor"
Ryan Amos writes "RMS has replied to the article "The Stallman Factor," as
posted on Slashdot about a week ago. In specific, his replies deal with the University of Texas SIGLinux naming fiasco and Bitkeeper. As always with RMS, an interesting read."
He is talking about Linux, the kernel component of GNU/Linux. When talking only about the kernel, the correct name IS Linux, because thats what the kernel is called. The entire OS is the Linux kernel + a bunch of low and high level apps.. many of which the GNU project created. Hence the argument that the OS should be called GNU/Linux instead of just Linux.
because he was talking about the kernel - not the whole system.
And a prediction:
I drew up this list because I know I'm going to get annoyed at the RMS-bashing that will surely follow. Many of the bashers won't even bother to read the article, because it is long and requires some effort to follow. I present this summary so that people understand that it is not just about RMS seeking credit. He makes a cogent and logical distinction between his point of view and (eg) Linus's point of view, and gives an example of why he thinks his own principles are important. You don't need to agree with him, but simply insulting him is unacceptable if our community is to continue to move forward.
Reality is defined by the maddest person in the room
I was just about to argue the point, then realized that SuSE is a perfect example.
In SuSE, in particular, the kernel distributed in binary doesn't have a corresponding source distributed, or downloadable. I found this to be quite annoying. I couldn't download patches for the kernel I had. I would have had to patch the source I had, which was already a couple minor revisions behind the binary they distributed.
But also, remember, they don't need to distribute the source, only make it available. If they have a download available somewhere, then they are in compliance.
Lots of people have CONTRIBUTED to GNU applications. Where would emacs be today without Lucid's fork? Where would gcc be without cygnus? How many completed applications were donated to the GNU project...software that "GNU" had nothing whatsoever to do with? Look at CLISP. It is "part of GNU" but why? Originally they weren't. They used readline in a way they thought was compatible with the GPL. RMS came down on them and told them they were breaking the law and gave them the choice of licensing under the GPL or removing that part of their code.
What does the "GNU team" have to do with such "GNU software" as Sather, dylan, SmallEiffel, kawa, or zlib (which isn't even licensed under the GPL)?
Yet all of those things are listed as GNU software on the FSF website. Were the people who wrote that code consulted as to whether they wanted projects using their software called "GNU/Project"? Or is it the FSF that decided that? Then shouldn't it only be code that the FSF themselves have written that counts when making that decision?
My point is this: The GPL ALREADY gives "credit where credit is due." Do you think that any GPL'd code in the Linux kernel exists without the GPL headers, and the appropriate attributions? HELL no.
Why is RMS changing the rules? GPL'd code BY NATURE gives credit where credit is due. It does NOT require anything else, especially the silly prefix of "GNU/".
Nope. The difference is that *BSD is a complete distribution - kernel, libraries, posix command line tools. [although gcc and GNU emacs are used]
Linux is a kernel only, and [GNU/]linux systems use GNU libc and GNU posix command line tools.
Do you even lift?
These aren't the 'roids you're looking for.
This isn't a flame. Really. I'm not going to call you names or insult you.
But you are hugely off base here, and it is *exactly* people like you that Stallman is trying to help... what's the word... "educate" I guess is reasonably emotionally neutral.
(And before we start - no, not everything Stallman says is gospel. In particular, while his point that Linux would be nowhere without GNU and the FSF is completely valid, the whole "GNU/Linux" naming deal is just sour grapes and entirely asinine)
With that out of the way...
Back in the day, computer science was treated like a SCIENCE, in so far that every program was treated almost like a scientific discovery. You don't hord discovery in a truely scientific world; you share it.
The program you wrote may advance the art in some manner - it solves a problem that was previously unsolved, it provides a service that was previously unavailable (or was somehow suboptimal), and the code you used to do it with may introduce new techniques to solve similar problems.
In a world where all code is shared freely, the net effect that every new program increases the size of the solution space (or, if you prefer, the "utility space") of computing in general.
It is important to understand that for the longest time, the free sharing of code was the normal state of affairs. There were no secrets between coders.
You only have to look at UNIX to see this. UNIX-as-kernel is just another operating system, but start adding in sed, grep, awk, troff, perl, TeX, vi, (even - gasp - emacs) and so forth (each of which is a component that freely plugs into the others and usually designed by other than the kernel maintainers) and you start increasing the size of the solution space of the OS.
It's not enough to just build the tool and release the binaries. The tool may be in some way incomplete for someone, and they need the ability to modify it (and then contribute their modifications back to the larger community, further increasing the solution space)
The concept of "commercial software" (ie, "software-as-product") as embodied by Microsoft and Billy Gates, nearly erased this state of affairs. For the generations of coders raised in the "software is something you hord and sell" world, you've never seen the way things were in the "software is shared freely as part of the Art" world. This really is an alien concept to you, and it's not really your fault.
Luckily for the state of the Art, code produced and freely shared is superior in every way to code produced by those that hord-and-sell. For a while, the hord-and-sell people had the upper hand, but now that almost every computer is networked, network effects are taking over, and things are slowly starting to right themselves. In ten to fifteen years, software-for-sale (with perhaps the exception of games) will be nothing more than a historical asterix.
But I digress.
The bottom line is that, yes, we all DO have the right to your code - but that is not a very good way of phrasing the situation or thinking about it. Instead, YOU have the OBLIGATION to _release_ your code, to advance the state of the Art.
Stallman is not demanding you give up what is yours, he is reminding you of your obligation to the Art.
The absolute best modern example of this I can think of is the work done by John Carmack and the rest of id Software. John's code truely advances the Art. Every new program that springs off of John's computer brings with it new techniques, new optimizations, new ways of thinking about the problem space he has tackled. And he has released the source code to every program as soon as it is commercially viable to do so, thus fulfilling his obligations to the Art. It's a very good example to follow.
But the trap you cannot fall into is to assume you have the "right" to hord code. You have no moree "right" to keep your code locked up on your hard drive for eternity than you would to hord a cure for cancer, or a working GUT equation. It is essential that you (and people like you) start to realize their DUTY to share code - if you do not, then you will eventually be marginallized and discarded, to the net loss of all.
DG
Want to learn about race cars? Read my Book
No, not very arguably at all. You can choose not to install X and still have a useful system.
RMS's point is that, if we use our terms properly, the "Linux experience" means that you have a kernel. What you put around that kernel to make a functional system - i.e., the rest of the operating system - is GNU. Then on top of that you might install X, Mozilla, etcetera.
Linux (in the strict sense) is the kernel, not the whole OS. RMS's whole point is that GNU had been working for many years to create a free (as in speech) OS, had it all together except for the kernel (IMHO because they got too ambitious with the HURD), along came Linus and provided that kernel.
Tom Swiss | the infamous tms | my blog
You cannot wash away blood with blood
> The S in BSD stands for SYSTEM.
That's the weirdest spelling of "Standard" I've ever seen. (Berkeley Standard Distribution.)
rant
"Change your name before I come" is RMS' personal requirement when accepting a speaking engagement. Actually, other FSF speakers often speak to groups that call themselves "Linux" groups. We ask only that the advertising and press material about our particular speech call the system, GNU/Linux.
Of course, when I and other FSF speakers make a speech, one of the items on our agenda is to ask such groups, as a favor to the GNU project, to change their name and/or documents to say "GNU/Linux" consistently. While it is RMS' personal demand that the name change occur as a term to accept the engagement, the FSF does not, as an organization, demand such name changes. We simply request them.
Comparing it to Microsoft's tactics is out of proportion. FSF firmly stands for free speech rights. We assert your right to call the operating system anything you like; we request as a favor that you call it GNU/Linux.
RMS is a highly sought-after speaker. As it turns out, since he is not (nor never has been) paid a salary by the FSF, he collects speaker fees to help pay for his living expenses. As with any speaker, it's his prerogative to set the terms of his speaking engagements. Indeed, every speaker has his or her own set of requirements. (AAMOF, ESR's are available online.) Personally, I have a rule that there must be vegetarian restaurants that someone can take me to in the towns I visit. Of course, FSF doesn't take a position on vegetarianism, but it's a personal need of mine that I can't ignore---even when I am speaking on behalf of FSF.
While RMS won't come to speak for your group if it's called a "Linux" group, I'd be happy to come, as would many of the other FSF speakers. While I am there, I am, of course, going to ask you to change the name of the group. But, please note the key point here: just because RMS sets a personal rule doesn't mean it is ipso facto FSF policy.
While it is RMS' personal demand that the name change occur as a term to accept the engagement, the FSF does not, as an organization, demand such name changes. We simply request them.
Bradley M. Kuhn, Executive Director of the FSF
Because the GNU tools are add-ons. BSD (Berkeley Software Distribution) is a complete operating system. If you prefer the GNU tools over the Berkeley tools, you can use them. However, they aren't necessary.
GNU/Linux needs the GNU tools to be useful.
Rich
Get every bit of GNU software off your systems. Then see what your "linux" system is worth. Sure, you can get by without gcc, gimp, gnome, ncurses, emacs, bash. But you can start by getting glibc off your systems. And after you delete it, reboot.
Ncurses doesn't use a GNU license. It uses an MIT style license.
You are missing the point here. XFree86 is packaged independently in many GNU/Linux distributions, but it is not an inherent part of the system that you call Linux.
Linux the kernel is minimally distributed with either GNU components or GNU add-ons. Further, it requires GNU packages in order for it to work. This is the distribution that should be called GNU/Linux. Try not to confuse yourselves, will you?
Anyone who wants to understand RMS should read Free as In Freedom first. It's free as in beer and free as in speech.
The most chilling part of the whole book is a reference to Bill Gates' Open Letter to Hobbyists. It's chilling really, to think of young Bill, writing this letter, unaware of where he will be in 25 some-odd years.
Two quotes from this letter are very interesting:
Nothing would please me more than being able to hire ten programmers and deluge the hobby market with good software.
I think he succeeded in the hiring and deluging, but some might want to qualify "good".
But ultimately, he got one thing terribly wrong...
What hobbyist can put 3-man years into programming, finding all bugs, documenting his product and distribute for free?
It turns out that the Internet makes that thing possible. A math whiz at MIT, a graduate student in Finland, a couple hundred thousand computer hobbyists all over the world can do this.
Before you attack RMS for his wacky views, remember that he has committed his life to his views. He's committed his life to proving Gates' letter wrong.
Of course, you can buy tools from someone, and IMHO, that's okay. I can't grow bananas in my back yard, so I pay money for them. I suppose RMS would do without bananas.
But the GNU project is all about making the quality tools for free. The GPL enforces that notion with our own crazy copyright law. To extend my analogy, I do grow basil in a window box. You're welcome to some of my basil, but if you want to a whole lot of it, please give me some of your fine pesto sauce. You don't want to share, OK, then get your mitts off my plants.
More than many other modern occupations, programming is a craft: like gardening, or woodworking. Many people do it for fun. Some are lucky enough to get paid. The freedom RMS is fighting for is the freedom to share your code openly. You get my basil, I get your pesto, and everyone can get bananas. The conventional rules of scarcity don't apply with code. As we share, our tools get better, we become better craftsmen and perhaps we get our $208 back.
My father is a blogger.
As pointed out you are wrong about the S. Your basic premise is also wrong -- Historically, BSD was a set of patches to AT&T UNIX, not a complete stand-alone system. It took years of development to create a runnable version of BSD with no Bell code in it.
You are correct about GNU's goal to be a complete replacement for UNIX -- partly as a reaction against Berkeley's incremental improvement approach that still left you depending on a Bell licence.
Where is the Catch 22? What rights do the employees have regarding the tools they use to perform their job? They are not the consumer in this case, they are employees. The consumer is the company. If the company changes and redistributes the software within the company but not outside the company, it is not violating the GPL License. There is NO End User License Agreement (EULA) in the GPL, only provisions for redistributing modified code and/or binaries to the public or to customers. Unless such action is taken, the company is the end user, the consumer. It would certainly be nice if the company provided its employees with information on how to obtain and install free software for themselves, but they are not obligated to do so.
assert(expired(knowledge));
> That's the weirdest spelling of "Software" i've ever seen.
l does indeed say "Software" and they ought to know.
Bother. At least it's an old and common one.
My reference books are at home, so I just did a Google on "Berkeley Standard Distribution" to check my memory.
However http://coe.berkeley.edu/labnotes/history_unix.htm
rant
A lot of the objectors to "GNU/Linux" think RMS is some random guy who wants to rename their system, not realizing it's the original name of the system they've been running all along that's been whitewashed out of history by people who value function but not freedom.
Stallman's claim is that GNU doesn't get enough credit.
And for people who keep saying "HuH HuH What about GNU/XFree86/Linux/Apache... OH DAMN IM SO SMART", XFree86 is part of the GNU system. The GNU system is a bunch of stuff from a bunch of different groups. Not all of it written by the GNU Project members and not all of it is copyrighted under the GPL. It's not even very hard to find RMS saying this himself.
The incredible irony is that Stallman should have just called GNU + Linux by the common name "The GNU System". He was already calling XFree86 + GCC + TeX by the name "The GNU System" and nobody complained about that. But instead Stallman recognised the huge contribution that the Linux kernel provided to the GNU system and graciously called it the GNU/Linux system instead of just the GNU system.
I get the impression that the people who insult RMS - like you - simply don't understand what he's saying. Admittedly RMS is not a very good communicator of his ideas but ingrates like you don't make it any easier.