How RMS is vital to the free software movement, ESR looks like he's sold out and Linus isn't so special as we all make out.
How RMS is vital to the free software movement, ESR looks like he's sold out and Linus isn't so special as we all make out.
Since 1984, RMS has been working on writing software for, and promoting the GNU project. A project to provide software which is not only royalty free (free like beer), but that allows the user to modify, pass on and generally screw up said software (free like speech).
The GNU project has always had the aim of replacing UNIX with a workalike system (it could be argued that this is the aim of emacs alone). By the early 1990s GNU was providing a complete set of development and user tools to run on top of many commercial operating system. The only part missing was the kernel.
GNU have been working on their own UNIX like kernel. Built on top of the Mach Microkernel, HURD aims to compete with the most advanced and modern operating system kernels to date. However, development (which of course had to be done using entirely GNU tools) has been slow and even now HURD is not ready for any sort of production system.
In the early 1990s Linus Torvalds, appeared from nowhere with a working rewrite of the Minix kernel written under the GPL, Linux. The Linux kernel is heavily based on tried and tested designs, old technology. However, it works, is fast and incredibly reliable. This was the spark on the arms dump that was GNU. Suddenly there was available a completely free operating system with all source code and a range of user and development tools.
In media terms it appeared overnight. One minute there is a bunch of obscure hackers writing compilers for UNIX, an OS that had not even been heard of by most computer users. The next, there are a few distributions of "Linux", providing the kernel alongside sets of GNU tools.
Linux took off, picked up by many students wanting to get their hands dirty with something that they could work on and learn about it was propelled into teaching institutions, ISPs and the hands of even more hackers. By 1998, Linux was being touted as "the last best hope" against Microsoft just as the Apple Macintosh had been before they went into their long dark period of flaming Powerbooks and buggy Finders.
Linus Torvalds will not be remembered in history as an innovator, he will be remembered as in implementor. As his discussions on Minix with Andy Tanenbaum show, Linus wasn't concerned with new technology, taking advantage of powerful hardware or dealing with the problems of tomorrow. He seized the opportunity to apply textbook principles and build an OS kernel using 60s concepts. Linus should not be hailed as a great hero, who boldy coded where no man had done before. The reason that Linux is now so good is the work of thousands stabilising and improving the system. Linus should, rightly, be congratulated for sitting down and doing a dirty job that nobody in operating systems wanted anything to do with, writing a working system using old technology.
Next came the ugly bits. Industry wasn't interested in an operating system written by "hackers" thrown together from whatever was available. They refused to provide device drivers for Linux, mainly because they were concerned that they might give away trade secrets by providing free source code under the GPL. Throughout the 2.0, 2.1 and 2.2 kernels, Linux changed constantly. Providing binary only drivers for it became impossible (was this on purpose). Companies had no choice but to provide code that could be compiled against the a kernel of choice. This meant opening up precious source code.
Source code was released under a variety of licenses. There was GPL code, BSD code, XFree86 code, Apache code, Artisticly licensed code and all sort of other weird things. The only common factor was that each provided source code and allowed users to at least distribute untampered versions of source code and binaries.
So, in an effort to tidy up the situation, the "Open Source Movement" began. Fronted by ESR and Bruce Perens it brought together all code fitting a common denominator of source code availability and freedom of copying under the banner, "Open Source". Initially, opensource.org claimed to, and did, act as a marketing campaign for the GNU project. It generated amazing amounts of publicity.
However, when opensource.org started to class software such as QT under the same banner as GCC and other GNU software, RMS took issue. He denounced open source as not being purely free software and distanced himself from the movement.
Open source is the power hungry brat child of GNU. Concerned with short term publicity and gain, they abandoned the principles that have given GNU such a strong foundation. After RMS split from opensource, there were various other internal squabblings, most visibly over the use of the trademark "Open Source". Next came the talks at Microsoft from ESR and the killing he made by being on the board of VA Linux. In the space of a few months he managed to suddenly move from the editor of the Hacker's Dictionary and hacker icon, to sold out betrayor of GNU in the eyes of many.
In a sense, ESR not only distanced himself from the hacker ideal. But showed software developers and marketeerers just what potential for cash-in existed in open source software. Since then, it seems, open source has been the latest and greatest buzzword. Everyone (even Microsoft) has either released open source software or talked about it. Suddenly, there is a vast amount of code available to normal users.
RMS argues that it is wrong to call the "Linux distributions" "Linux". Instead he favours GNU/Linux, to show that the system is comprised of both GNU tools and the Linux kernel. This will probably never happen as the term "Linux" is so well established in the media now (when HURD comes along, things may be very different though). A much better name for most of today's Linux distributions would be opensource/Linux. For example, Mandrake comprises binary only versions of software such as netscape while providing open-source software with restrictive licenses such as QT. The only distribution which could realistically be called GNU/Linux is Debian (but only if they finally ditch non-free).
Recently there was a Slashdot interview with RMS where questions were submitted by users. The story carried a health warning. RMS is accused by many of being a zealot who wants to see all programmers starve. He is not.
RMS provides a much needed figurehead for the FSF. A group devoted to providing and fighting for free software. Much like Marx, Machiavelli or Neitsche everything he says should be taken with a pinch of salt for life in the real world. But without these people, without the purist ideals they promote we would be stuck in a realistic world of pragmatists ready to sell out at the first opportunity, hardly role models.
It is possible that someday RMS will receive the Nobel Price for his contributions. The place of RMS in history should be someone that changes the world, like Washington and Einstein. Neither ESR or Linus would ever be remembered in that way. RMS has weaknesses: he is not a good manager and his interpersonal skills leave something to be desired. But overall, his positive points far outweight his weaknesses.
(BSDers: if you flame, please direct that to /dev/null)
I'm afraid I'll have to object to saying that calling Qt (at least Qt2) "open source" is a sellout. Its license meets the Debian Free Software Guidelines. You can write free software that links to it. (Qt1 wasn't open source, and I don't see anyone at OSI, or Troll for that matter, who pretended it was.)
Feel free to attack ESR and Bruce Perens for whatever other failings they may have, but the acceptance of the QPL as a free license isn't one of them. The QPL may suck for other reasons, most notably the apparent incompatibility with the GPL, but it is free software, even in RMS's world.
All I can say is people like to call RMS eccentric, crazy and even
unfriendly. I will say one thing he is consistent. He devoutly
preaches freedom in regards to software and for over 15 years has stood
by that stance. Funny that you bring up ESR and Linus.
ESR seems to be the loon to me. He changes his definition of Open Source on a daily basis and has some wild ideas as to WHY businesses need to Open Source. Then you have Linus, here is a guy that could actually make real money doing GPL works for about any GPL producing software house in the world, yet he decides to go to Transmeta to do work for prop systems. Obviously he didnt choose the GPL for Linux because he valued freedom.
I think probably the most frustrating thing for me is the ease in which people drop freedom for the almighty dollar. One example is PHP. Here you have a piece of software that is very well put together and a poster boy for open source. Only now with the creation of Zend they have altered their license are propogating closed solutions. : ( Such a shame. At least MySQL woke up and reversed their position. From being non-Open to free over night. : )
So the question I think all free software developers need to ask in introspection is when the big buck comes a knocking will you answer and leave your freedom at the door, or will you ignore the knock and keep turning out free software. : )
Not to turn this into yet another flamefest, but it itches that the author of the article makes the assumption that the goal is more "free software".
It might be for some, but for some of us it's not about free software as much as about better software. Open source (in a broader definition than the one RMS uses) obviously happens to be an effective way to get better software.
Kris Kennaway have said it better than I could:
"Why do I prefer the BSD license? Because it helps to raise the bar of software quality for everyone, everywhere. BSD code can be used anywhere with no real strings attached, so it makes a logical foundation from which to build yet higher. BSD software makes the world a better place because people don't have to worry about reimplementing the exact same damn piece of code from scratch, simply because the license terms are unacceptable for their project."
I think Apache is the best example of how software with a BSD like license can help making the world a better place to be. Without Apache as a reference, http might have been embraced and turned into some nightmare by Microsoft or Netscape like they did on the browser side. A lot of the "Apache culture" and the work on Apache comes from the fact that it is not under a restricted license like GPL.
RMS is a visionary. He has a vision of what he considers would be a better society, and he is not willing to compromise it. Personally I consider him as an ideological leader and a lot of his ideas actually make a lot of sense to me. We may never live in the society he envisions but we can at least try to approach it. He has been showing the path for 15+ years now, and I think he deserves a lot of respect for that.
Linus is not a visionary in the ideological sense, but he is a damn good programmer and project leader. He has a lot of practical sense.
ESR is wierd. I dont quite understand him. He is a great hacker, but as a would-be-leader he leaves me much more sceptical. He tries to be practical and sell the idea of open source to corporations, however I think he makes a lot of compromises along the path and I dont quite like that because I think that when you start to compromise your ideals you soon discover that you have completely lost them. His ideas probably have some value to them, but I think its pretty bad that he misrepresents them as being the ideas of the free software community in general.
Overall I really appreciate RMS and linus for different, but complementary reasons. ESR tries to be a mix of both and it does not really cut it for me :-)
The author misrepresents RMS:
RMS argues that it is wrong to call the "Linux distributions" "Linux". Instead he favours GNU/Linux, to show that the system is comprised of both GNU tools and the Linux kernel.
This is simply false.
RMS has explained, over and over again, why he would like it to be called GNU/Linux, but his explanation is routinely ignored. Probably it's inconvenient to remember the truth.
If you have any interest in the truth, read:
RMS argues that it is wrong to call the Linux-based distributions "Linux". Instead he favours "GNU/Linux" because the distributions are a direct outgrowth of the GNU project, the first-ever project with the goal to produce a Free operating system.
This is why he doesn't suggest "GNU/BSD", or "GNU/X/Linux". He also doesn't ask for it to be called "RMS/Linux": by promoting "GNU" he is honoring the hundreds of contributors to the GNU project who wrote the dozens of critical programs according to the master plan for the operating system before Linus ever thought of coding a kernel that would run in it.
Linux is an important program, but it's just a program. An operating system is composed of much more than just programs. The GNU Project includes the programs and everything else, and what we so blithely call just "Linux" depends on all of it.
We all have loose bundles of motivations when we write free/open/gift software. I've used a variety of licenses over the years. I'm confortable having used something BSD-ish for one project and the GPL for another. Each seemed appropriate in the setting.
Things sometimes seem a little complicated as we decide what projects to create or support, especially as licenses collide. Personally, I think the important thing to remember is that more than one motivation can exist. I am very hesitant to label anyone else's motivation as wrong, or even misguided.
In my opinion, diversity is alright.
In my opinion, diversity is alright, and even needed.
Because of this paradox :
i.e. but we only can understand the novelty thru the things we already know.
And How do we cross this Gap to the unknown ? Who is the best guide ?
A visionary is not necessarily a good back-office man (man of detail and habit/repetition), a back office man is not necessarily a good teacher, a good teacher a great orator/preacher, an orator a negotiator.
These specifics qualities brings "side effects" ( like in RPG, where pushing one quality weaken another one). This is why a group of people with "Extreme qualities", is powerful, but last as long as they can respect their differences, which are also their global strenght. Dismissing one is weakening the group.
Visionaries may not like the negotiator's arrangement, they see it as treachery to the "sacred goal", or teachers's simplification, which becomes "compromises", or lobbying, regarded as "sold out", but all these are needed for the "others", the unbelievers, the unconvinced "yet" to be reached by the "faith".
A vision becomes reality if you can share it, but the harder the width of the gap.
Then the faith is can become a "culture", when enough can understand "their" benefit (and not "your" goal).
Bottom Line, I am personnally more interested by what these great people bring to me, rather by their "negative qualities". I never expected them not to have some . If what they "sell" (or give) does not interest "me", I just dont buy. But why should I prevent others from buying ? This is not selfishness, just free-will.
dneighbours, please note that PHP is fully open source. The fact that a company will be selling closed-source add-ons to PHP has nothing to do with PHP. And calling the developers of PHP sellouts is just plain ignorant.
There are a number of interesting points brought up in this discussion, but I'm not going to reply in substance, because I don't like the tone of the original post. To me, it's not respectful of the difficult work that these "figureheads" have devoted to the cause of free and/or open source software.
So, in the spirit of name-calling, where the hell is your fantastically innovative yet carefully engineered and thoroughly optimized operating system kernel, written with no desire for any kind of profit or benefit? And don't start on how you ate at the soup kitchen because you couldn't afford to buy your own food - that just shows how much of a leech you are compared with normal, productive members of society.
Methinks maybe you should reconsider that whole casting the first stone plan.
(and who said that Web discussions aren't like Usenet? *plonk* :)
I really hate to see "think pieces" like this. Making personal observations about public figures like ESR or RMS based solely on their published writings and speeches is pointless. Furthermore, referring to ESR as a "Open Source sellout" is needlessly inflammatory. ESR must live by his own decisions, as we all do.
The Linux community as a whole is suffering from an excess of testosterone, and this is due in part to the influx of high schoolers and college kids with little or no experience in the larger world. These kids don't have mortgages to pay, families to feed, or long-term careers to worry about. (They'll probably change jobs every two years, so they don't really give a whoop about longevity in one place.) I happen to agree with some but not all of the tenets of Open Source, but it is unfair to attack people who believe differently.
I don't want to paint all young hackers with the same brush; I'm sure there are more level-headed and thoughtful Linux programmers than not. However, pieces like this one make me think that most of the "new generation" programmers are not idealists, but simply punks with an axe to grind.
My advice to them: grow up.
Anyone who claims that the Linux kernel is merely a rewrite of 1960's technology obviously has no understanding of the work that has gone into it. In addition, it would be interesting to see them explain why other Un*x vendors are looking to Linux for new ideas in kernel-space.
I am not too sure linux should be said inovative. It is not exactly the greatest most inovative most intelligent piece of code ever. It is a good stable operating system which has some known problems but its greatest merit is to exist. Simply to exist.
Linux' greatest merit to me is to have brought mainstream college boys as someone caled them sooner into free software development. Because someday, these young people will grow up and rule the world and at least some of them who have the power to make decisions will remember what free software is like.
Just my rant.
The open source movement needs RMS, even if it doesn't like him. The reason: he prevents parties opposed to open source from manufacturing our consent to a watered down version of its core tenets.
Noam Chomsky, in his political analyses, describes a fascinating propaganda technique known as "the manufacture of consent". It refers to the consent of the governed to the actions of government. In a democracy, the tools of political censorship and control must be much more subtle than in an overt dictatorship. The manufacture of consent relies on producing the illusion that the centre of a debate lies somewhere other than between its natural poles.
For example, if you're an anti-abortion media spin-doc trying to skew a debate on abortion, you first portray yourself as a fearless and impartial journalist. You then frame the debate between the poles of a Pat Robertson (on one hand) and a George Bush Jnr (on the other). You are implying -- to the uncommitted members of your audience -- that anyone more liberal than George Bush is a whacko lunatic, because of course George defines the left-wing position and Pat defines the right-wing. Committed anti-abortionists will go along with you, while pro-choice viewers will feel confused and isolated because they've been implicity consigned to the extremist fringe of the debate.
In this way, the illusion will be created that that thedebate is between moderate anti-abortion types and total abolitionists, and that a consensus exists that permitting abortion is undesirable, while also maintaining the fiction that the debate is impartial (and anyone who disagrees is an extremist).
In the software debate, RMS acts as a vital anchor on the fundamentalist open source side. His existence means that it is not possible to re-frame the debate as being between, say, Bill Gates and John Young (or Bill Gates and Larry Ellison -- gack!).
I do not agree with all the positions RMS stakes out. Nevertheless, I am very glad that he is there to stake them out -- if he did not exist, we would have to invent him, or we would be lost. He defines our extreme position in a sane and sober manner, without compromise. It will be a very sad day for open source if he ever gives up.
If they can keep people arguing left-wing free software vs. right-wing free software, then we win, because everyone takes it for granted that the base assumption is free software!
Another part of the manufactured consent game is what Rick Moen likes to call "moderate-fu". You simply point out extremes on either end of a debate, and show how moderate you are. By contrast to the extremists, your argument appears to have merit. Thus, Linus Torvalds appears to be a level-headed pragmatist who is driven by a desire to do The Right Thing. What more could you ask for?
Yes, he's a bit of a figurehead, however deserved or undeserved. But in playing the game of free software PR, he's invaluable. He sits down and calmly talks about the coding he's done and his personal motivations for using the GPL, and all observers get warm feelings in the pits of their respective stomachs.
The whole process whereby ESR suddenly became seen as a figurehead of the open source movement is amazing. Eric has really built on three things, and it was all borrowed glamor.
1. The New Hackers Dictionary.
3. The phrase "Open Source".
TNHD is where he showed up in the public eye first. The people who contributed to and maintained the original Jargon File and Hacker's Dictionary were often horrified at the way he ripped out what they considered important references, replacing them with definitions they often strongly disagreed with.
Fetchmail. He picked up an open source project and shepherded it for a while. Important work, but there's thousands of others who did it as well.
And Open Source is just an evolution on the Open Systems / Software Tools approach many of us were touting in the '70s and '80s. And in a lot of ways Open Systems are more important than Open Source. I don't want to see the source to Windows, or Netscape, or Star Office. I have better things to do than dig through yet another pile of crap corporate code. So long as the interfaces and protocols are stable and well defined, like TCP and HTTP, it doesn't matter whether the web server you're talking to is written in Visual Basic on Windows or C on Linux.
A lot of open source software is moving steadily away from Open Systems. Look at the pile of software you have to bring in to run Gnome software, because there's all these Gnome-specific interfaces and protocols. KDE brings in their own. And then there's Star Office's 21 private libraries.
It's definitely time that we dropped back and stopped worrying about RMS and ESR and Linus, and rethought the lessons that DMR and BWK and Ken Thompson taught us.
I agree with Raph: there's a whole lot of thoughtlessness here. And simple ignorance of facts: that RMS rejected the Linux kernel because he thought that hurd had a better future, and switched only when it became obvious that hurd failed; that ESR's deserved fame came not from his free software projects but instead from his The Cathedral and the Bazaar paper.
Mocker: n. Someone who mocks others.
Figurehead: n. 1. A person given a position of nominal leadership but having no actual authority. 2. Nautical. A carved figure on the prow of a ship.
What I see here is a bunch of people who, instead of going out and doing something, are busy criticizing the people who are doing the work. RMS this, ESR that, Linus something else. Sheesh!
Yes, to a certain extent, RMS and ESR are figureheads, but given the chaotic nature of the Open Source community, how could it be elsewise? RMS created countless pieces of software ... and then gave up control of all of them by releasing them under the GPL. If the FSF spoke for and controlled the actions of hackers around the world, then RMS as its titular head might avoid the label of "figurehead", but they don't and can't, so RMS remains a figurehead. While occasionally ESR may claim to speak for the Tribe of Hackers, in point of fact, he doesn't, which makes him a figurehead too. Most of the time ESR just claims to be able to describe the features and benefits of open software development, in such a fashion that non-OSS developers and managers can understand those benefits.
In my opinion, both RMS and ESR do important work that goes far beyond mere coding -- they're doing public relations for a bunch of people who apparently, are confusing PR for credit. Neither RMS nor ESR claim to have created all of this software by themselves, or in a vacuum. While they may both disagree on a few minor issues, they both claim to be good friends, and they both believe that openness is good, and more openness is better. What's the harm in that?
New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.
Keep up with the latest Advogato features by reading the Advogato status blog.
If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!