Last week, I profiled Richard Stallman, a central figure in the ongoing saga of the free software movement. Within minutes of the publication of the story, "The saint of free software," my e-mailbox began filling up with a provocative outpouring from Stallman's colleagues, friends and enemies.
Assorted free software/open source zealots contributed their own perspectives on Stallman's role in the free software movement. Tim O'Reilly, the president and founder of O'Reilly & Associates -- a computer book publisher criticized by Stallman for publishing "non-free manuals" -- asked for a chance to present his rebuttal. And Eric Raymond, the high-profile open source evangelist, chimed in with a detailed critique that he simultaneously forwarded to Stallman himself -- sparking a vigorous, eloquent e-mail debate.
Numerous correspondents objected to Stallman's attempt to impose the "GNU-Linux" name on the Linux-based operating system. As many readers pointed out, the operating system commonly referred to as Linux includes several elements -- such as a free version of the X-Windows system -- that were not produced by the Stallman-founded Free Software Foundation/GNU Project. To relentlessly insist on one particular terminology, suggested readers, was pointless nit-picking.
"As far as the whole 'GNU-Linux' thing," noted reader Andrew Norris, "the one sure bet out there is that getting hackers to add gratuitous syllables to terms is a losing proposition. When was the last time somebody told you about using 'Microsoft Word for Windows 97 Service Release 1' to edit a document? 'Linux' is the shortest name to adequately describe the OS, therefore that's what hackers use. Anything else sounds like a legalism."
Even worse, suggested reader Nathan Mates, Stallman's insistence on dictating the course of the free software movement contributed to the increasing irrelevance of the GNU project.
"GNU has floundered over the past few years precisely because of Richard Stallman's hyper-controlling mentality," wrote Mates. "The culture of Linux/Apache/etc. development is friendlier to developers: Linus Torvalds loves to work with others, admits his faults and failings, and accepts stuff that's better than his own. Richard Stallman's development philosophy is almost antithetical -- he must control, and so his projects suffer."
But the commentary was by no means all negative. Several readers wrote in to express their own appreciation of Stallman. Chris Hanson, a research scientist at MIT who says he has known Stallman for 20 years, contributed the most telling appraisal.
"Most people that I know are seriously alienated by Richard's politics and by his uncompromising attitude; I'm often uncomfortable around him as well," wrote Hanson. "But he has a knack for getting to the heart of things, and once you understand where he's coming from, the things he does make perfect sense. In fact, it's hard to understand how else they could be done. It's sad that so many people reject him out of hand, often while mouthing some empty boilerplate phrase about how they admire him for his programming skill or something. As if one part of him could be separated from the other."
"I don't always agree with him," added Hanson, "but I always listen carefully to what he has to say. Richard is a genius, a man with a clear and unusual vision, and like others before him, he comes in a quirky and difficult package. Mozart wasn't too well-liked among the cultured people of his day, either; perhaps someday someone will make a movie about RMS, his dry humor, temper tantrums, and beautiful vision of people working together."
- - - - - - - - - -
For the past year, Tim O'Reilly has been playing an increasingly prominent role in boosting the profile of the open-source movement; his company, O'Reilly & Associates, organized both April's "Free Software summit" and August's Open Source Developer Day. But O'Reilly is in a unique position: As he admits, no company has made more money from "free" software than O'Reilly & Associates. O'Reilly's manuals for Linux, Perl, the X-Windows system and other pillars of free software are widely admired in the computer publishing business -- a fact that has led Stallman to argue that O'Reilly is actually a negative influence on the free software movement, since the quality of O'Reilly manuals discourages people from writing free versions.
I caught up with O'Reilly in person, and he gave a spirited defense of the O'Reilly business model. O'Reilly says that he has experimented with publishing free versions of O'Reilly books under Stallman's "copyleft" protection -- a license designed to ensure that anyone can copy or modify a particular information product. As a case in point, O'Reilly noted that he had published a copylefted version of the Linux Network Administrators Guide. But the experiment didn't work out.
"'The Linux Network Administrators Guide' was originally a book that was copylefted," says O'Reilly. "We took it, we worked with the author, and we developed a new edition which was about twice as long and we copylefted that. What we found is that several of our competitors basically took the book, published it also, and it sold much less well than our other Linux books that weren't copylefted."
"The fact is, changing the business model is hard," says O'Reilly. "We grew up in a business which is not 'let's put these books out for free and see what happens.' We grew up in a business where we were selling books, and it is a very competitive landscape. [The publishing house] Macmillan actually had a group, a formal group that was called 'The O'Reilly Killers,' whose sole purpose was to try to figure out how to put us out of business. So we look at that competitive landscape and say, gosh, if we made our books available in such a way that Macmillan could republish our books, they'd do it in a heartbeat."
"Stallman's argument is simply that if there weren't good proprietary manuals then there would be good free ones," say O'Reilly. "I think that is just bogus. I think people write free software because it solves a problem, their own problem -- they are typically trying to fix something that is broken. People who know something don't try to write documentation for themselves. Writing documentation is fundamentally an activity that you do for other people."
Ultimately, O'Reilly says he parts ways with Stallman not on the particularities of the debate over free manuals, but on the larger issue of diversity within the free software/open source movement.
"I guess my biggest issue with Stallman overall is that he doesn't really have a lot of room for other models," says O'Reilly, "and at bottom the thing that I like best about free software/open source is that it allows people to do whatever they like."
Eric Raymond is probably the most articulate spokesman for the pragmatic wing of the free software movement --which argues that the open source development methodology produces the best software products. As such, he is evolving into the antithesis of Stallman -- who argues repeatedly that free software is superior on moral grounds, rather than technical grounds. Here are excerpts from the Raymond-Stallman e-mail debate, edited for reasons of length (as well as to excise some of the more overly personal ripostes).
RAYMOND: There's a difference between wanting to write RMS [Stallman] out of history (which nobody wants to do) and wanting to shut out his rhetoric and tactics in the present (which many people do desire). Many of us count him as a personal friend (I personally have known and valued him for longer than the GNU project has existed) but find that his public behavior continually exasperates us.
That [April] Free Software Summit conference makes a perfect example. I argued strenuously for inviting RMS. I lost the argument; Tim O'Reilly and other invitees on the pre-conference mailing list feared RMS would use the event, and the press conference afterward, for exactly the kind of embarrassing and divisive display we saw at the Town Meeting at the end of Open Source Developer Day.
It pains me to admit it, but they were right and I was wrong. RMS's kind of acting-out would probably have destroyed the fragile advantage we took away from the Summit and subsequently built on with good effect. This is not the first time Richard has had to be handled with tongs. His "Saint Ignucius" persona seems revealingly symbolic not just of dedication but of a thirst for martyrdom -- a self-destructive drive towards positions so purist and extreme that, often, he renders people who want to be his allies unable to support him.
RMS is right to point out that the GNU tools underlie and have enabled Linux and more recent efforts. But RMS himself rewrites history when he claims to have founded the free-software movement; important strains of it (such as the BSD Unix tradition) predated him and remain both technically and ideologically independent of his Free Software Foundation. His dismissal of these strains, and his St. Ignucius act, show that the position he truly craves is not that of arch-hacker but arch-ideologue. RMS wants to be the pope of a flower-children's crusade.
I would not be in my present role of public evangelist for "open source" without RMS. This is true both in a positive and negative sense. Positive because he created many of the technical facts on the ground that made "open source" as we know it, and successfully mobilized the hacker community itself. Negative because he failed to present (or even formulate) a message that would be accessible to the wider world outside it.
So beware of conspiracy theories or divagations about "certain people." It suits Richard's propaganda purposes to cast himself as a saintly victim of some sinister cabal conspiring to sell out the hacker community. But the truth is that just as Richard's victories are very much of his own making, so are his defeats.
STALLMAN: Free software developers disagree on philosophy and politics, but we have some views in common. It could be useful, for certain purposes, for us to work together on the goals we share, putting aside temporarily the issues where we disagree. This way we could have unity, temporarily, on the issues where we agree.
Open Source Developer Day was not designed to do this. I was asked to keep silent about my views that the others disagree with, but they had no intention of holding back their views on the same issues.
Several long speeches during the day were [pervaded] by the assumption that non-free software that relates somehow to free software constitutes "value added" -- an assumption which is the direct opposite of what I am trying to tell people. I was not supposed to state my side of this issue; I was supposed to talk about another topic. I brought up this issue anyway, during my speech, because I was incensed at how the agenda had been set up to present only the other side.
RAYMOND: When the purpose of the event is to sell our ideas to the trade press and business, there are times when the speeches of people you disagree with are functionally helpful and yours are not. Therefore, if I am trying to get victory for all of us, I may have to put pressure on you but not on the people who disagree with you -- even if my private views are actually closer to yours.
I'm not being a hypocrite when I say this, because I myself have positions that I keep quiet about for political and marketing reasons. If the Open Source Definition completely reflected my personal convictions it would be a bit different than it is. But I've left it alone because it works. The fact that it works, and the consensus around it, is more important than the points on which I differ with it.
You [Stallman] did a lot of damage, more than you probably know. I've seen three different articles with the basic theme "those flaky hackers can't get it together and won't grow up," every one citing your diatribe. Next time, it's going to be a lot harder for me to argue that you ought to be included. And a lot harder for me to be upset when I lose.
STALLMAN: The idea that "growing up" means neglecting political issues, and jumping on the bandwagon of the day, is a common form of shallowness in our society. Shallow thought is widespread; for it to be entirely absent from journalism is too much to hope for.
I might prefer to stay away in the future. If an event is going to be telling people that non-free "value added" software and manuals are a good thing, and I won't get a chance to debate the point, then my presence would be counterproductive. I can refute it more effectively by speaking freely elsewhere, than by being present but muffled at the event.
RAYMOND: Either open source is a net win for both producers and consumers on pure self-interest grounds or it is not. If it is, you cannot lose; if it is not, you cannot (and should not) win. Either way, the moralizing you do about how things "ought" to be is at best useless, and at worst actively harmful.
Our most effective lever has always been the "propaganda of the deed" -- the actual demonstration by example that open-source software is superior to closed. When you do that, you address (and sometimes alter) the economic facts on the ground. When you stray away from it, you generally end up doing net damage to your own cause.