The community of software developers who hang out on the IRC chat channel #abiword is small; the number of nicknames registered as present rarely breaks two dozen. But the room is open, and active, 24/7, 365 days a year, because the membership stretches from Australia to France. When the hackers in Cambridge, Mass., finally drag themselves away from their terminals to crash, their counterparts in Melbourne are just settling in for some serious coding and chatting.
#abiword is a friendly, casual watering hole, a place to gossip about upcoming exams or the perils of Microsoft Word-formatted job applications. It's a place to talk about a relative's illness, or to slap virtual backs when one of the group does something noteworthy, such as getting accepted as a member of the Gnome Foundation, a group of developers dedicated to making the desktop safe for free software. And, naturally, it's the place to share notes on how to overcome nasty problems such as getting embedded tables to display properly in AbiWord, a free-software word processor.
That AbiWord even exists, much less that it is thriving through the steady contributions of this band of programmers, is one of the delightful mysteries of the free-software world. I can recall wandering the LinuxWorld convention floor back in the spring of 1999 in San Jose and stopping to talk with a man from AbiSource, a small, privately held company that was attempting to drum up venture capital for its ambitious plan to create a free-software office suite.
Like so many software start-ups in those hothouse days, AbiSource was aiming for the big score, the IPO that would catapult it to financial heights. I was dubious. While the free-software approach made obvious sense for those working in the truly geeky domains of software -- Web servers, mail-transport applications, operating systems -- I just couldn't see the business model in supporting a free-software word processor, and I was skeptical that building such an application would generate the kind of developer enthusiasm that projects like Linux or Apache enjoyed.
It's one thing to hack an operating system kernel, but endlessly tinkering with file formats? Who wants to do that? It was all well and good to dream of world domination, but free software's chances of challenging Microsoft for desktop supremacy were slim, I thought.
I was right -- but also very, very wrong. As it turned out, venture capitalists agreed with the thesis that a free word processor did not sound like a hot IPO slam-dunk. After sinking half a million dollars into the project, AbiSource gave up, changed its name to SourceGear, turned the code over to the free-software developer community, and looked for other ways to make money. (Which, by the way, it is apparently doing with some success. In its October issue, Inc magazine named SourceGear one of the fastest-growing privately held companies in America.)
At the same time, without any corporate backing, AbiWord has flourished. At the AbiSource Web site, anyone can download a perfectly usable cross-platform word processor. It's unlikely ever to match Microsoft Word feature for feature, and it still lacks in some prime-time aspects, but for most normal purposes, it's good enough.
And as is free software's habit, it's getting better all the time. Bit by bit, patch by patch, revision by revision, AbiWord is reaching the point where it just works. Not because there's money to be made, and not even because its developers necessarily believe that free software will inevitably conquer all. They're doing it, as Eric Sink, the founder of AbiSource, observes, "because it's fun." They're doing it because, while booms and busts come and go, hackers have to hack.
Martin Sevior, an experimental particle physicist at the University of Melbourne, in Australia, is one of the lead developers of this hardy band. He's not a typical member -- he's 43, which, he acknowledges, makes him "quite old for an open-source developer." For months, Sevior, who sees his hacking as the kind of "contribution to the community" that should be expected of an academic, has been focusing on the troublesome task of getting tables to work in AbiWord.
A word processor that cannot display tables, says Sink, just isn't "credible." Sevior agrees, but he thinks he has the problem licked. "We're not finished," he says, "but the end is clearly in sight."
Sevior started playing with computers in the late '70s. He began running Linux in 1993. But while he was happy writing his papers using the LaTex word processing program popular with scientists, he found that he was frequently receiving Microsoft Word documents from "admin types in my university."
"It was a real pain to read these," he says. Every time he found a solution that would work on his Linux system, Microsoft would upgrade its file formats and he'd be stuck again.
Finally he stumbled upon AbiWord. It imported documents that other free-software applications were choking on. It looked good on his Linux box. It was moving forward with admirable rapidity. "I started lurking on their mailing list and was impressed with how people would out-of-the-blue announce a patch for a really cool feature and then after a day or two it would be incorporated into the code base."
So he jumped in, feet first. "I really liked the AbiWord community, so during a vacation in 1999 I dived in and started to learn how the program worked and to start hacking."
He's never regretted the move. "We have a fantastic team of programmers working together on AbiWord," he says. "It truly has been a very enriching experience working with them. I've learned so much and have had so much fun along the way too. It really is wonderful to work with the brightest people on the planet."
During the dot-com boom, when hundreds of millions of dollars were being thrown at software start-ups, and anyone with a passing familiarity with Linux or Perl, it seemed, could walk right into a lucrative job, the overall production of free and open-source software skyrocketed, subsidized by the deep pockets of companies such as VA Linux. After the bust, I don't think I was alone in wondering what would happen. Would development momentum be sustained, or would free software, after having hit a peak of media and investor attention, sink back into obscurity, never again to challenge the powers that be?
The progress visible in a project like AbiWord, or its close GNOME relation, the Gnumeric spreadsheet, or the suite of office productivity apps that are part of the KDE desktop, or a host of other efforts, is sending a clear signal that the bust and boom were, at most, a distraction. And in an ironic twist, the ups and downs of the tech economy don't have the kind of effect you might imagine on free-software production: Unemployed programmers have even more time to hack than employed software engineers.
Not only is contributing to a free-software project a good way to keep your skills sharp and build a résumé, but, Eric Sink observes, "People keep doing it because they like doing it. It compares very, very favorably to working for stupid corporations who ask developers to build projects that they then throw away, or to writing really good code and than have bad marketing put in front of it. Those kinds of things are very unsatisfactory for developers."
What is satisfactory is to receive the plaudits of your peers and to create code that people actually use. It's not about beating down Bill Gates (although Microsoft, by every indication, is increasingly obsessed with beating down free software), it's not about making oodles of bucks, it's not even about being the best.
At one juncture in the nonstop get-together that is #abiword, one contributor to the effort had a moment of near despair as he became overwhelmed by the magnitude of a task he was trying to solve. Another developer, Dom Lachowicz, who along with Sevior is one of the primary engines of AbiWord's locomotion, reached out a soothing hand.
"Try not to feel discouraged or burned out," he typed. "Focus on the things that make you feel good. Ignore the noise."
Sage words from a young programmer. And worth savoring. Because, in the end, seen from the perspective of a hacker trying to get something to work, all the rhetoric about world domination, about booms and busts and IPOs and market share, is just noise. Whereas the camaraderie of developers sharing their anxieties and triumphs, working together to solve problems -- the chitchat that fills the hours spent scrolling down the terminal windows of the hackers hanging out in #abiword -- that's sweet music.