May 24, 2021
The Luck of Open Sourceopen source without great men
War and Peace weighs so much in part because it’s two works in one: first an epic novel of tsarist Russia, then a forceful essay on what that novel means. War and Peace dramatizes the folly of understanding history as a drawing room play of a few “Great Men”, flanked by casts of thousands in frame-filling roles, as flotsam minions.
Personalizing history at a level we can understand, like a tennis championship or a family feud, makes for better stories. But the arc of human experience traces the competing desires, choices, and struggles of great multitudes of individuals—Great and overwhelmingly not too Great—linked by familial and sundry associations, both masters and subjects of massive currents of change. Currents beyond the full understanding of any one person, even a Russian epic novelist.
The dominant story of open source is a story of Great Men, symbolic Actions, and Big Ideas. The Battle of the Printer. The Rout of the HURD. The Open Source Schism. The Mozillian Gambit. The conversion of the flagging, initially hostile Empire of Armonk. We can place ourselves, however tentatively, in bit parts of this drama. Like extras in Ben Hur, we can try to point ourselves out in the action, if not in the credits. In a very serious way, that’s getting history backwards, or at least too simple.
What is the War and Peace of open source? What were the lived experiences and greater forces that Stallman, Torvalds, and Raymond responded to, and came to symbolize? How would we tell grandma the story of open source without those convenient names? With the dust of propaganda and journalism settled, the revisionists are due. We could and should see a people’s history of open source in our lifetimes, without the selection bias of an editor.
There are some popular “alternative” tellings of free software and open source. But largely in the competitive vein of claiming open source as some small part of a still broader, more momentous story, with its own pantheon of even greater Greats. Such is open source as a side quest in the grander Internet saga, for example. Many compelling story beats, from the BSD network stack and DNS through Apache and Mozilla to wherever we stand today, afford convenient hooks for relegating open source to subplot status. Open source was open applied to source, after all, a mere specialization of a more universal, more ambitious, more humanist notion of public-private laissez faire, better grounded in the hardware revolution.
Nonetheless, telling people who are very good at the Internet—software developers—that the Internet is Very Important and Very Powerful still tends to go over pretty well, even if the story’s no longer mostly about them. But I am here to suggest another story of open source, less peculiar, more mundane, non-ingratiating:
Open source is a story of money, hype, and fortunate timing. The luck of open source lay in straddling an epochal market moment, created as much by venture capitalists and speculators as engineers: the incredible froth and ballyhoo of the Dot Com Bubble and the sudden, involuntary shift to thrift once it burst.
Through the middle to late 1990s, free software, eventually open source, achieved mindshare and recognition as a codeveloped marketing asset of a crop of technology firms, largely pickax-in-gold-rush types. These firms benefited from a cohesive, revolutionary spin on what was largely a budget play in a time of easy money.
Some movement folks, like Raymond, got stock in these companies for their troubles. Some, like Tim O’Reilly, built or buoyed their own businesses by pitching in. When the high-flying, speculator-compatible firms successfully went public near the peak of the frenzy—Cobalt, Red Hat, and VA all IPO’d in 1999—a few true believers, propagandists, and soothsayers got the boon of paper wealth and a strong case of follow-on fever. Not so different from the Bitcoin millionaires of the last decade pouring money into “blockchain” and “decentralization” to this day.
Such was open source as “revolution”. So far, so very good.
Then reality came calling. The hype cycle ran out of new money. People looked down. The stocks tanked. Paper gains evaporated. Paper friends disappeared.
In those ashes, the substance rather than the style of open source—its low-as-can-go price tag and insatiable thirst for idle hacker cycles—asserted itself in the clear. The “free” in “free software” did mean “free as as in beer”, after all. At least, as they said, if your time was worth nothing. Which suddenly it was: a lot of competent geeks had time to spare, and a lot of precocious children got dial-up at home.
A lot more people got the price signal than any essays on “philosophy” or license terms. So the phrase “open source” became technical jargon for “free”, the way we mean “free” five nines of the time, with some looming valence of “some assembly required”. As more time piled in, the open source started to mostly Just Work, at least as well as the closed stuff did. Which left the story at “free” most of the time.
Looking back, I’m struck by how few of the dot-com companies built on open source. Yes, they bought a lot of servers, routers, and overpriced colo space, because that was the only way at the time. But the funny money bought a whole lot of license keys, too.
Microsoft was the world’s most valuable company. Oracle was selling hand over fist. Even smaller, specialized shops, like Macromedia, were killing it. In short, the not-open incumbents the open “rebels” railed against—Windows v. Linux, Oracle v. MySQL, ColdFusion v. Perl and PHP—were still very much incumbent. The Internet Eats Everything story of open source, which boasts the virtue of looking back this far, doesn’t leave much of a narrative gap for this transitional period. But so it was.
As we’ve seen since, the collaborative aspects of open source have reliably waned, while its potential to distort markets by zeroing out price has seeped its way through business schools and board rooms.
Many true believers from the pre-bust hype era still maintain that open source isn’t “real” open source unless it’s collaboratively developed by diffuse hobbyists or employees of competitors. But open source projects, especially projects maintained by paid devs, especially projects maintained by well organized teams of paid devs, such as those working for a single firm, need less and less help from their “users”, who really are a set apart. Languages, static analysis, testing tools, continuous integration infrastructure, and development practices have taken leaps. Crash reports and even usage telemetry have been normalized, if you know where to look. Automatic, even transparent self-update is gaining ground, for server-side and installed applications. Eyeballs make bugs shallow, but we’re automating eyeballs, too.
This telling of open source has value, beyond merely escaping the traps of early 2000s propaganda and Great Man fandom. It directs us to ask what bigger trends arguably compelled what became open source, and how much of that was compelled, and how much of it was happenstance. Where are those trends heading now, and how will that likely play out? What similar trends might shift mindshare and muscle memory in the industry?
Your thoughts and feedback are always welcome by e-mail.
back to top — edit on GitHub — revision history