Cheap talk

Why Open Source and silence don't mix

Late last year, after spending too much of my life treating the word market as a verb, I decided it was time to burn down Marketing as Usual.

It amazed me that the Internet had already been Hot Stuff for longer than anybody goes to high school, yet most companies were still "targeting messages" at "consumers" as if the only instruments at anybody's disposal were remote controls and toll-free numbers. Despite massive evidence to the contrary, marketers were treating the Internet as TV for myopics, who they referred to as "eyeballs" worthy only of "capture." They brought to mind this apocyphal line about one of Silicon Valley's most arrogant employers: "The clue train stopped there four times a day for ten years and they never took delivery."

Well, that quote floated into a conference call between myself and three like-minded chaps, and within seconds we had registered and a viral meme was born.

In the spirit of Martin Luther, we arranged a pile of clues into 95 Theses and nailed it up on the Web as The Cluetrain Manifesto, which spoke in the second person voice to clueless companies everywhere. "If you have time for only one clue this year, this is the one to get," it said. "We are not seats or eyeballs or end users or consumers. We are human beings — and our reach exceeds your grasp. Deal with it."

One of the first Cluetrain passengers was Eric S. Raymond, president of the Open Source Initiative. He wrote, "The Cluetrain is to marketing and communications what the open-source movement is to software development — anarchic, messy, rude and vastly more powerful than the doomed bullshit that conventionally passes for wisdom."

Within a few more weeks Cluetrain had been translated by volunteers into several languages and was the subject of a major piece in The Wall Street Journal. Today well over a thousand people have signed the Manifesto, its theses are being quoted all over the place, and authors have a nice book deal. Not bad for an act of arson.

But nothing special for an open source project.

If you're in the open source world, you already grok most of what Cluetrain is about, because open source is inherently clueful. In fact, open source succeeds in large measure because there is nothing anti-clueful about it.Open source has no secrets. It doesn't come from distrust.

If the whole Cluetrain Manifesto could be reduced to three lines, it would be these:

  1. Markets are conversations
  2. Talk is cheap
  3. Silence is fatal

Open source implicitly trusts and relies the conversations that comprise its markets. This is what makes open source fundamentally different than closed source. Not only can you do more with it (and to it) because everything about it is exposed, but it trusts you enough to disclose all of itself to you. Remember that the Linux movement began with an act of disclosure by Linus Torvalds. The explosive Linux conversation we have today — which now includes the whole world of computing — began as a conversation about source code among the hackers who helped Linus create Linux.

Contrast that with another worthy operating system: BeOS. By many measures, BeOS is better than Linux, especially on clients. In's Be View, Scot Hacker says, "Here you've got an operating system that's easier to install, faster, and more stable than Windows, and that leaves Linux in the dust when it comes to ease of configuration and ease of use. The more time I spend trying to get friendly with Linux, the more I come to agree with Jamie Zawinski (founder of, when he said 'Linux is only free if your time has no value.'"

Yet BeOS is still struggling while Linux and open source are burning down Development as Usual. Why? Is it just because open source has more Goodness than closed source? No. It's because open source has no secrets. It is inherently disclosing. And disclosures start conversations — and then do nothing to stop them.

BeOS was developed in deep secrecy. Few people outside the company knew anything about it, and if they did they were sworn to say nothing. As a result there was no demand for Be boxes or the BeOS at the time when they were announced — even though the announcement itself was a storybook success. When Be founder Jean-Louis Gassée demonstrated the first Be box at a major industry event, the demo was a mind-blower and Jean-Louis got a heartfelt and tearful standing ovation. But that was it. Press interest in Be lasted barely longer than the applause.

Now, years later, we are finally starting to see a groundswell around BeOS. But that's despite a lack of openness, not just to development but to conversation about OS code by the people who matter most when a technology needs to establish itself — the nerd community. Witness the Slashdot conversation about Scot Hacker's article. When BeOS succeeds, it will owe much of its success to adoption and blessing by nerds who hack on Linux — and nothing to Be's original secrecy (or the secrecy that persists around its source code).

The world is full of secrets that never should have been. For example, can you name the key developers of the PowerPC chip? Well, there's IBM, Motorola and Apple, right? But who else? Try Groupe Bull, the French systems company.

IBM brought Bull into the PowerPC development project early on, because Bull had world-class expertise at something IBM's original POWER chip lacked: the ability to do symetrical multiprocessing (SMP). So PowerPC owes much to Bull's SMP design expertise. But Bull, like Be, was an extremely secretive company. So they stayed mum about their involvement with PowerPC until it was time to announce their own PowerPC-based SMP product line. As an event the announcement went well; but as a conversation it failed. Today almost nobody knows — or cares — that Bull's fingerprints are all over PowerPC. And we never hear about SMP-based PowerPC systems. In other words, Bull's silence not only prevented interest in their own products, but denied high-end success to PowerPC as well.

So here's the clue we're talking about here: Outside the secret-keepers themselves, there is no demand for secrecy. No market for it. And since markets are conversations, you can't use secrecy to make a market. Only to prevent one.

This makes Linus' employment by the hypersecretive Transmeta an extremely ironic matter. No doubt Linus' involvement will drive plenty of interest in whatever it is that Transmeta does; but when the company finally announces something, it will find little in the existing conversation to sustain the applause.

Let's hope, for our hero's sake, that Transmeta's silence isn't fatal.