I’m teaching a course this Fall entitled Web Development 101: Hacker Theory and Practice, with the Saxifrage School. For the theory side of the course, we’re going to read The Cathedral and the Bazaar (CatB) by Eric Raymond (ESR). I started my programming career in late 1999, soon after the first edition was published (the core essay dates from 1997). The second edition landed in 2001, soon after the dot-com bubble burst. Until reading the book, I actually didn’t realize the “open source” brand was invented as late as 1998. I accidentally grew up as a hacker right alongside the open source movement.
I finished reading CatB a couple nights ago, in time for Poul- Henning Kamp’s ranting lament for the Cathedral (discussed extensively on Hacker News). I’m a bit of a PHK fanboy, to be honest, having used FreeBSD and especially jail(8) for my first years as a hacker (now it’s mostly Mac OS and Heroku). I used his beerware license for some early projects. However, I don’t think it’s smart to fight with reality, and 50% of the population is below average. To me, the transcendent wonder of the bazaar is in including all contributors and all contributions in something bigger than the sum of its parts.
I think we need more bazaars, and better ones. We need people with PHK’s taste to somehow figure out how to organize and direct the energy of the supposedly unwashed masses. I admire Linus for having done this with Linux (prickly though he might be). I admire Guido for having done this with Python. And to an extent, I admire Jimmy Wales for having done this with Wikipedia.
CatB concludes as follows:
I expect the open-source movement to have essentially won its point about software within three to five years (that is, by 2003–2005). Once that is accomplished, and the results have been manifest for a while, they will become part of the background culture of non-programmers. At that point it will become more appropriate to try to leverage open-source insights in wider domains.
I see Wikipedia as the first non-programming open source success. Based on the following graph of Wikipedia’s growth, ESR’s prediction seems to have been spot on: Wikipedia really started to take off between 2003 and 2005.
Wikipedia is interesting both for what it is—the free encyclopedia that anyone can edit—and for how it is managed: openly. Not only can anyone edit Wikipedia, anyone can edit the Wikimedia Foundation itself. That’s new. That’s big. It’s a best-of- both-worlds mashup between a corporation and a government. It’s nominally a charity, but the 501(c)(3) is a red herring: the preponderance of Wikimedia’s mass is in the community of people who are not on WIKIMEDIA FOUNDATION INC's payroll.
Now, imagine if Jimmy Wales had Steve Jobs’ taste. Imagine if Wikipedia weren’t painfully ugly. Imagine if the editing experience were delightful. Yes, there is Project Athena. To be honest, early signs (PDF) leave me underwhelmed:
Imagine if the GitHub team were building Wikipedia. The tagging interface would be awesome, if they decided we needed it at all.
ESR is prescient here as well:
I believe the problem for 2001 and later is whether we can grow enough to meet (and exceed!) the interface-design quality standard set by the Macintosh, combining that with the virtues of the traditional Unix way.
However, he turns out to have been too optimistic:
As of mid-2000, help may be on the way from the inventors of the Macintosh! Andy Hertzfeld and other members of the original Macintosh design team have formed an open-source company called Eazel with the explicit goal of bringing the Macintosh magic to Linux.
ESR actually speculated that Apple itself might “go open” with Mac OS X, because in his view “open-source peer review is the only scalable method for achieving high reliability and quality.” Well, I daresay that Google and Amazon and Facebook have demonstrated this to be false. These companies depend on open source software, yes. But there is surely a great deal of proprietary work that is essential to their reliability and quality. And far from “go[ing] open” with Mac OS, Apple did the opposite. It wrapped FreeBSD in a heavy coat of proprietary code, and captured the lion’s share of the hacker laptop market. Ubuntu’s user experience is nothing compared to Mac OS’s. Apple then used the iOS variant of Mac OS to basically create the mobile market as we know it ex nihilo, boosting its consumer laptop sales in the process. Meanwhile, Eazel folded in 2001, and after a few more years trying to make good on ESR’s vision, Andy Hertzfeld went to Google in 2005, where he remains to this day. ESR’s poster children are now bywords and also-rans: Netscape, VA Linux, SourceForge, RedHat. Mozilla is basically kept alive by Google, for it’s own strategic purposes. And then Google went and built Chrome anyway. Sometimes it feels to me like open source is less of a revolutionary than a lapdog.
What happened? Why are our best people building Google and Facebook and GitHub, and not Wikipedia? Or rather: Why aren’t Google and Facebook and GitHub open projects like Wikipedia? Simple: It’s not yet economically rational for the world’s top talent to work on open projects. But I think it could and should be.
If I don’t like the direction Project Athena is going, I should get involved in the Wikimedia movement and “edit” it myself. Right? In a way, I see myself doing that with Gittip. I want Gittip to shift the economy towards open projects like Wikipedia. Why? Because I was raised in the bazaar. I love working together as free equals, rather than being someone else’s cog. Though chains be of gold, they are chains all the same.
Fifty years from now, I want most economic activity to happen in the open. That’s a dream. I want open companies, with open governments to follow.
If you want to see this too, consider joining Gittip, and funding someone who works on Wikipedia. While you’re at it, why not fund the people who build GitHub? Would they open up their company, if we made it economically rational? Maybe some of them would even go help out on Wikipedia.