Leadership and Power in Open Source
Last week I waded into the Great WordPress Drama of 2024 with a post comparing Matt Mullenweg’s leadership of WordPress with Dries Buytaert’s leadership of Drupal. My conclusion was that Matt does not share power.
Besides Dries, another key interlocutor of Matt’s during this crisis is David Heinemeier Hansson, whose name I have yet to be able to spell confidently without checking, despite having followed his work and crossed paths a bit since 2004. DHH, let’s call him. Shall we?
Matt published a reply to DHH today [update: Matt took down the original], which sorta is-what-it-is, par for the course in the fight Matt is waging. What I found more interesting was this related post from 2019, about a conversation Matt had with DHH on the Rework Podcast entitled, “Open Source and Power with Matt Mullenweg.”
The context was Automattic’s Series D, and an emerging narrative around growing WordPress from 34% of the web to 85%. Today WordPress is at 57%, so Matt is right on track, I guess. At the time, David rankled at one company consolidating so much power. A Twitter exchange evolved into the podcast episode. tl;dr Matt really does not share power—except when he does, then he really does.
Evil Monopoly Stuff
There are some quotes in Matt and David’s call that are really juicy in light of the current crisis. Here’s the juiciest, at 13:32:
And the truth is that if I had an evil hat on—like let’s say I started doing evil monopoly stuff—people could easily fork the software, and they would … tomorrow! So there’s a checks and balances there.
David pushed back on the feasability of a fork against such a large incumbent, but today we have Valkey and OpenTofu as existence proofs that it can and will be done. Who among us will be surprised if and when Linux Foundation and WP Engine announce OpenPress or whatever?
Here’s the kicker: Any fork, especially if WP Engine backs it, will be a win for Matt. It means they’re finally contributing. Since it’s GPL their modifications are available to everyone else, including Automattic. And it’s what Matt has said all along. Here he is at 49:25:
I think in “the benevolent dictator,” which is an interesting sidebar, I really like the benevolent part. And again, I think on the dictator part, I don’t love that terminology. I don’t think either of us would be attached to that. But in other places, you know, companies, they typically have a CEO. Countries, they typically have presidents or something like that. And all of these systems are set up to have checks and balances, whether that’s a board or shareholders or voters or things like that. And if you don’t like your country for example, it’s really actually difficult to move to another one and opt into like a different governance system. But in software it is a hundred, a thousand, a million times easier than like uprooting your family and your friends and everything like that and moving to a different country.
Actually, Free Software Foundation is a good example of that. Myself and many others drifted away from them over the years. So, even though they started things, their influence got smaller and smaller and smaller as they took positions which didn’t really align with the future where people were going or where their constituents, be that developers, users, donors, whatever, were headed.
To the extent that Rails is doing great, WordPress, is doing great, et cetera, I think that we’ve been responsive to changes in the marketplace. I think that term started as a joke. There is something interesting that with a lot of these Open Source projects there’s almost nothing you can name on the list that has a committee-based leadership structure. And I think about that all the time. Like, why is it difficult for committees or an alternative governance structure to create really great software, backend software, frontend software. You typically have something more like a director of a movie or a conductor of an orchestra. There’s no perfect analogies because it’s software, it’s different. But you typically have a person with whom a great deal of decision-making power for determining the platform rests and that is often a good thing for the health and quality of the platform, and I don’t mind it as much because we have these checks and balances with Open Source forking or all sorts of different ways that people can opt out of that person’s power.
It’s hard for me to fault his main point. Personally, I can’t stand WordPress as a software product. It really frustrates me every time I have to use it. So I don’t, for the most part. That’s me opting out of Matt’s power. If I did love WordPress and didn’t like what Matt is doing, I would think about contributing to a fork. He’s right that Open Source provides a meaningful check-and-balance that, e.g., Mark Zuckerberg does not have. David finds this to be “uncharted and dangerous territory for open source,” but I’m not sure it’s not fine. Matt is being quite aggressive, yes. But WP Engine really was a “Taker,” in Dries’ terminology. Dries’ patient, thoughtful approach to resolving the Maker/Taker problem is inspiring to me. Matt’s tyrannical approach is not. But I’m free to opt out of Matt’s power, so what’s the problem?
In Matt’s reply to DHH [update: Matt took down the original], he called out David and Dries for holding the trademarks for Ruby on Rails and Drupal, rather than vesting those with their respective foundations. He has a point there, as well.
What I’m seeing here is a clarification of ways to lead in Open Source, ways to wield power. I’m grateful that I can opt out of people’s power in Open Source, and I trust that the best ways of wielding it will win out over time.
P.S. Yay For More Emotional Bandwidth
By the way, I love that we’re finally learning to move conversations from text-based online communication, which has low emotional bandwidth, to podcasts and video calls, which have higher emotional bandwidth. Just the other week I did two of these, with Peer Richelsen and with Pete from Hugging Face. I first experienced the power of this practice back in 2013, with DHH, in fact. We talked about mixing Open Source and money, which is still highly relevant.
The roots go back much further. The term emotional bandwidth is Mitch Kapor’s, from 1995. Mitch was wrestling with the tension between the promise of the early Internet to bring people together, and the reality of the flamewars and conflict that we ended up in, even back then. “The answer,” he wrote, “may lie in increasing the emotional bandwidth of communication.” But he could only dream of the future we’re living in today:
As telephone systems are upgraded and cable television becomes an access path to the Internet, higher speed connections capable of carrying voice and face will become much more common.
We’re pretty accustomed to dropping to a call in our work lives, inside our companies. It’s great progress that we’re starting to see more of this in public life as well. Hopefully someday national presidential politics will catch up. ;-)