When the Cathedral Owns the Bazaar

The classic "cathedral and bazaar" metaphor hardly holds up to reality. What is the WordPress project?

collage of computer devices and the wordpress logo

WordPress has focused it’s attention recently on the Five for the Future campaign, the banner name for all open source contributions made by the community at large. A few recent posts by project leadership has begun to examine what it means to actually “contribute” to a project that not only includes core software, but tons of other moving parts including design, documentation, education, translation, and more. We saw the need for these contributions expressed this past weekend at WordCamp Europe as contributors from some of these teams pressed the project leadership for more resources and attention.

Since writing about the “Notifications” feature project, I’ve been trying to find ways to contribute to WordPress beyond simply writing about it. When you begin the contributors journey, they suggest taking a few courses at Learn WordPress that teach the fundamental philosophies behind open source as well as general etiquette for collaboration.

When I took the Open Source Basics course, much of the material was drawn from the seminal playbook of open source: The Cathedral and the Bazaar. If you haven’t read it, it’s essentially a collection of insights into open source, framed around a developer’s experience contributing to, and ultimately inheriting, an open source library. It was written in the mid-nineties, so while some of the technical digressions feel antiquated, it actually exists as a very strong time capsule of the early internet ethos. I’d recommend reading it (online for free) if you haven’t.

The central metaphor of the book is the contrast between the cathedral and the bazaar. The cathedral represents top-down, proprietary software development, the sort of thing we see from Apple or Adobe. The cathedral is privately controlled and only released to the public once it is considered complete. The bazaar is open source software. Anyone can join the bazaar and contribute a piece to the larger whole. The bazaar is continuously improving and considers itself more secure and less buggy due to the public scrutiny of its source code.

The book was especially prescient with its command to “release early and often,” something we’ve seen become pretty much the standard for everything from software subscription models to Gutenberg to the bigger developer previews from today’s cathedrals. Similarly, pithy morals such “a truly great tool lends itself to uses you never expected” have proven beyond true for the platform model in general.

The reality, though, is that the overarching “cathedral and bazaar” metaphor hardly holds up to reality. Sure there are still cathedrals (Apple being perhaps the most ubiquitous example), but what about the bazaar? Most open source projects these days are open source in terms of license and code freedoms, but not in terms of contributions, funding, or leadership. React is open source, but is owned and heavily managed by the Facebook cathedral. Google’s AMP is technically open source and was potentially one of the most damaging ideas for the open web in recent history. And what about WordPress?

If I were to extend the metaphor further, I’d say WordPress is not a bazaar, even if it’s not really a cathedral. It’s something in between. WordPress is a shopping center.

The shopping center is not a singular entity, like the cathedral. Its populated by distinct merchants, and theoretically anyone with means could show up and contribute. But most of the real estate is leased by major players, and the department store, the grocery chain, or the Target/Walmart effectively control the landscape. Even further, the whole thing has an owner and a property manager who make the ultimate decisions, and they’re also dealing with regulations and permits and profit margins and god-knows what else. Contrary to this idealized bazaar, you don’t just pull up to a shopping center and open up shop, there are hoops to jump through.

In WordPress, companies like Automattic, Yoast, and GoDaddy dominated the contributor space for the latest release (6.0 as of this writing), alongside a handful of major development agencies. There’s also a very small number of extremely prolific contributors with an unclear sponsorship status and a very long tail of individuals making single contributions. But the idea that WordPress is created by a loosely affiliated band of misfits and volunteers is a nostalgic recasting at best. The decision-making behind WordPress comes from a company with over 2,000 employees that has raised almost a billion dollars of funding in it’s lifetime.

Sponsoring and paying contributors makes sense, because it’s never really free to contribute something. For starters, you need to have the free time and the knowledge. In recent years, with the shift towards JavaScript-first development, the buy-in costs have increased dramatically, especially as the complexity and bureaucracy of the project itself have naturally grown. Just like there are web development agencies who become more adept at navigating the bureaucracy of government RFPs than delivering quality projects, much of the future progress in WordPress will be left to those with the profit margins to pay full-time sponsors.

This is the central flaw in the “bazaar” analogy- there will always be gatekeepers and leadership in open source projects. It would be impossible if every single pull request was simply accepted and merged. At a certain point, any reasonable project needs leadership who will determine if your contribution is in the best interest of the project, and we can’t simply rely on the “worthy and widely shared goals [of] the project leaders and tribal elders”. While we all have our dream list of what the next release could be, the truth is that the aims of the project become more disparate as the project itself grows.

This leads to Matt Mullenweg’s recent comments at WCEU about WordPress market share (of which much ink has been spilled). As Matt correctly stated “market share is not really a goal, but it is a result.” This is very true, and I would agree that it’s probably not even an extremely reliable result to consider. Instead, I think something like an “ecosystem GPD” could be more valuable. Understanding how much economic value WordPress provides to users and developers would be infinitely more meaningful than spending years trying to steal market share from race-to-the-bottom competitors like Wix (who would even do that?).

However, Matt then went on to suggest that WordPress market share could simply be increased by migrating Tumblr’s user base and underlying infrastructure to WordPress, essentially juicing the numbers with pre-existing accounts. It was an off-the-cuff remark, but I think helpful in terms of how the current project leadership sees and values WordPress itself. Sure building Tumblr on WordPress would be “a big boost to the numbers”, but to what end? Unless this hypothetical Tumblr-on-WordPress could support the broader ecosystem of WordPress plugins, themes, and custom development, it’s unclear how this is meant to reassure a conference of WordPress developers that this would be something valuable to anyone outside of Automattic.

To horribly mangle the shopping center metaphor, how confident do you feel opening your boutique shop across the parking lot from a Target? Do you want to build your restaurant in a strip mall with three fast food joints and an Applebees? Mullenweg made references about page builders migrating to take advantage of Gutenberg’s structure, but I’m unaware of any actual examples of this happening. Instead, we’re seeing the souring of acquisitions, from Frontity being abandoned because of FSE’s complexity to popular plugins being bought and sold off.

Mullenweg also said that “Gutenberg may even be a larger contribution than WordPress itself” when discussing the use of Gutenberg in proprietary tools like Day One, the fantastic daily journal app acquired by Automattic. Again, while that sounds great for Day One users like me and Automattic in general, it’s unclear how this helps the goals of the larger WordPress project.

I’m a big fan of Five for the Future, and am onboard with finding more ways to promote it. I’d love to see more pressure to contribute placed on any development agencies with significant revenue from WordPress. And any attention to some of the other ways to contribute that aren’t as heavily sponsored (organizing and volunteering at WordCamps, for example) is a good thing. But As Joost de Valk pointed out in his comment to the original post on Five for the Future:

I’d say that for this definition to work, I think the part “and the project’s current big ideas” should be removed OR the process of creating these big ideas should be opened up to (a delegation of) the wider community. If not, as it stands, these decisions are made purely by people working for Automattic, and there’s no way for them to be prevented from being seen as for “the sole benefit of a company or individual”.

Joost de Valk

The language around these ideas is important, the metaphors and stories we tell ourselves about our work matter. To say WordPress is a “bazaar” is to make implications about it’s structure that may not actually be the case.

It’s pretty clear that any project needs leadership and decision-making. But there are a host of other features in WordPress that aren’t being worked on because many of us fear that our efforts will never be merged anyway. There aren’t enough contributors for the features that are a priority. And there are contributors who are literally begging the leadership for the basic technological necessities to make progress.

Open-source projects, when they founder, essentially never do so for want of machines or links or office space; they die only when the developers themselves lose interest.

Eric Steven Raymond, The Cathedral and the Bazaar

So next in my contributor journey, I’m taking the Learn WordPress course titled How Decisions are Made in WordPress. I’m pretty sure that this will answer all of my questions.

Author Profile Image

Brian is the Technology Director at Understrap and Howard Development & Consulting. Located in Southern California, Brian is a former college instructor and full-stack developer who brings his unique academic perspective to Understrap Academy. Brian is a graduate of California State Polytechnic University Pomona and California State University Fullerton. His work has included projects for Harvard University, The World Bank, and the Colorado Department of Public Health and Environment.

Subscribe & Share

If you liked this article, join the conversation on Twitter and subscribe to our free weekly newsletter for more 🙂

MasterWP contains no affiliate links. We’re entirely funded by the sponsors highlighted on each article. In addition to MasterWP, we own EveryAlt, WP Wallet, Understrap and Howard Development & Consulting.

Latest Posts