May 29, 2012
NOTE: Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words. You are warned, people.
Drupal is an open source content manager, Dries says. In the past twelve years, Drupal has “grown significantly”: 71 of the top 100 universities use it, 120 nations use it, the White House uses it, 2 of of the 3 top music companies use it, the King of Belgium uses it. [Dries is Belgian :) ] The NY Stock Exchange is converting from a proprietary Java solution to Drupal. Five of the 6 top media companies use it. One out of 50 wesbites run on Drupal. Drupal has 10,000+ modules, 300,000 downloads a month, 1.5M unique visitors a month at drupal. org. And it’s free as in beer.
Today he’s going to talk about: history, open source, community, the evolution of software, and how to grow and sustain it.
Dries began writing Drupal in his dorm room, more or less by accident. He wrote a message board for the Linux project, in part to learn PHP and MySQL. About a year later he released Drupal 1.0 as open source, as “a full-featured content management/discussion engine…suitable to setup a news-driven comunity or portal site similar to kuro5hin.org and slashdot.org” (as it said in the original annoucement). “It took me about 30 seconds to come up with the name Drupal, a terrible name.”
Three years later (v.4.1) he says it still looked “pretty crappy.” Two years laer,in 2005, 30 develoeprs showed up for the first DrupalCon, in Antwerp. There are now several year. By 2011, it was looking quite good, and 3,200+ developers showed up at DrupalCon. There are now weekly meetings around the world.
There were growing pains, he says. He tells us about The Big Server Meltdown. In 2004, the servers failed. Dries put up a blank page with a PayPal button to raise $3,000 for a server. Within 24 hours, they’d raised $10,000. One of the CTOs of Sun shipped him a $8,000 machine. Then Open Source Labs in Portland OR offered to house the servers. “That’s just one anecdote. In the history of Drupal, it feels like we’ve had hundreds of these.” (There are currently 8 staff members. They organize conferences and keep the servers up. )
But, Dries says, this shows a weakness in open source: you suddenly have to raise $3,000 and may not be able to do so. That’s a reason he started Acquia, which provides support for Drupal.
Drupal is open source: It’s gratis, anyone can look at the source code, they can modify the code, and they can share it. The fact that it’s free sometimes let’s them win bids, but open source “is not just a software license. It’s a collaboration model.” “Open source leads to community.” And “ultimately, that leads to innovation.”
Dries shows photos of the community’s embrace of Drupal (and its logo). “Drupal is successful today because of the community.”
Q: How do we know there will be enthusiastic support a few years down the road? How do we know it won’t have a Y2K problem?
A: There isn’t an easy answer. Things can go wrong. We try to keep it relevant. We have a good track record of innovation and keeping the right trends. And a lot of it comes down to keeping the community engaged. We have a large ecosystem. They volunteer their time, but the are all making money; they have an economic interest in keeping Drupal relevant.
“Drupal doesn’t win just because it’s cheaper. It wins because it’s better.” It is technically superior because it has thousands of developers.
Evolution of software
Dries points to a common pattern: From innovation to bespoke systems to products to commoditization. In each step, the reach becomes wider. Proprietary software tends to stop at the products stage; it’s hard to become a commodity because proprietary software is too expensive. This is an important opportunity for open source.
Growing large projects
Is Drupal’s growth sustainable? That’s a reason Dries founded the Drupal Association, a non-profit, in 2006. It helps maintain drupal.org, organizes events, etc. But Drupal also needs companies like Acquia to get it into new areas. It needs support. It needs people who can talk to CIOs in large companies.
Open source Joomla recently hired some developers to work on their core software, which has led some of the contributors to back off. Why should they contribute their time if Joomla is paying some folks? [Joomla's experience illustrates the truth of the Wealth of Networks: Putting money into collab can harm the collab.] Drupal is not going to do that. (Acquia develops some non-open source Drupal tools.)
IBM and RedHat are the top contributors to Linux. What companies might make that sort of strategic investment in Drupal? Instead of one or two, how about hundreds? So Dries created “Large Scale Drupal,” a membership org to jointly fund developments. It’s new. They contribute money and get a say in where it’s spent. The members are users of Drupal. E.g., Warner Music. Module developers can get funded from LSD. Two people run it, paid by Acquia. There has not been any pushback from the dev community because there’s no special backdoor by which these projects get added to the Drupal core. In fact, the money is then spent to fund developers. Dries sets the technical roadmap by listening to the community; neither the Drupal Association or LSD influences that.
Of these collaborative projects often start as small, volunteer-driven projects. But then they become institutionalized when they grow. Trade routes are like that: they were originally worn into the ground, but then become driven by commercial organizations, and finally are governed by the government. Many others exhibit the same pattern. Can open source avoid it?
If you’re thinking of starting an open source commercial company, you could do dual licensing, but Drupal has not made that choice.
Q: How much does Drupal contribute to the PHP community?
A: A little. There are tribes: some are active in the PHP tribe, others in the Drupal tribe. It’s unfortunate that there isn’t more interaction. Dries says he’d love to grow Acquia enough so that it can put a couple of people on PHP, because if PHP isn’t successful, neither is Drupal.
A: We don’t have a lot of decision-making structure. I’ve always been opposed to formal voting. We work through discussion. We debate what should be in the core. Whoever wants to participates in the debate. Ultimately we’re structured like Linux: there are two people who are committing changes to a core version of Drupal. For every major version I pick someone to work alongside me. When we release the version, he or she becomes the maintainer of it. I move on to the next version and select someone to be my co-maintainer. The 15,000 modules are maintained by the community.
Q: Do your biggest contributors agree to programming standards?
A: We are strict about our coding and documentation standards. I make the final decisions about whether to accept a patch. Patches go through a workflow before they reaches me.
Q: What advice would you give to someone trying to attract people to a project?
A: If people can make money through your project, it will grow faster. We built a community on trust and respect; we make decisions on technical merit, not dollars. We have a darwinian model for ideas; bad ideas just die. See what rises to the top. Include it in the next version. Then put it into the core, if it’s worth it. The down side is that it’s very wasteful. I could tell people “If you do x, it will get in,” but I try to get out of the way. People have taken Drupal in sorts of directions, e.g., political campaigns, elearning platforms, etc.
Q: [me] How important are you to Drupal these days?
A: I think I’m more important as the face of Drupal than I used to be. In the governance sense I’m less important. I was the lead developer, the admin for the servers, etc., at the beginning. The “hit by a bus factor” was very risky. Nowadays, I don’t write code; I just review code. I still have a lot of work, but it’s much more focused on reviewing other people’s work and enabling them to make progress. If I were to die, most things would continue to operate. The biggest pain would be in the marketing . There are a lot of leaders in Drupal. One or two people would emerge or be elected to replace what I do.
Q: What’s hard for Drupal?
A: One of our biggest risks is to keep nimble and lean. It takes longer to make decisions. We need to continue to evolve the governance model to encourage us to accelerate decision making. Also, we have some real technical issues we need to address, and they’re huge projects. Volunteers can only accomplish so much. LSD is perfectly positioned to tackle the hardest problems. If we did it at the pace of the volunteers, it would take years.
Date: May 29th, 2012 dw