Swapnil Bhartiya | 25 September 2017
The Linux operating system was created some 26 years ago by a young Finnish engineer, and it now powers the global economy. Not only has Linux survived for more than quarter of a century, it continues to grow its influence and dominance.
Not all open source software projects thrive, however; many promising projects die untimely deaths. So, what’s unique about projects like Linux that thrive where others fail? What’s the secret sauce that sustains one project over others? Is it the community? The license? The code? The organizations backing it?
We talked to open source veteran Brian Behlendorf, co-founder of the Apache Software Foundation (ASF) and current Executive Director of the Hyperledger project, for some answers to these questions. Here is an edited version of the interview conducted at Open Source Summit North America in Los Angeles.
What are the core components of sustainable open source projects?
Brian Behlendorf: By definition, any open source project that is still alive needs some critical mass of developers contributing to it. The Linux kernel is 25+ years old, and it still sees 5,000 new lines of code every day. It’s still such an incredibly active project.
In my book, that means you need this body of maintainers and contributors who are willing to continue to nurture the project even as it goes into adolescence and later life.
For me, the only way to more or less guarantee that happens is to see that there are companies out there who are making money off of open source software. They have embedded it at the core of their business. And even if it’s not what they do as a business, it’s still something that they need. So they’ll provide feedback, contribute, and continue to invest in shepherding it forward.
So, having companies use and contribute to your project and in return inject resources does help. What role do non-profit organizations like The Linux Foundation and ASF play?
Behlendorf: What The Linux Foundation, I think, has figured out, is how to identify these technology spaces, bring companies together around them, and then help them make money from it and profit from it.
But it’s not the only viable model. The Apache Software Foundation model is entirely volunteer driven, with developers even doing things like running the books or doing marketing.
There’s an incredibly empowering side to that, but it doesn’t always work. There weren’t enough developers who showed up around OpenOffice, for example, for that to work for the Apache OpenOffice community.
It’s almost hard to say if any model is better than the others. They’re all very unique for the kind of software being built and the developers who are attracted to that software.
You talked about commercialization of open source, yet we have seen that some open source communities are averse to the idea of any commercial or corporate links.
Behlendorf: I don’t think there was really ever a truly long tradition of a battle between open source developers and commercial interests. I think many of the people I know who were contributing to open source even before me were building businesses on top of it. Michael Tiemann built Cygnus on top of the GNU compiler suite. So this template, and every ISP, every web business is building on top of open source web components.
I think the real battle might have been between proprietary software and free software. And the real question was, did we need to vanquish proprietary software in order for free software to flourish?
Do licenses play any role in sustainability of open source projects?
Behlendorf: I tend to think of companies that have played games with licensing. There’s not a lot of successful examples out there. Why don’t we just put these kind of games to the side? Let’s build the software we need together, and go out and build great applications and great websites and great other things on top of that.
And this is what we carried forward in the Hyperledger project as well. All the Hyperledger code is under an Apache license. All of it is designed to be embedded inside of other people’s products and services.
We want to see lots of cloud hosts running Hyperledger technology. We want to see a lot of application developers embedding this inside and, say, putting it inside of cars or IoT sensors or those sorts of things. The less time that we have to spend with lawyers and with MBAs explaining to them how and why they can make money with this code, the better off we all are.
Diversity is necessary for the survival of organisms, can the same be said for open source projects?
Behlendorf: If your community doesn’t look like the global community, then something’s wrong.
The blockchain movement is a great example of diversity. India and China and Europe have been running as fast with this technology as anybody in the United States. We are constantly looking at what countries are we visiting. Where are our companies based? How do we go and empower those companies in a country like China or a country like India, to go and be champions of what they’re doing, of the technology that they’re building?
What about culture?
Behlendorf: I’d say the final thing I’d throw out about sustainability is if your project isn’t comprised of people who are nice to each other, it’s not going to be very sustainable. Even the smartest people, even the most enthusiastic people will burn out if the dynamic in the community is very harsh, or if every time a good idea is brought up you hear crickets or somebody talks it down. You need to be nice to each other on an open source project in order to have any hope of being sustainable.