Blockchain shows open source’s fatal flaw—and a way forward

Open source usage has skyrocketed, but not the number of developers working on projects. Those who benefit need to pay developers to keep it all going

Blockchain shows open source's fatal flaw—and a way out

“26,000 new blockchain projects last year!” screamed the headline. “But only 8 percent remain active!” The implication is that blockchain’s future is at risk, given the high mortality rate among its offspring. Yet nothing could be further from the truth. If anything, we need many more blockchain projects to fail to clear out some of the noise, leaving room for “Linux of blockchain”-type projects to remain.

And yet there is cause for concern, though not in blockchain specifically. Instead, the greater concern should be for open source, which has never been more popular with software users even as the developer population feeding it has remained flat. Unless we can find ways to encourage more contributions, open source efforts like blockchain threaten to crumble under the weight of user expectations unmet by developer productivity.

Open source is the perfect petri dish

Years ago, open source was an imitator, aping the best proprietary software. Today, open source tends to be where innovation happens, such as in machine learning, mobile, cloud computing, and big data. As Cloudera cofounder Mike Olson has declared, “No dominant platform-level software infrastructure has emerged in the last ten years in closed-source, proprietary form.”

He’s right, and it’s why Deloitte is equally correct to highlight, in its “Evolution of Blockchain Technology” research, that “open source could be the ideal petri dish for attracting a critical mass of blockchain coding efforts, talent, and overlapping objectives that accelerate an ecosystem with common standards.” Open source not only affords the freedom for blockchain developers to tinker, but also to collaborate.

And so they have.

According to Deloitte’s data, GitHub now boasts 86,034 blockchain-related projects, averaging 8,603 new projects every year—with an astounding 26,885 new projects in 2016. Although at first these projects were launched by individuals 99 percent of the time, the percentage of projects pushed by organizations has climbed to 11 percent. That’s the good news.

The bad news is that blockchain projects stick around for a mere 1.22 years, with 92 percent of those new projects in 2016 moribund by 2017. When a corporation is involved, the survival rate jumps from 7 percent to 15 percent, but the mortality rate remains high.

This shouldn’t surprise anyone. Open source has always been like this.

Open source projects are born to die

Twelve years ago, citing academic research, I wrote: “Open source projects ... tend to be small (82 percent suitable to one or two developers) and young. 60 percent of open source projects ... had been in development less than a year, 22 percent from one to two years, 15 percent two to three years, and around 2 percent more than three years.” In 2017, the same trends prevail, with two-thirds of all GitHub projects lucky to get one or two maintainers. Broadly developed and used open source projects like Linux and Kubernetes tend to obscure the reality of the reality of open source: Overwhelming numbers of tiny projects where a solo developer (and her sidekick) lost interest in maintaining and so abandoned them.

Why is it like this? Because ... people.

It turns out that developers are people, too—a point made by GitHub’s Nadia Eghbal in her excellent “Rebuilding the Cathedral” talk. One of the wonderful things about open source, she says, is its ability to deliver immediate feedback on the code you’ve submitted.

This is, however, also the terrible thing about open source. Find a project you like and contribute code, only to discover that “your contribution [is] lost in a sea of hundreds of unanswered issues and pull requests that are piling [up].” From the project maintainer’s perspective, “It’s fun at first and then the notifications start piling up so [you] start responding faster and then that leads to even more notifications,” resulting in “an odd productivity paradox.”

But this is a good problem, you insist. More contributions equals more good! Well, yes. But as Eghbal highlights, open source was a bit easier to manage when the total user population (measured imperfectly by SourceForge) was 200,000. Two decades later, it’s more like 20 million, resulting in a heck of a lot of notifications to filter.

Add into this the fact that though the number of users of open source software has gone up, the number of open source contributors has remained flat—all while the current crop of users tends to come from Fortune 500 companies. It’s become a mess worth Richard Stallman getting angry about.

Which brings me back to blockchain.

More filthy lucre is what open source needs

In general, more money in open source tends to bring more good. Not always, of course (OpenStack, anyone?), but on balance money is a good addition to an open source project. Why? Because money lets code committers remain committed to a project. Open source project maintainers have been coping with the increasing burdens on their time by limiting how much they invest in their projects, among other coping strategies. Turning open source into a developer’s full-time job offers a better way.

Eghbal calls this a return to cathedral-style development, arguing that the bazaar model mythologized by Eric Raymond back in 1999 could never work once open source hit scale. In the bazaar model, developers flock to projects for fun and contribute copious quantities of free time to making the world a better place. Meanwhile, back in the world of reality, documentation needs to get written, unsexy code needs to be developed, and companies are betting their businesses on it. It’s a long-term project, like building a cathedral.

As Eghbal puts it, despite the bazaar myth, “There is no bazaar. There are just a couple of people struggling to manage a bazaar-sized workload.”

So more people need to start contributing, and contributing more.

Blockchain’s biggest audience today is the financial services market. Let it start paying developers to contribute code full-time. Wal-Mart loves OpenStack, so maybe Wal-Mart should start contributing code—oh wait, it does! And so on.

Waaaaayyyy back in 2008, Red Hat CEO Jim Whitehurst urged enterprises to get involved:

The vast majority of software written today is written in enterprise and not for resale. And the vast majority of that is never actually used. The waste in IT software development is extraordinary. … Ultimately, for open source to provide value to all of our customers worldwide, we need to get our customers not only as users of open source products but truly engaged in open source and taking part in the development community.

In other words, we don’t need more software written. The software is already being written, but being hoarded behind the firewall. What we need is for enterprises to recognize that their own self-interest depends on giving that code to the open source projects upon which they increasingly depend.

This isn’t a request for charity: as Coinbase cofounder Fred Ehrsam posits, improving the Ethereum platform behind blockchain to make it easier to use could generate $3 billion in value. With billions in value at stake, it’s time for enterprises to feed their self-interest and pay more of their developers to contribute open source software—not merely use it.

Copyright © 2017 IDG Communications, Inc.

How to choose a low-code development platform