Open source is all about experimentation and iteration, which is why a 98% failure rate may well be the best sign of its success.
Though open source has boomed over the last 10 years, the success rate of individual projects has not. If anything, it has cratered.
Back in 2005, researchers analyzed FreshMeat projects and discovered that 90% of the projects hadn't been actively developed in the previous six months, and 83% of projects weren't able to interest more than two developers. Today, these numbers have gotten worse, with more than 98% of all projects on GitHub not seeing any development beyond the first year they were written, as Redmonk's Donnie Berkholz found.
Has open source failed?
Exponential open-source growth
While the percentage of "failed" open-source projects has remained constant, one thing has not: the total number of open-source projects has skyrocketed, as CA's head of developer relations, Ian Kelly, points out. With this in mind, if we simply count the total open-source software in existence, open source has been a tremendous success. Back in 2008, researchers Dirk Riehle and Amit Deshpande found that "every year (plus a month or so), both the total number of projects and lines of code double."
That is truly exponential growth.
Since their paper was published, open source has roughly maintained this pace, doubling every 14 months or so. This has been true, regardless of the platform. For example, Microsoft's Codeplex has seen hugely impressive growth, despite Microsoft's historical antipathy toward open source.
In fact, this has been one of open source's most important traits: it evolves. Once the domain of Sourceforge-loving hackers, open source has embraced a wide variety of platforms, including a critical shift to mobile platforms, as Black Duck Software reports (Figure A):
Cumulative open source projects.
In a world where it's a given that most projects will fail, it's exceptionally important that the volume and breadth of open-source projects has boomed.
As ever, the trick is to remember that "failure" is the norm in open source. That's the source of its greatest beauty, in fact. In open source, we celebrate the principle that "given enough eyeballs, all bugs are shallow," as well as the spirit of iteration. Open source isn't about creating one pristine project. It's about experimentation.
Such experimentation depends upon a culture that encourages new projects, even if they individually seem unlikely to prosper.
The reality is that most open-source projects never get any outside code contributions. Most fail. This has always been the case. For every Drupal, with its incredible community involvement, there are 98 other projects that are an abandoned labor of love by one or two developers.
And that's OK. In fact, that's exactly why open source, as a whole, succeeds.