Intel has been active in open source for decades, but (like the rest of the industry) we’re still learning and improving. Lately we’ve been taking stock of the whole company picture on open source contributions so we can contribute more effectively across all the communities we’re part of – from Linux to cloud-native infrastructure to machine learning and artificial intelligence.
Recognizing Open Source Sprawl
Intel’s developers are active across thousands of upstream projects, and we’ve started quite a few projects of our own. Some of those projects are active today, with clear governance and participation guidelines. Others… are not.
The lack of friction and barriers to creating new open source projects on sites like GitHub makes it easy to engage in open source sprawl. It’s easy to spin up a repository and commit code, which is good, but the tooling that makes it easy to launch doesn’t enforce maintenance or communications. That’s less good.
While more code under open source licenses is almost always a good thing, we know that code is just part of the picture. What’s really valuable is healthy open source communities that users, developers, and organizations can participate in and plan against.
That means, for example, having a roadmap and contribution guidelines. It means explicit governance that lets contributors understand the rules of the road for a given project. Not just the mechanics of submitting a patch, but what it means to contribute – how one moves from casual contributor to being part of the community. Who decides what’s accepted and what’s not? How do releases happen (if they do) and at what cadence? How are vulnerabilities handled?
Lots of questions –if only there was a standard yardstick for measuring projects! Turns out, there is: The Community Health Analytics Open Source Software (CHAOSS) metrics.
Using Metrics to Measure Community Health
CHAOSS, a Linux Foundation project, has defined categories of metrics that we’ll be using to track the health of our open source efforts. With, literally, hundreds of projects to watch, we need a streamlined way to observe and track progress.
But, to cut down on open source sprawl, we’re also looking at these metrics as a way to prepare for success and ask whether we should be starting a specific project to begin with. If you look at the metrics and work backwards, it gives you an idea whether there’s a feasible plan to succeed.
If you want to see contributor growth, do you have a full plan for attracting, empowering and retaining contributors? CHAOSS metrics ask those questions and help provide a strong frame of reference for assessing the health of projects.
We don’t want to inhibit participation in open source or hold back great software from being released as open source. But we do want to participate thoughtfully and with consideration for the larger community. That means having a clear picture of what success looks like, what resources might be required – and how we meet users and contributors who’d like to see those projects succeed as well.
Archiving Inactive Projects
Right now, Intel is looking at the repositories our developers have spun up on GitHub to see what projects are still active and which ones are no longer maintained and are unlikely to be revived.
Soon, we’ll be setting a number of projects to a read-only state and making sure they have a clear message that these are considered “archived” by Intel. Because they’re open source, the code is still there for anybody who’d care to examine or work with it under the terms of its license. We just want to be really clear to the community when there’s no expectation of future development, updates, security announcements or any other activity you’d expect from an active project.
We’ll also be working with active projects to make sure they clearly communicate their governance, how contributions are accepted (if they are) and the terms and norms for contributing.
Watch this space for updates on Intel’s open source efforts and leave a comment here if you have thoughts or questions.
About the author
Joe Brockmeier (jzb) is the Director of Open Source Community Architecture in Intel’s Open Ecosystem group. He first started working with open source technology in 1999, was the first openSUSE community manager, helped to establish Apache CloudStack as a Top-Level Project in 2013, and has contributed to CentOS, Fedora, and a number of other projects. He lives in Durham, North Carolina with his family, a menagerie of cats and dogs, and a superb music collection.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.