Bridging the gap: Viewing your project through a downstream packager's lens

Have you ever published a project under an open-source license? If so, do you know if it’s used downstream? Has it been packaged to a Linux distribution? Who brought it there, and who maintains it? What are the needs of someone integrating your project into their ecosystem - be it an operating system or an environment manager, like conda? Have you ever wondered how projects appear from that perspective? I won’t be offended if the answer is: “Well, no, why would I?”

That’s the blind spot I’ll address in my talk. I’ll share the unique perspective of a person who has worked with thousands of Python projects and integrated them into an operating system. You’ll learn:

• Which project features make a packager’s life easier - and which make it harder,
• Whether tests are a blessing or a curse (and when),
• Why we care about the package metadata and included contents so much,
• How to simplify the process for others to contribute to your project,
• What are the challenges package maintainers face,
• And how downstream packagers can enrich your project. My talk will bridge the points of views of both the project authors and downstream integrators, and will offer a set of good practices you can consider adopting.

About me: In my job, I integrate over 4000 Python packages into Fedora Linux with every new Python release, from the first alpha to a final version - a process that spans an entire year. I’ve seen a lot. About you: If you’ve ever written a small application or library, packaged it and published it on PyPI, or even considered doing it, you’re the target audience of this talk. No specific prior knowledge is required beyond creating Python projects and publishing them via version control systems.