September Python Leiden meetup summaries¶

Maintaining Python packages to attract free and open source software contributors - Steve Piercy¶

Steve is in the Netherlands for next week’s Plone conference in Maastricht (NL). He’s been involved in open source software for over two decades.

What are contributors? Contributors to open source software? Why do you do it? Learning (I myself got a good programming education out of contributing to Zope/Plone while still at the university). You might want to give back. You might like the community, you like to belong to the community. Finding your tribe. Perhaps build a resume or CV. Help other people.

Why not? Negative online reactions. You want to really make money out of it. AI slop. You might not have time. Too busy. Life happens. You might enjoy walking your dog or doing gardening more. Or perhaps you just don’t know how! Or you don’t feel experienced enough (imposter syndrome). Perhaps you don’t think your English is good enough (he advocates just using Google Translate).

(He asked how many people had already contributed to open source: about half the room raised their hands.)

There are actually lots of ways to contribute. It is not all code! Read this page for some examples. You can report security vulnerabilities or issues. You can comment on issues and help getting them forward. Review pull requests. Extend the documentation. You can join discussions in forums or on Matrix. Write a blog post on a project. And… you can financially sponsor a project.

(He again asked who had contributed: now most hands went up!)

Where to start? Start with what you use and what you find interesting. This helps staying involved and getting into the community more and more. He himself liked web applications and especially forms: making sure they’re intuitive and safe and well-validated. And he liked documentation, so some of his first contributions were documentation fixes for open source projects.

Recently, he volunteered to take over maintainership of an open source project (a sphinx extension). There were lots of open pull requests and he asked the owners to look at it again and perhaps tweak it a bit and… everybody responded and a few weeks later he could make a new release with lots of fixes.

What is a maintainer? Well, basically maintainers are contributors who make releases. But ideally you also have to “tend your garden” and try to get your small community to thrive. One thing to keep in mind: you have to put your ego aside. If it is open source, it is really owned by your users, not by you. What do you have to do? Show up when someone has a question. Document your project (how to contribute, how to report bugs and security issues, etc.).

He especially mentioned https://djangonaut.space , that’s a great initiative to help people get involved with contributing. Same with https://djangogirls.org/ .

As a maintainer, you might also have to look for funding. There are options like “github sponsors”. But also NLnet (from the Netherlands) and the German sovereign tech agency that sponsor lots of projects.

A best practice of good maintainers is to have good tooling. Automatic tests + coverage. Build documentation. Formatting checks. zest.releaser for good releases.

The most important part of being a maintainer: how do you treat people.

There’s one big problem: AI. There are just too many pull requests and too many issues. You can never get through them if your project is reasonably popular. Dealing with it is hard. Several projects made the choice to disallow AI contributions. Also read https://leidendeclaration.ai/ , for the same problems are entering math and science, too.

But on the other hand… in the icalendar project that he helps maintain, they got a new contributor that made his first contribution based on AI. He liked that: the person got enough confidence through AI. They now have an AI guideline in their contributor guide now.

There’s are automated tools on github that can help you check pull requests and identify slop-generating accounts, for instance. A tip when confronted with suspected AI work: ask vague, clarifying questions. Just ask for a bit of clarification in a vague way: AIs can’t respond to vague questions, but humans can.

“Once, men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.” - Frank Herbert, Dune. Funny that he could write that 50 years ago.

Something mentioned in the discussion: people are now bidding on real github accounts in order to let AI bots use accounts that appear real…

Watch out with AI. One of the links he shared: https://pivot-to-ai.com/2026/09/08/students-who-dont-use-ai-are-ahead-on-every-measure/ . And, look at the kids: it was bad enough with social media, but AI is much worse. And freelancers are getting buried with soulless AI slop cleanup: artists get to clean up AI-generated images, for instance.

Choose wisely. A quote he wanted to give us to think about: teach the world you want.

Monitoring my washing machine - Michiel Beijen¶

Michiel has a 2021 Samsung washing machine that plays Die Forelle by Schubert when it is finished washing… But the washing machine sits in the garage so he can’t hear the sound. It is a smart machine, so there’s an iphone app for that! “SmartThings”. But the list of data it wants to collect about you is horrendeous. And the app is almost 1GB! No…

There’s also “home assistant”, written in python. There’s even a SmartThings API connection. But… from October 2026 onwards you need a $5/month samsung subscription!?!

There’s also “Matter”. They call it vendor neutral, standards compliant. Samsung supports it. But only for connecting Matter devices to SmartThings, not the other way around.

Next try: put a small camera in front of the washing machine’s display, add a bit of OCR text recognition to detect the remaining time from the image and add an http interface to it. An old webcam couldn’t provide a good camera. A raspberry Pi Camera 3 did the trick. (He later thought it would perhaps have been better to pick some old smartphone.)

He showed some graphs detailing how the washing machine estimates the time it will finish, including the adjustments it made halfway (adjusting for load or so).

Another approach that a colleague of him took: use a monitoring plug that measures the electricity used. Once the electricity usage drops off, the washing is finished.