Dependencies and maintainers February 6, 2020 on Drew DeVault's blog

I’m 34,018 feet over the Atlantic at the moment, on my way home from FOSDEM. It was as always a lovely event, with far too many events of interest for any single person to consume. One of the few talks I was able to attend1 left a persistent worm of thought in my brain. This talk was put on by representatives of Microsoft and GitHub and discusses whether or not there is a sustainability problem in open source (link). The content of the talk, interpreted within the framework in which it was presented, was moderately interesting. It was more fascinating to me, however, as a lens for interpreting GitHub’s (and, indirectly, Microsoft’s) approach to open source, and of the mindset of developers who approach problems in the same ways.

The presenters drew attention to a few significant crises open-source communities have faced in recent years — left-pad, in which a trivial library was removed from npm and was unknowingly present in thousands of dependency graphs; event-stream, in which a maintainer transferred project ownership to an unknown individual who added crypto mining; and heartbleed, in which a bug in a critical security library caused mass upgrades and panic — and asks whether or not these can be considered sustainability issues. The talk has a lot to dissect and will frame my thinking and writings for a while. Today I’ll focus on one specific problem, which I called attention to during the Q&A.

At a few points, the presenters spoke from the perspective of a business which depends on up to thousands of open-source libraries or tools. In such a context, how do you prioritize which of your thousands of dependencies requires attention, for financial support, contributions upstream, and so on? I found this worldview dissonant, and asked the following question: “why do you have thousands of dependencies in the first place?” Because this approach seems to be fast becoming the norm, this may seem like a stupid question.2 If any Node developers are reading, scan through your nearest node_modules directory and see how many of these dependencies you’ve even heard of.

Such an environment is primed to fail in the ways enumerated by this talk. Consider the case of the maintainer who lost interest and gave their project to an untrusted third party. If I had depended on this library, I would have noticed long ago that the project was effectively unmaintained. It’s likely that I or my peers would have sent patches to this project, given that bugfixes would have stopped coming from upstream. We would be aware of the larger risk this project posed, and have studied alternatives. Earlier than that, I would probably have lent my ear to the maintainer to vent their frustrations, and offered my help where possible.

For most of my projects, I can probably list the entire dependency graph, including transitive dependencies, off of the top of my head. I can name most of their maintainers, and many of their contributors. I have shaken the hands of these people, shared drinks and meals with them, and count many of them among my close friends. The idea of depending on a library I’ve never heard of, several degrees removed via transitive dependencies, maintained by someone I’ve never met and have no intention of speaking to, is absolutely nuts to me. I know of these problems well in advance because I know the people affected as my friends. If someone is frustrated or overworked, I’m right there with them trying to find solutions and correct the over-burden. If someone is in dire financial straights, I’m helping them touch up their resume and introducing them to companies that I know are looking for their skillset, or helping them work on more sustainable sources of donations and grants. They do the same for me, and for each other.

Quit treating open-source projects like a black box that conveniently solves your problem. Engage with the human beings who work on it, participate in the community, and make it healthy and sustainable. You shouldn’t be surprised when the 3 AM alarm goes off if the most you see of a project is a line in your package.json.

  1. And strictly speaking I even had to slip in under the radar to attend in the first place — the room was full. ↩︎

  2. If so, you may be pleased by a Microsoft’s ridiculous answer: “we have 60,000 developers, that’s why.” ↩︎

Articles from blogs I read Generated by openring

Status update, May 2024

Hi! Sadly, I need to start this status update with bad news: SourceHut has decided to terminate my contract. At this time, I’m still in the process of figuring out what I’ll do next. I’ve marked some SourceHut-specific projects as unmaintained, such as…

via emersion May 21, 2024

Automatic case design for KiCad

I don't generally get along great with CAD software with the exception of KiCad. I guess the UX for designing things is just a lot simpler when you only have 2 dimensions to worry about. After enjoying making a PCB in KiCad the annoying for me is alwa…

via BrixIT Blog May 15, 2024

The floor is lawa!

And now for something completely different… When was the last time you were excited about a simple window with nothing but a single background color? Well, I currently am. Let me tell you about it… This window is notable, because it was created using the ”pu…

via blogfehler! May 8, 2024