The software community as a whole is becoming increasingly more dependent on the work relatively small collection of open source software (OSS) developers. This wonderful group of people who give their free time to write that software is increasingly burning out doing work that everyone benefits from. Companies of all sizes, whether directly or indirectly, are profiting from the value they get out of this OSS. The developers themselves, however, rarely get anything in return for the value they are providing to these companies. This should change, but the question of how has remained unclear.
I’ve spoken with many OSS developers and users about their opinions on this topic. The general outlook is bleak. Everyone sees the issues, and no one sees a good solution. It’s a hard problem! Most of the developers producing widely used software still don’t have access to a generally approachable and successful business model. For users, there’s very little incentive to donate one’s hard-earned money for something that doesn’t cost money.
What makes things even harder for these developers is, as Andrew Nesbitt puts it in this excellent talk:
Maintainers are working mostly in the dark
This is an important point. Package maintainers have little insight into how their packages are used, which makes their job difficult. The maintainers typically don’t know which platforms their users are on, which packages are used together with theirs (and with which versions), what errors might be happening, or which features of their packages are actually getting used. Github issues alone are a noisy and still incomplete picture to the actual user base. Some of the most popular OSS projects learn about their users through an annual survey. It’s notable if not ironic that the group who’s life’s work essentially boils down to automation still prefers manual data collection in this way.
As Nesbitt points out, this is a largely solved problem in other software domains, especially the web. Tools like Google Analytics have made it simple and easy to understand who is coming to your website and how they use it. The analog for open source packages varies by the community but is generally still much less developed. Some package managers like Homebrew actually do collect installation statistics, as do npm and others. Unfortunately, the data they collect is still limited, and whether the data is even exposed to package maintainers directly varies between package managers.
So what can we, the software development community do about this?
One thing we could all do would be to convince our employers to support the awesome people that build and maintain the OSS they rely on. But you knew that already. One free thing we can do is:
Build, use, and tolerate OSS that reports usage data back to the maintainers. Not everyone can financially support OSS, but anyone can report usage statistics. This would help maintainers be better informed about their software. In a world where it is not uncommon for a 3rd party library to break your code in a patch release, having well-informed maintainers benefits everyone. It will help you deliver better software, and it will improve all the other software you use.
Today, Scarf is rolling out a new library to help with this exact problem, targeting the node ecosystem.
For Maintainers
As a package author, simply adding a dependency on Scarf will kick off data reporting anytime your package is installed, giving you key insights that can help make your life just a little bit easier.
- Which packages are commonly used along with yours?
- Some insight into your users:
- Usage trends (Is my package gaining popularity or losing it?)
- Operating systems (Did the newest release break Windows?)
- Country (What language should I consider translating the README to?)
- Company (Which companies might be interested in supporting my work?)
The library itself has zero dependencies, transparently logs everything it’s doing, and logs on a best-effort basis so it will never interrupt the user. If you configure Scarf to be opt-out by default, you won’t risk collecting data from anyone who didn’t explicitly give you the OK.
For Users
If you are a user of a package that’s using Scarf, you can feel good about knowing that you are doing your part to support the amazing people who build the OSS you rely on. Now there’s a concrete way to support OSS developers without paying money!
Concerned about privacy? While Scarf keeps data as safe as possible with up to date best-practices, it is certainly understandable that reporting data is not appropriate in all situations. You might be against it altogether. That’s ok! Whether the package authors choose to use Scarf with users opted in or out by default, you can enable or disable Scarf analytics easily. Your data will not leave your computer, and you can still be happy that the software you use can be more effectively maintained.
I hope that by giving open source developers standardized tools for safe data collection and analytics, we will have a community of better informed, less burnt out maintainers. A general movement in this direction would help both makers and users of open source! Whether or not you yourself are willing to report data, being supportive of developers simply asking for usage data can help the situation. No matter what your relationship with open source software is, there’s something you can do to make things better for everyone!
—
If you’re interested in adding Scarf to your npm package or learning more:
- Scarf homepage: https://scarf.sh
- GitHub: https://github.com/scarf-sh/scarf-js
- npm: https://www.npmjs.com/package/@scarf/scarf
UPDATE: a more recent post which expands on some of these ideas: https://about.scarf.sh/post/analytics-and-open-source-sustainability