Corporate Engagement and Contribution Guide
Along with welcoming contributions from individuals, we also welcome contributions to JupyterLab from corporations. Over a number of years of working with a wide range of corporations, we have discovered a set of principles and practices that enable these collaborations to be healthy, productive, and sustainable. These principles and practices are detailed here, along with a mental model that sets the stage.
We would like to offer the following mental model of open-source projects, set as a parable. This idea was first introduced by a core Python contributor, Brett Cannon, on Twitter:
Pull requests can be like someone trying to give you a puppy you didn't ask for; they mean well, but they can forget a puppy is a decade-or-more commitment…
Let's expand on this idea a bit more.
An open-source software project is like a big yard full of puppies. Puppies are amazing!!! Lots of people stop by to look at the puppies, play with them, take them on walks, and play fetch. These people are like the users of an open-source project. A few people, the maintainers, live in the yard with the puppies. The maintainers feed the puppies, take care of them, wash them, take them to the vet, get their shots, and keep the yard clean. Just to keep the existing puppies happy and healthy requires a ton of work and money.
Working in a yard full of puppies is CRAZY! At times, there can be a line around the block of people wanting to play with the puppies and talk, or complain, to the maintainers. Sometimes, people show up with more puppies. These are like pull requests to an open-source project. While the maintainers love puppies and would love to be able to have more of them for everyone to enjoy, each new puppy has to be cleaned, fed, trained, taken to the vet, etc.
Practices and Principles
The following principles and practices should be followed by corporations wanting to contribute to JupyterLab. These apply to both corporations as a whole, as well as individual contributors within those corporations.
Empower individual contributors. Encourage your employees to help at the puppy yard. In the daily development and design work of the project, everyone contributes as individuals. Because of this, if your corporation wants to engage with and contribute to JupyterLab, the primary means will be through allocating, empowering, and encouraging your individual employees to contribute in an ongoing manner. Put more bluntly, pay your employees to contribute to Jupyter.
Make balanced contributions. You can help take care of the puppies in a variety of ways. We expect all contributors to make balanced contributions that match their skill level and level of participation. Balanced contributions include work on new features, fixing bugs, reviewing pull requests, writing documentation, issue triage, answering users' questions, attending meetings, helping with releases, and other maintenance tasks. Balanced contributions build trust with the team and scale the human side of the project in a healthy and sustainable manner.
Help review pull requests. One aspect of this balanced contribution approach deserves special attention: reviewing pull requests. Submitting pull requests with new features is an exciting and high-profile way to contribute. However, reviewing such pull requests is probably the biggest bottleneck in the development of open-source software. If you want your pull requests to move through review quickly, proactively review and test the pull requests of other contributors. Helping us to keep the pull request queue moving quickly accelerates the work of everyone.
Use public communications channels whenever possible. JupyterLab is developed openly on GitHub. We have weekly video meetings on Wednesday at 9:00 am PT on Zoom (https://calpoly.zoom.us/my/jupyter). Finally, we have a public chat room on gitter (https://gitter.im/jupyterlab/jupyterlab). As much as possible, all work and communications should take place on these public channels. We welcome you to join these channels and introduce yourself. We are willing to have private conversations once in a while, but we try to minimize their frequency and always summarize them in our public channels.
Hire our core maintainers and give them more time to contribute. Consider paying the regulars around here to keep looking after the puppies. Our core maintainers love working on Jupyter, but have to balance those contributions with the reality of paid employment. If you are interested in hiring a core maintainer, we encourage you to do this in a way that gives them more time to contribute than they currently have. Put this into their employment contract. This is an incredibly effective way of turbocharging the work of the project and building trust in the community. Conversely, if you hire a maintainer who wants to work on the project and give them less time to contribute, you may damage the project and your organization may find itself alienated from the community.
Commit to maintaining your contributions. Don't abandon puppies here! Contributions, especially those for new features, add to the ongoing maintenance load of the project. New features bring new users to the project, and along with them, bug reports, ideas, discussions, questions, refactoring, etc. If your company wants to contribute new features, we ask that you commit to maintaining those features for a period of at least 2-3 years. This isn't a legal contract, but we do want you to carefully consider the maintenance of your contributions. If a new feature is abandoned and we are unable to maintain it, it may be deprecated or removed.
Don't surprise us. Don't build us a new dog park, when what we really need is just a couple more kennels. Before beginning major work, build trust through balanced contributions, and talk to us about your plans (see public communication channels above). See if there are existing issues on the topic, or open a new issue describing the work. There may be other people already working on it, or the work may be blocked by other factors.
Be patient. Puppies require training, and complex new tricks can take years and the right motivation to learn. Open-source software projects offer a unique path to innovation by bringing together users and contributors with diverse needs and usage cases. Even if fully resourced, the resulting innovation can be slow, possibly much slower than your corporation's desired timeline in the short term. However, in the long-term, sustainable open-source can lead to rapid growth and development. Help us build a sustainable project through long-term thinking, strategy, and resource allocation.