Nhost - Backend-as-a-Service with GraphQL for modern app development - Interview with Johan Eliasson
When developing a project, you should consider its goals and timespan. Some projects are one-off and meant to solve an immediate problem and are then thrown away. Others have more longer term focus and evolve over a long time. The question is, how to enable this long term evolution?
You have to consider at least the following aspects:
In single person projects you develop the project, publish the package, and forget about it. One day someone might find the project and notice that something is missing. As discussed in the Processes chapter, this could lead to a pull request and to a new contribution to the project. You gain contributors this way.
A model like this could be enough for one-off projects. You solve a problem and revisit it only when it’s relevant. If the problem is big enough, then it might require a sustained effort of longer time to resolve. It could be possible that there is no clear way and you will have to tune the approach as time goes by. Projects like this benefit from a community especially if the problem solved is important to a group of people.
In community projects initial users may evolve into contributors that give something back to the project. If they are active and fit the project culture, they can become project maintainers that shepherd the development of the project and help contributors.
Sometimes new maintainers may pick up over the project completely as older maintainers fade away and find something else to do. Interests tend to vary and people move across projects. Spending some time beyond a project help in avoiding burnout or at least help you gain perspective on the project. When you look at a project too close for too long time, it can be difficult to see how to improve since it’s so clear to you.
Although open source software is freely available, this doesn’t mean it’s free to develop and maintain. At least it takes the time and effort from a developer to write and publish it. Especially the maintenance effort tends to require a lot of energy especially for popular projects that cannot keep up with the demand. Some do this for fun or as a hobby, but there are also commercial models that enable developers to dedicate time towards development while getting paid.
The models tend rely either on benevolence (donations), added value (consulting, merchandise), licensing models (dual license), and grants. Lemonade stand by Nadia Eghbal, open funding list by ralphtheninja, and Producing Open Source Software by Karl Fogel cover common options in greater detail.
A project may have explicit goals or it may evolve organically and find its purpose. Regardless of how it evolves, the maintainers of the project have to shepherd it into a good direction.
In some projects, all design decisions go through a certain person. This model, Benevolent dictator for life (BDFL), has proven successful in many cases and it relies on the leadership capabilities of the leader. If there are disputes within the community, the BDFL will have the final say on them.
A project can have a group of core developers that steer the direction of the project together. The model combines with BDFL and doesn’t necessary exclude the existence of such in a project.
In the worst case, a project becomes abandoned by its maintainers and new contributions cannot make it to the project. A fork may be needed to rejuvenate the development. The problem is discussed in detail in the Where to Start Packaging chapter.
To keep projects sustainable over longer term, especially the bigger ones focus on attracting new contributors and maintainers. The interests of the different parties need to align for this to work.
At its core you have the project maintainers. Their work is normally split between reviewing new contributions and contributing themselves. The availability of maintainer time tends to be the biggest constraint in large projects as they have to take care to grow contributors into future maintainers.
To even gain contributions, the project should be contributor friendly. Various aspects discussed in this book, such as documentation, are important here because if project is too difficult to contribute to, you won’t get contributions. Certain development practices, such as testing and continuous integration, can make it easier to contribute while pushing the quality of contributions higher and avoiding maintainer work.
As a project grows in complexity, it can become difficult to keep track of it. Popular services, such as GitHub, work as starting points. As the complexity grows and there’s more to worry about, solutions such as Octobox become valuable as they focus on these problems at scale.
Sometimes a project becomes popular even if popularity isn’t a goal. This can happen if the project solves an important problem and the right people find it. They will evangelize the project and make it popular whether you wanted or not.
Popularity comes with its challenges. Even though it might sound like a great thing, it comes with responsibility. Popularity tends to lead to increased demand for maintainer attention. The project development process needs to be solid as discussed in the Processes chapter and has to deal with the pressure to remain popular.
Sometimes the pressure of popularity is too much and people give up on project maintenance. This is true if the incentives aren’t there. Once it becomes too much work for too little reward, it can feel waste of time. Thinking about project longevity and development model is important because of this reason.
Project longevity is one of the key concerns you should think about when developing a project. This can happen if you think the project has a community-driven future for it. Becoming a community project isn’t the only way, though, and developing something in small scale can be worthwhile as well.
You’ll learn how to market your project in the next chapter.
This book is available through Leanpub. By purchasing the book you support the development of further content.