I'm considering starting an open source project. Initially this would be a set of language specific modules (python and javascript) incorporated into ML/AI pipelines to improve observability.<p>- How much time should I spend structuring the project and building tooling to make development easier, vs just diving in and building the project?<p>- What are the most important areas to focus on make sure its as useful as possible (documentation, license, community, other areas I haven't considered?
If you are overthinking how to start, you will never start with it.<p>Just get coding and start prototyping.<p>More importantly, you are building an open source project for yourself. If you are building it for others, you won't be happy with it.<p>Always try to solve your own needs and problems.
>How much time should I spend structuring the project and building tooling to make development easier, vs just diving in and building the project?<p>Imo, at most 8hrs.<p>>What are the most important areas to focus on make sure its as useful as possible (documentation, license, community, other areas I haven't considered?<p>My order:<p>1. Demos
2. Good README with ability to get started in <4hrs
3. Documentation
4. Community
5. License (If it's supposed to make money, I vote BSL)
Microsoft has this which contains some nice pointers <a href="https://microsoft.github.io/code-with-engineering-playbook/" rel="nofollow">https://microsoft.github.io/code-with-engineering-playbook/</a>