Attention: Here be dragons (unstable version)

This is the latest (unstable) version of this documentation, which may document features not available in or compatible with released stable versions of Redot.

Content guidelines

This document outlines what should be included in the official documentation. Below, you will find a couple of principles and recommendations for writing accessible content.

We want to achieve two goals:

  1. Empathize with our users. We should write in a way that makes it easy for them to learn from the docs.

  2. Write a complete reference manual. Our goal here is not to teach programming fundamentals. Instead, our goal is to provide a reference for how Redot's features work.

Guidelines and principles

Below are the guidelines we should strive to follow. They are not hard rules, though: sometimes, a topic will require breaking one or more of them. Still, we should strive to achieve the two goals listed above.

Writing complete and accessible documentation

A feature doesn't exist unless it is documented. If a user can't find information about a feature and how it works, it doesn't exist to them. We should ensure that we cover everything Redot does.

Note

When adding or updating an engine feature, the documentation team needs to know about it. Contributors should open an issue on the redot-docs repository when their work gets merged and requires documentation.

Do your best to keep documents under 1000 words in length. If a page goes past that threshold, consider splitting it into two parts. Limiting page size forces us to write concisely and to break up large documents so that each page focuses on a particular problem.

Each page or section of a page should clearly state what problem it tackles and what it will teach the user. Users need to know if they're reading the correct guide for solving the problems they're encountering. For example, instead of writing the heading "Signals", consider writing "Reacting to changes with signals". The second title makes it clear what the purpose of signals is.

Note

Long section titles lead to long entries in the side menu, which can make navigation cumbersome. Try to keep headings five words long or less.

If the page assumes specific knowledge of other Redot features, mention it and link to the corresponding documentation. For instance, a page about physics may use signals, in which case you could note that the signals tutorial is a prerequisite. You may also link to other websites for prerequisites beyond the documentation's scope. For example, you could link to an introduction to programming in the getting started guide, or a website that teaches math theory in the math section.

Limiting cognitive load

Limit the cognitive load required to read the documentation. The simpler and more explicit language we use, the more efficient it becomes for people to learn. You can do so by:

  1. Introducing only one new concept at a time whenever possible.

  2. Using simple English, as we recommend in our writing guidelines.

  3. Including one or more concrete usage examples. Prefer a real-world example to one that uses names like foo, bar, or baz.

While many people may understand more complex language and abstract examples, you will lose others. Understandable writing and practical examples benefit everyone.

Always make an effort to put yourself in the user's shoes. When we understand something thoroughly, it becomes obvious to us. We may fail to think about details relevant to a newcomer, but good documentation meets users where they are. We should explain each feature's capabilities or intended uses with the most straightforward language possible.

Try to remember what you first needed to know when learning about the feature or concept. What new terms did you need to learn? What confused you? What was the hardest to grasp? You will want users to review your work, and we recommend you practice explaining the feature before writing about it.

Note

Programming fundamentals are a prerequisite for using a complex engine like Redot. Talking about variables, functions, or classes is acceptable. But we should favor plain language over specific terminology like "metaprogramming". If you need to use precise terms, be sure to define them.