What is `documentation`?

documentation is a documentation generation system that targets JavaScript code; and is itself implemented in JavaScript. It exposes multiple interfaces for users:

with npm i -g documentation, it provides a binary for command-line usage
install documentation with npm to use the node-facing interface

documentation runs in node.js but supports JavaScript that runs in any environment. You can use it to document browser libraries, server libraries, and wherever RequireJS or another module system is used.

How do I use `documentation`?

There are two main ways:

You use the documentation command on your command-line to generate docs from your source code.
You use one of the integrations with a build system like Gulp to generate docs from source code.

How does `documentation` differ from JSDoc?

JSDoc is both a standard syntax for documenting code and an application, also called jsdoc, that processes that syntax into documentation.

documentation uses the JSDoc syntax and provides an alternative to the jsdoc application.

Why use `documentation` instead of JSDoc?

documentation aims to modernize and simplify the process of generating JavaScript documentation.

Beautiful defaults for HTML & Markdown output
Supports CommonJS require() syntax so that node modules can be documented by giving their main file
Extensively documented internally: all public and private functions in documentation are documented. JSDoc is not well documented internally.
Robust ES6 support
No Rhino cruft
Uses JSON literal objects for data representation instead of the abandoned and untagged TaffyDB project.
Uses high-quality node modules for syntax parsing, argument parsing, and other tasks: separates concerns so that we can focus on a robust solution

Why use `documentation` instead of writing a Markdown file by hand?

documentation can generate multiple formats. When you create a website, documentation can take your documentation and generate beautiful HTML output.
The JSDoc syntax exposes a powerful, standardized type syntax to, for example, express parameter types like 'an array of strings'. as Array<String>, and to support custom object types with inter-linking
The eslint valid-jsdoc rule makes it possible to require documentation as part of your linting step, ensuring that new code doesn't lower documentation coverage.

Which files does documentation.js include?

By default, documentation.js follows dependencies within your source tree and excludes node_modules from results. This is meant to process your application code automatically but avoid documenting the npm modules you're using.

This means that if you point documentation.js at your index.js file and that file uses require or import to include other source files, those source files will be documented too.

You can customize this behavior by specifying the --shallow command-line option. With --shallow specified, dependencies aren't followed: documentation.js processes only those files you explicitly name.

If you're using ES modules, you enable the option --document-exported to automatically document all exported bindings in your project, even if they don't have JSDoc comments. This also ignores non-exported items, even if they are commented.

Will adding JSDoc comments slow down my code?

The short answer is "no".

As far as execution performance - how fast your code runs - all JavaScript implementations like V8 or SpiderMonkey will remove comments from the generated low-level code that they run. In other words, your browser does not run JavaScript as a string of code - it parses your code into an intermediate representation that ignores comments, and in this system comments, as well as whitespace, have no effect on performance.
As far as download performance - whether these comments add kilobytes to website's download time - any typical code minifier like UglifyJS or Closure Compiler removes comments by default when compressing your code.