TL;DR Computer Words
I write a lot of documentation. I do it for two reasons: I like writing, and I
like to be understood. Documentation is what takes software from “something
@irskep put on GitHub” to something you can actually use.
But I’m not happy with the available tools. Open source documentation for young projects tends to fall into one of three modes if they have it at all:
- A GitHub wiki
- A directory of Markdown files
- A proper web site, but requires a tool no one wants to install or a syntax no one knows, except one maintainer
Some ecosystems have good, fairly well known tools (see Python and Sphinx), but with the prevalence of Markdown and the convenience of hosting Markdown files on GitHub, I suspect people have been tending toward the solutions with less friction.
I want it to be easier to produce good documentation sites, without GitHub branding, with a bare-minimum learning curve, and a sky’s-the-limit plugin architecture, so that more projects have better documentation.
I think I’ve succeeded. Computer Words is a tool written in Python 3 that lets you turn your directory of Markdown files into a beautiful web site, without sacrificing important features like semantic cross-referencing. It vaguely resembles Sphinx in its design, so it has the potential for powerful plugins, but I did my best to keep it small and easy to use.
If your docs are currently in Markdown, these features might interest you:
- All extensions use HTML tag syntax, so your editor’s syntax highlighting won’t break. Other tools may or may not like it, depending on how they parse HTML within Markdown.
- Semantic linking. Rather than baking an HTML reference into your links,
<heading-alias name="blah" />above any heading and insert a link to that heading with the correct name as
<heading-link name="blah" />.
- Automatic multi-document table of contents auto-generated from your headings.
- A reasonable configuration can be as short as 4 obvious lines copied from the Computer Words home page.
If you want to auto-generate docs from a programming language besides Python, and you are willing to write a docstring parser for your language (in your language!), I would love to help you. The interchange format is very simple.
Version 1.0 is a groundwork-focused release. The theme isn’t very fancy, but most of the pieces are in place to grow the system cleanly.