Consolidate documentation

[MattiSG]

Create a new codebase, generating documentation as a static website from a series of manually written files.

Technical choices

We prefer to create a new codebase because the layout will be different than the current main one, the current bus factor is 1, and the maintainer who has knowledge of the current system is off for a long time.
Over time, we are likely to migrate progressively the whole website to this new generator system. In the meantime, we will use a reverse proxy to serve documentation under the /doc path on the main website.

We intend to use Hugo because of its first-level support for i18n. Jekyll is easier to use, but its best support for i18n is still insufficient.

Contents

Manually written

• What is Open Terms Archive.
• Political statement (why and what for).
• Glossary.
• How to create a new instance.
• What are document types and how they evolve.
• How to contribute new documents.
  • Integrate https://github.com/ambanum/OpenTermsArchive/wiki/Usual-noise into it.
• How to navigate the history.
• How to subscribe over RSS.
• How to write a tweet.
• How to write a memo.
• Governance.
• How to access passwords.

Automatically generated

• Generate API documentation with JSDoc.

Consequences

• Update links in all versions repositories “about” section.
• Update templates and instances with links.
• Update README.

[MattiSG]

Repo created in https://github.com/OpenTermsArchive/docs.

After discussion with @Ndpnt and @clementbiron, the ambition to migrate this codebase to the docs one is cancelled, as the governance of both these sets of documents is very different. Another consequence is that the docs subdomain can be used, since there is no more ambition to unify the UI and paths, easing deployment.

A first pass of document writing is done in OpenTermsArchive/engine#971 and will mean mostly moving the contents of the docs folder to https://github.com/OpenTermsArchive/docs/.

[MattiSG]

The documentation has been deployed! 🎉
I leave this issue open for now as the contents should be checked against the list in the original description, and the contents erased from their original location as they are checked.