Convert Markup

Many Kinds of Documents

Markdown
⟷ DocBook ⟷ DokuWiki markup ⟷ EPUB ⟷ Emacs Muse ⟷ Emacs Org-Mode ⟷ FictionBook2 ⟷ HTML ⟷ Haddock markup ⟷ JATS ⟷ Jupyter notebook ⟷ LaTeX ⟷ LibreOffice ODT ⟷ MediaWiki markup ⟷ Microsoft Word docx ⟷ Textile ⟷ reStructuredText ⟷ roff man

Document AST

data Inline
    = Str Text
    | Space | SoftBreak | LineBreak
    | Emph [Inline]
    | Code Attr Text
    | Math MathType Text
    …

Why Haskell?

Barrier of Entry

I don’t know enough Haskell to create a pull request.

Comparison Projects

Drive-By Contributors

Same-Same

Contributors

Language Popularity

Level of Involvement

Non-code Contributions

Accommodate

Non-Haskell extensions

function Meta(m)
  m.date = os.date("%B %e, %Y")
  return m
end

[
  {"t": "Str", "c": "Hello,"},
  {"t": "Space"},
  {"t": "Str", "c": "World!"}
]

Libraries?

• Then: sparse
• Now: acceptable

Non-Standard

Libraries!

Purity ⇒ Sanity

Testing

• Unit tests
• Property checks
• Easy mocking

Refactoring

Case-Study: String to Text

data Inline
    = Str String
    | Space | SoftBreak | LineBreak
    | Emph [Inline]
    | Code Attr String
    | Math MathType String
    …

Live-Demo

Thanks

John MacFarlane
and
everyone involved with pandoc

Support pandoc

🤎 Sponsor

Summary

• Leverage Haskell for its strengths.
• Mitigate its weaknesses.
• Plenty of talent around.
• Use pandoc for all your document conversion needs.