Introduction

Two closely-related conversions produce the lightest-weight output PreTeXt offers: plain text, and markdown. The markdown version is the plain text version, plus the small set of conventions that make the structure of a document recognizable to a renderer: headings, emphasis, links, code blocks. Either way you get files a human can read with no tools at all beyond a pager or an editor, and that a program can consume without any parsing to speak of. A whole document can be a single file, or the chunking election of the publication file will produce one file per division, each with a trailing navigation line, along with summary pages (Subsection 45.1.1).

Some uses we have in mind. Websites that host source-code repositories will render markdown automatically, so a project can provide a browsable version of a book right alongside its source, with working cross-references, images, and mathematics, and no web server involved. Markdown is also the native language of AI assistants and their tools: a compact, faithful rendering of your book—or of one section of it, chunked—is exactly the right thing to place in front of such a tool. And plain text is the universal solvent: it can be searched with the simplest tools, quoted in email or a discussion forum, compared line-by-line between versions, or fed to the next program in a pipeline.