The evolution of markup languages traces a fascinating arc from the complex to the simple, reflecting a recurring pattern in computing where powerful but cumbersome systems eventually give way to more accessible alternatives that sacrifice some capability for usability.

The story begins with SGML (Standard Generalized Markup Language), standardized as ISO 8879 in 1986. SGML was designed as a meta-language for defining document markup languages — it was not a markup language itself but a system for creating them. SGML's power came with extraordinary complexity: the full specification supports features like tag minimization (omitting closing tags), SHORTREF (defining custom delimiter characters), and CONCUR (allowing multiple concurrent document structures). Very few systems ever implemented the full SGML specification.

HTML emerged in 1991 as Tim Berners-Lee's simplified SGML application for hypertext documents on the World Wide Web. Early HTML was deliberately simple — a handful of elements for headings, paragraphs, lists, and crucially, hyperlinks. As the web grew, HTML accumulated features through browser vendor competition (the "browser wars"), leading to inconsistent implementations and increasingly complex specifications. HTML 4.01 (1999) attempted to formalize the language, and XHTML 1.0 reformulated HTML as an XML application, imposing XML's strict syntax rules on web documents.

XML (1998) represented an attempt to create a simplified, strict version of SGML for data interchange. Unlike HTML's fixed vocabulary of elements, XML allows users to define their own elements, making it suitable for arbitrary structured data. XML spawned an ecosystem of related technologies — XPath for querying, XSLT for transformation, XSD for schema validation — that provided powerful document processing capabilities but at significant complexity cost.

Markdown (2004) emerged as a reaction against this complexity. John Gruber recognized that for the most common writing tasks — paragraphs, headings, lists, links, emphasis, and code — the full power of HTML was unnecessary. By using conventions already common in plain-text email (asterisks for emphasis, blank lines for paragraphs, leading hashes for headings), Markdown created a format that was simultaneously human-readable in source form and convertible to HTML for rendering. This "lightweight markup" approach proved enormously successful.

Markdown was not the first lightweight markup language. Wiki markup (used since Ward Cunningham's WikiWikiWeb in 1995), reStructuredText (2001, used in Python documentation), Textile (2002, used in Ruby on Rails), and AsciiDoc (2002) all predated Markdown. However, Markdown's simpler syntax and association with blogging (Gruber's Daring Fireball) and later GitHub gave it dominant market share. The emergence of these lightweight formats reflects a fundamental insight: for most writing tasks, the overhead of XML angle brackets or HTML tag pairs creates friction that discourages authors from using structural markup at all, resulting in less structured documents than a simpler syntax would produce.

Converting between these formats is non-trivial because they differ in expressiveness. HTML can represent structures (like colspan in tables, or arbitrary nesting of inline elements) that Markdown cannot. Converting HTML to Markdown requires making lossy decisions about how to represent these structures, and different converters make different choices. The reverse conversion (Markdown to HTML) is lossless since Markdown is a strict subset of HTML's capabilities, but the specific HTML output can vary between parsers for edge cases not explicitly covered by the specification.