- King of Qin Hero (2002)
Marle began as an effort to design a better authoring language for the web. Early iterations explored custom markup syntaxes, parsers, and publishing formats that could extend beyond the limitations of existing solutions. While each approach introduced new capabilities, they also introduced additional syntax, complexity, and long-term maintenance concerns.
The project ultimately arrived at a different conclusion. Rather than inventing another language, authored intent could be represented directly as structured data. Articles become collections of explicit content blocks, each describing what something is rather than how it should be written. Individual websites then decide how those blocks are interpreted and rendered, allowing the same underlying content to evolve alongside changing designs and technologies without requiring migration or re-authoring.
Today Marle exists less as a product than as a publishing methodology. It provides a simple, practical pattern for storing canonical content as data, together with reference implementations that translate those structures into semantic HTML. Most of my websites are built around variations of this approach, extending it with site-specific components while preserving a shared authoring model that remains lightweight, easy to implement, and resilient over time.
Because Marle stores content as structured data rather than markup, it can be rendered by any system. In my own projects I typically pair it with Lydio to output HTML, allowing articles to integrate naturally with reusable components, semantic layouts, and site-wide presentation patterns without coupling the content to a particular rendering technology.