Rubenerd

derived is an add-on to RSS 2.0 XML namespace that provides context about the source code for textual media.

At the moment there’s one attribute, used like this in RSS and RDF:

<derived:code type="text/markdown" 
    url="[github repo]/master/content/post/2016/derived.markdown" />

Or inline in HTML:

<meta property="derived:code"
    content="[github repo]/master/content/post/2016/derived.markdown" />

Multiple attributes with different mime-types could be used, though having a single canoncial source is strongly encouraged.

Why?

Like many sites, Rubenerd is written in Markdown and pre-rendered for clients in HTML. This link means smart clients could take the original markdown source, and render it for their specific viewport or client.

Facebook have shown with their stories feature that there’s demand for refactoring/optimising content. I want the same ability for the open web using standard RSS and XML.

How is this different from Dave Winer’s source namespace?

Good question. Dave’s source namespace includes the outline attribute, but that carries an implication that the source is an OPML or similar outline. I wanted an attribute that linked directly to the original source, regardless of structure.