as a living document (akin to a playbook)

[urbanjost]

Depending on the markdown varient it is to read till a begin string like matrix is encountered and then just go back to read and copy when a stop string like is found. combine that with ignoring lines in that block beginning with # and prefixing all output back with a # like a diary file so it is replayable and you can create simple files with github metadata in them.

Applied to HTML the <xmp> and </xmp> strings could be the delimiters, albeit HTML5 deprecates <xmp>.

One of the nicer things about NAMELIST is input is in named blocks so all lines outside of the blocks are simply ignored and can be anything from *roff to markdown and so on.

Should this be built-in?