Landed on this way of taking notes years ago and have refined it into a final system with minimal functionality. While the overall vision is not complete, the underlying "markup-like" syntax for Link Text is complete and wanted to share for reference. It's nothing fancy, just an indentation-based note-taking scheme for defining specifications or even writing code. It's parsing/handling is no more sophisticated than a JSON parser, as it literally is just parsed into a simple "term and text tree". From that point it is up to the person to resolve how to interpret the tree. On some level, it is just like a trimmed downed cleaned up version of XML, without attributes. You can, for example, write DSLs in Link Text format, and take the parse tree and serialize it into objects, which is what we are doing for some modeling of things like dictionaries and creative works and such. You can go further and define a programming language DSL using it, but then you have to figure out how to compile it into a programming language runtime. We are experimenting with that currently. The end goal with creating DSLs is, because it's a tree of terms, with text and values at the leaves, you can easily write a "tree-parsing grammar" to transform the DSL tree into objects for use as data in some programming language. So you take some notes in a DSL you just made up, refine it a little, then write a quick pattern-matching "grammar" to transform your notes into data objects. From that point you can pass the objects through transformation/processing layers and figure out interesting things to do with them.

Link Text

A data modeling language

A data modeling language