plain text & feature creep

[Albert Skye]

Basic Markdown (and other languages with even less syntax) may be semantically sufficient for many applications but as semantic requirements increase, plain text inevitably requires augmentation with additional metadata/syntax.

I imagine most prefer (as do I) that semantics be inferred, in preference to introducing syntactic metadata, and that careful attention be given to the form of whatever syntax is necessary.

For my own purposes, the following are essential:

- primitive container elements (inline and block)
- the ability to attach attributes to any element
- general substitution (to include more than only images)

Though a language may be used for displaying styled text and much more, plain text remains useful as a source format, and if one wishes to forgo extended semantics, superfluous metadata need not be present; i.e., a language which provides rich semantics may have simple syntax when used simply, as syntactic load may increase/decrease in natural proportion to semantic complexity.

For those who find Markdown insufficient, it may be useful to modify it (and even throw some things away) instead of creating an extended superset. Perhaps Markdown ought to remain as it is, and the greater problem of structured text, abstracted and addressed directly, informed by Markdown instead of bound by it.

askye