Processing:
  * Optionally cap over-long titles in headers.
  * Mark bad links (not only warn at build) in flightcheck+draft mode.
  * Warn about possible HTML escapes in content (e.g. &#\d+; or %{HEX}).
  * Maybe comply with PDF/A-1b (or newer).
    + Include to preflight online test against e.g.
      <https://www.pdf-tools.com/pdf/validate-pdfa-online.aspx>
  * Generalize pandoc-memoir filter.
    + Enable matter injection conditional to environment variable.
    + Define Headline regexes via environment variable.
  * Verify that pandoc-citeproc is applied only once.

Content modelling:
  * Use -- for en-dash and --- for em-dash, and update EDITING.
  * Consider use typographic enhancement tunings:
    + icomma - avoid automagic space after comma (e.g. money amount).
    + ellipsis - ensure balanced space around ellipsis.
    + xspace - ensure space after superscript/subscript characters.
    + numprint - space/comma separated enumeration.
    + cite - enhanced citation handling.
  * Support numbered lists with complex starting point.
    + "Mandatory Characteristics" list ideally should be 4.1.1-4.1.6.
  * Redirect bibtex parsing warnings/errors to STDERR.
    + Example: Title section containing doublequote.
    + ideally switch to Biber.

Styling:
  * Use standardized layout for bibliography.
    + Obey csquotes in citations.
  * no quoting around origin link string.
  * Maybe visually wrap URLs in <> to distinguish trailing punctuation.
  * Adapt colophon for each flavor (and re-add varying type measure).
  * Penalize pagebreak between single line and list.

Repository:
  * Add README
  * Add proper copyright and licensing
    + Code: AGPL-3+.
    + Design: CC-by-sa-4.0.