There's plenty to do. Beyond the obvious, here are some things to tidy up:
- Handle xref with unexpected content
- Handle xref to element with no title
- Handle whitespace only and empty link and ulink elements
- Handle duplicate ids in source docbook document.
- Markdown and HTML documents should be treated as articles, not as books.
- Generated HTML links should have a title attribute.
- Tidy-up output file naming scheme.
- Add some static analysis to the build.
- Log a warning when an Error element is added to the document.