I agree – FWIW my personal preference is for preserving the distinction between structure editing and serialization.
There’s a clear case for multiple lines of text within a single content block in an export format, but I think we begin to lose the power and clarity of the outline model if we zip a sequence of peers / siblings together into a monolithic text block too soon.
The trick, I think, is to work on structure in a small node outliner, and then serialise out to custom document formats when we need them.
(It’s already possible, incidentally, to write such serialisations (export formats) from Bike html outlines, using standard xml parsing tools, with
pandoc, for example downstream)
Very happy to sketch any examples that might interest people, perhaps this weekend, if there are any suggestions.
(An example might be to export the top N
bike levels as headers, the next level as sentence groups run together into paragraphs, levels below that as bullets etc, and to write out to Markdown, or pipe through to