Scot Becker <scot.becker@gmail.com> wrote:

He wants to write up a document using org-mode's outline facilities as

a skeleton to help him build up, navigate and visualize his document,

but then he wants only to use SOME of the headlines but ALL of the

text when he actually makes a printed version for others to read.  

I've wanted this as well, since when you think about it the structure

you need as a writer may not be the structure you want to pass on to

your readers.

OK, thanks for the clarification: that was one of the interpretations
that did not make much sense to me, but I was looking at it from a much
more rigid point of view, where the headlines are what give structure to
the document, so you want to preserve them at all costs; I wasn't
thinking about different meanings that they might have for different
people.

However, it still sounds like an ill-defined problem to me: in
particular, your manual algorithm would wreak havoc on an outline with
headlines at multiple levels. What is the "real" algorithm supposed to
do with something like this:

,----
| * foo
| text1
| ** foo1
| text2
| * bar                  :omit-this-header:
| text3
| **bar2                 :omit-this-header:
| text4
| ***bar3
| text5
`----

Or are we supposed to imagine headings at a single level only?

I suspect that one would be better off with two (or more) outlines: one for
the writer, one for the reader (perhaps one for each class of readers), with
some way to pick text from one outline and plug it into the other(s).

Nick