On Sat, Jan 4, 2020 at 17:56 Steven Penny <svnpenn@gmail.com> wrote:

> On Sat, Jan 4, 2020 at 4:35 PM Trey Ethan Harris wrote:
> > At some point in 2016 I see from my Git history that I un-filled all my
> Org
> > files (i.e., changed “paragraphs” interspersed with newlines near
> fill-column
> > to a single line per paragraph no matter how long it is), so for some ten
> > years before that I was wrapping lines and just got tired of the hassle.
>
> Nope. I have been dealing with long links for many years, since at least
> 2011.
> Last month I learned that both reStructuredText and AsciiDoc can deal with
> them
> pretty easily. Now that I know this, I am more included to use Markups
> that have
> that capability.
>
Not sure what you mean by “Nope”—I don’t publish the Org files I’m
talking about, but I could show you some sample diffs if you don’t believe
me for some reason...

But my larger point was that for me, _stored_ line length (as opposed to
displayed line length) should be dictated by export, not for some arbitrary
purpose. By doing it this way, source blocks get wrapped according to their
modes’ ordinary behavor (which is what I want if I ever publish the code),
and prose text is stored in “paragraph, 2 × newlines, paragraph, 2 ×
newlines, paragraph” form—which ,for pretty much everything except email
these days, is what you want.

A common purpose I put Org prose to is editing text I will post as web
content, like in a forum; I’ve found a distressingly high proportion of
comments parsers can’t deal with newline-wrapped text and turns each
72-column line into a paragraph. But almost all of these treat blank lines
delimiting paragraphs correctly. So storing my text this way requires the
least post processing—blocks get post processed by content-aware dedicated
exporters, while interstitial text may or may not—but it’s at least always
in a  form that 95% of the time can just be cut-and-paste, if nothing else.