On Mon, 11 Jul 2016 23:10:57 -0400
Hendrik Boom <hendrik@???> wrote:

> On Mon, Jul 11, 2016 at 10:19:03PM -0400, Steve Litt wrote:

> I'm kind of curious about the format you seem to be designing or
> developing.

Let me start with my priorities:

1) Fast, easy authoring, from a text editor. I'm not saying this rules
out things with end-tags, but it sure puts them at a big
disadvantage. I don't feel like maintaining a GUI just to deal with
my format. LyX did that, and 2 decades later, as I see it they're
painted into a corner.

2) SEMANTIC writing. As an author, I understand the need for styles.
The styles I apply should be kept as styles through all layers of
export, and converted to appearance only at the last possible
moment. This is what will allow write once, read multiple formats.

3) Human readable/parsable native format. This enables writing in the
text editor, proofreading in the text editor, and all sorts of
exports by anybody who can write a program.

Here's my first attempt at a specification. I never carried it out
because I'd bitten off more than I could chew:

http://www.troubleshooters.com/projects/stylz/

One of the first thing I'm getting rid of is the directory/file
configuration setup. It's so djb-correct, and so much work for so
little gain. I'll be using YAML for all of that stuff.

What I'll be doing is a small subset of what I identified. Paragraph
styles and character styles for sure. Other stuff might just be in
Xhtml. I have absolutely no trouble doing the trickier stuff in
Xhtml. I'd rather get this thing working than win the Nobel Prize.

Fancy stuff like bibliographies, footnotes, etc, will have to wait for
version 2. I have no idea how to make them, and don't want to risk
their inclusion in any way slowing the authoring process.

> I've build a completely ad-hoc notation that doesn't
> satisfy me at all any more. I'm looking to convert it to something
> else. The good thing about it is that I *can* convert it, sonce
> I've implemented it myself.
>
> I'm tinkering with writing the project notes (as opposed to the main
> text) in markdown, using omd, a rather fast markdown converter
> written in OCaml) to convert it to html, ad then looking at the
> organisational notes in a browser (currently Midori).

I flirted with Markdown, MultiMarkdown and the like. Trouble with those
is that arbitrary styles are difficult or impossible. I want to avoid,
at all cost, premature conversion of semantic styles, like the STORY
paragraph style, to its appearance, like indented italics. You know how
when you solve an algebra problem you leave the variables in place
until the bitter end, rather than substituting known numeric values for
known variables beforehand? Same idea.

>
> But the main text remains in the ugly ad-hoc notation.

Ugly isn't a problem. Slow would be a problem. Unreadable would be a
problem. Hard to parse would be a problem. Ugly --- I'd rather write
fast than work with pretty tools.

>
> I may have written a diatribe on the design of proper notations for
> source text a while ago, long after I had developed the ad-hoc one.
> ONe thing I have discovered is that it should be based on separators,
> not wrappers. The Open Document types (used by OpenOffice) is based
> on wrappers -- XML does bracckets all the time. I'll see if I can
> still find that diatribe.

I'd love to see both your diatribe and some content in your ad hoc
native format.

SteveT

Steve Litt 
July 2016 featured book: Troubleshooting Techniques
     of the Successful Technologist
http://www.troubleshooters.com/techniques