💾 Archived View for idiomdrottning.org › path-through-prose captured on 2022-06-11 at 21:25:09. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2022-06-03)
➡️ Next capture (2022-07-16)
-=-=-=-=-=-=-
One real key to writing an unruly, hard to organize text is transition sentences. There’s no perfect path through a complex topic, and sometimes no very good path either. Put together an acceptable one and write great transitions.
That’s not my jam. My method is to start with two outlines.
On the right, the outline of the genre (for a science paper, “abstract”, “introduction“, “premises”, “conclusion“ etc. For fiction, the events chronologically and escalatingly, with any mysteries pulled out and saved as later reveals. For technical or political texts, and I need to get better at this, start with what you want and then follow up with the how and why).
On the left, my own mindmap of everything you found out—mindmaps can be converted to outlines. Then move everything over one by one and then write it clean.
While this is still good for longer texts, I don’t use this very often anymore for blog-length texts because I often just do it mentally instead. I imagine the underlying tree structure skeleton without having it explicitly written out and then move things around in my head to guide my prose editing.
I use the path through the text as a test for “does this work”. The reader is going to experience the “tree” of ideas as a depth-first tree-walk and if nodes don’t fit together, I need to reorder, move things around, or even remove or add nodes.
If I can’t find a good path without segues, I shelve the text.
This happens around one or two out of five texts. Often they get rescued later. A path will reveal itself, the missing nodes will have shown up, the superfluous nodes will have outstayed their welcome. Other times the idea is dumb and unworkable and I leave it in the shelf. Even then, the story of the idea (“I used to think that…”) is a good addition to a later text.
I edit a lot, even small posts and replies on Fedi and IRC, where rcirc’s multiline buffer (C-c C-c) is a godsend. That’s why I’m so much clumsier when talking in person. (That and the accent.)
There are a lot of patterns to my editing. I’m like “OK, I can remove the parens around this parenthetical, it belongs in the main text” and “I need to move this up” and “OK, this became redundant” or “OK, this entire topic needs to be its own post, and it needs to be published first so I can refer back to it”.
The reason my art is progressing so much slower, and is so much more terrifying, is because I can’t find good feedback. With writing, I could, as long as my writing was in programming languages (or Lojban, which also has a tree parser). The compiler is a harsh mistress. It made me comfortable with writing and refactoring (without fear, since every step of the way I could compile and see if it still worked and tests still ran), and I brought those editing patterns back to English.
When working with trees this way, there’s an important limitation to keep in mind compared to graph trees.
If you start a subsection, you can’t go back to the section you started it from. You can start a new subsubsection, a sibling subsection, or a new section at any higher level, but you can’t resume the section you were in. This used to trip me up all the time when I was writing. I’ve learned that I need to hold off on starting subsections until I’m done with what I wanted to say in the section itself.
(The hacky workaround, which I sometimes resort to, is sidebars or footnotes.)