Wednesday, February 11, 2009

Back in the snowflake again

I'm reviewing the ideas presented in the Snowflake Method for organizing a novel, by Randy Ingermanson. I thought Ingermanson had a lot of good practical advice in that article. The key things I thought he got right were:
  • an iterative process with layers that lets you manage the large and small structure of a novel; iterative meaning you feed back and correct each layer as you develop the next
  • he emphasizes that you should use what works from his ideas. He doesn't think he has all the right answers.
Ingermanson's web site is a bit marketingy and salesy and maybe oversells him. I'd never heard of his writing before encountering the Snowflake concept through NaNoWriMo. He calls himself an award-winning writer, but the awards he cites are for Christian writing, which is a niche that doesn't give him any particular juice for me.

Still, I've gotten a lot of use out of these ideas already. But I've been writing short stories, and I've never done all the detail involved witha full Snowflake. Now I'm working on a novel and I plan to go all the way, baby. And all the way, with the Snowflake method, is a long way: you end up with pages and pages of notes.

I generate pages and pages of notes anyway, though. They just aren't usually as organized as the Snowflake method would have them.

Here's the Snowflake in the smallest nutshell I can distill it down to:

Write progressively more detailed descriptions of your story, like this:
  • One sentence.
  • One paragraph summary.
  • Goals and one-paragraph storylines for each major character.
  • Expand the one paragraph summary to a page.
  • One page per major character.
  • Expand the page summary to four pages.
  • Detailed character charts per character.
  • List of every scene needed to make this plot work.
  • Description for every scene, totalling about a page per chapter.
  • Start your first draft...but halfway through, update your design docs.
The updating thing is more important than anything else. Later steps will provide ideas that will change earlier steps. You should use those, and you should go back and make everything consistent; that's how you can make a complex novel work. And man, is it ever work.

Image made using the Make-a-Flake flash tool. I wanted a snowflake image and it occurred to me that there must be some art tool out there for making snowflake pictures, and it was only a Google away. The zeitgeist is a wonderful thing.


  1. Maybe you could use some of your XML ideas for tagging essential plot scenes, essential character scenes, etc.

  2. Yeah, I could do that. I'm kinda afraid it would devolve into tagging-instead-of-writing.

    I periodically revive an idea to do a very simple xml for prose. But you know, you could tag character names, I'd love to demark things like new day starts .... it quickly becomes madness. ;)

  3. My challenge with the Snowflake method is that I've never felt like I could complete all the preliminary outlining and notemaking that comes before the actual writing, because I've never had that complete an idea of a story in my head.

    Then again, maybe that's a warning sign.

  4. So, I keep raving about this method, but I never go all the way either.

    (Probably totally consistent with my whole life, to be reluctant to go all the way.... :) )

    But my excuse before has been that I was working on short stories, which didn't seem to need the whole shebang. I still think it would be a good idea to try to take a short story through iterative phases where you worked your way through more and more detailed outlines until you suddenly had...the story!

    When I work on the snowflake, and I've been working on this one for the last 5 days solid, I tend to work on one chunk, then if I get tired of it, move ahead. I haven't done all my character storylines, but I drafted a 1-page summary anyway. And I think that's okay: I'm not sure the current design needs all the same characters from the story design, anyway, and why not explore up and down the chain? I think that, reading a procedure like this, it's hard to get the proper emphasis on the real iterative up-and-down the scale process.