Via the always-affable Austin Kleon, here’s a post from one Mason Currey reporting on a revision technique used by Oliver Burkeman¹, whereby in order to get a fresh outlook on a piece of work, Burkeman will…

… get a draft going on the computer, print it out, delete the version on-screen in front of him (!), and then retype it from the printout.

As he’s retyping, Burkeman inevitably makes changes—and these changes, he says, “come very easily and naturally,” much more so than if he were to just fiddle with the draft on-screen.

I’ve never tried that one myself, but I quite often print out drafts to annotate and edit them with pencil on paper, and type up insertions in a separate document (or, these days, Obsidian) before splicing them into a fresh version’d copy of the original; indeed, I was doing this just yesterday with a short story. Editing in-doc and on-screen is hell; the only thing worse is editing in-doc and on-screen and having to keep track changes running².

Further down Curry mentions the nuclear version of this practice: Lauren Groff writes her first drafts by hand in a spiral notebook, then puts the notebook in a banker’s box and rewrites the whole damned thing entirely from memory. The late Christopher Priest used to do this; I believe that Nina Allan and Matt Hill both still do, and I’m sure there’s plenty of others.

I can see the logic, and even feel the call, but I’ve always been both too scared and too short of time to try; it’s not an approach to try on anything with a tight deadline, I think. If you’ve got a year to write a novel, sure, why not! But my writing life usually ends up feeling a bit like the Iron Council, frantically laying track just ahead of its own front buffer—and that’s not conducive to this start-from-scratch approach.

1. This sort of “X says that Y says that Z…” thing kinda drives me nuts, but I’m enough a product of blogging’s Gilded Age that I can’t bear not to cite my vias, and once you’ve started it’s hard to know where to stop. ↩︎
2. I usually say to my proofreading clients that they should take the track-changes version, make a copy, accept all changes and then read that, the edited no-tracking version: if you can’t tell where a change has been made or might still be needed without having a bunch of visual cues there to tell you, then no further changes are required. Track changes has nothing to do with making written work better, and mostly serves to make editing slower and more frustrating for everyone involved. ↩︎