I have a tendency to get stuck on the start.
And for this, iteration is my friend. I'll edit pretty much ad infinitum, so it's never a stretch to tell myself that what I need to write (or do) is a rough draft.
Recently, there was something at work that I wasn't terribly excited about doing. But I wrote up a draft (in Word, people!) and sent it out. I've been editing for about a week, and the thing is looking just fine now. Thanks, iteration!
I find that the promise of improvement doesn't always get me unstuck, and I'm not really sure where the difference lies between the successes and the failures.
And in the interest of attacking a problem - in this case, never writing a damn thing - from every possible angle, here I am writing. My thinking is that practicing, even in a different format, might lessen the barrier for scientific writing. Actually, I prefer to start with an easier format, such as writing non-consequential posts not intended for colleagues. If I were doing something equally difficult, I'll bet I'd never do it.
This discussion feels a bit rambly to me, but that's okay. Maybe tomorrow I'll do some editing and make a better version.