I fear writing; a result of my perfectionism. The prospect of defective prose frightens me, so I don't try. A thought I put to paper last month:

"I am imprisoned within my lack of clarity and lack of style. Fucking failure. Chained to a quiet and passive existence.

I want to express; I want to free my ideas; I want to communicate my message; but I fucking can’t. I can’t clear this painful fog."

I just decided to write a lot. Even if it doesn't make sense (http://blog.uncool.in/2009/03/26/episode-1-perspective-shift...). I suppose after a month or so of writing crap, I'll be able to loosen up my rusty writing skills a bit and my writing will become more natural.

You might want to try that.

I agree. Commit to writing 3 pages a day, even if it is just repetitive nonsense and ramblings. Feel free to burn the notebook at the end of the month. After a few months of this, you just may find you have loosened things up a bit in terms of the old "the best is the enemy of the good" problem.

I have undertaken that same task; working at it for about a month. It gets easier. (http://dailymis.com/?p=55)

Rudolf Flesch recommended

http://www.amazon.com/Write-Speak-Think-More-Effectively/dp/...

choosing someone whom you'd like to write to and who would write back (a parent or a grandparent, perhaps?) and then writing that person a letter with lots of description of what you're thinking about and doing once a week. After a year of that, you'll write a lot better.

I'm often faced with the following choice: write a paper about Algorithm 1.0, or try to figure out Algorithm 1.1? Or perhaps tweak the code to make it cleaner/faster/more general?

Writing the paper usually loses.

Write paper 1.0. Then work on 1.1. Then, instead of having one paper, you have two, and are that much closer to tenure. :)

In theory, that is the optimal procedure. But writing code is so much more interesting right now.

It's my most productive form of procrastination.

This is why deadlines can be good things.