"If I start early, it'll eat my life"

The late start has one real virtue, which is that it caps the job. The proposal took one night because there was one night. Begin two weeks early and it could take two weeks, every evening, fiddled with endlessly. In 1955 C. Northcote Parkinson opened an essay in The Economist with the line that work expands so as to fill the time available for its completion. He was joking about the civil service and it isn't a research finding, but everyone recognises it. If you lean towards perfectionism, you have special reason to fear it. For you the deadline does more than get you started. It's the only thing that ever makes you stop.

I take this seriously. If the alternative to the last night really were a fortnight of open-ended fiddling, the last night would be the better choice.

But look at what the deadline was doing for you. It limited the hours, it forced a decision about how big the job would be, and it told you when to stop. Those are three separate decisions, and you can make all three yourself, in advance, in daylight.

Decide the size. Before the second session, ask what this piece of work needs to be. Some things need to be excellent, many need to be good, and some only need to exist. A rough test is to ask what happens if it's merely adequate. If the answer is "nothing much", decide now that it's a four-hour job, and then do a four-hour job. The rushed version was always a small version. The only change is that you've picked what goes in it, with a clear head, instead of discovering at one in the morning what fell off.

Decide the hours. Give the work a budget, taken from the honest estimate in chapter 7. It might be twelve hours across the fortnight. Put them in the diary as sessions with an end time, and stop at the end time. Stopping on time is what teaches you that starting is safe. If any session might swallow the evening, you'd be right to avoid sessions.

Decide when it's done. Set your own finish a couple of days before the real one, and make it real by promising the draft to somebody. It could be a colleague who'll read it on Wednesday, or a friend you've told. A date that exists only in your head is easy to move, as you know. One that sits in someone else's diary is much harder to shift. I can't offer you a trial showing that this works, and the notes explain why a well-known one is missing. It's my reasoning, and it uses something you already know about yourself. You turn up for other people.

Then there's the perfectionist's objection, that you can't hand in something that's merely good. There's a whole book's worth to say about that, and for now I'll make one observation. You already hand in work that's merely good. A first draft written overnight can't be anything else. Starting early doesn't lower your standards. It's the only arrangement under which your standards ever get applied, because it's the only one with a second pass in it.

There's a gain here that's easy to miss. With a bad version done and sessions in the diary, the evenings between sessions are properly free. Chapter 4 described how an unstarted job follows you around. A job that's been started, with its next step written down and its next session booked, mostly leaves you alone. The habit was costing you leisure as well as quality, and this is how you get the leisure back.