I try to quit when I'm stuck. Because when I'm stuck, I'm not really stuck, just tired. I think about my problem in bed until I fall asleep, and wake up often much closer to the answer.

If I get stuck significantly before bedtime, I take an extra shower or a walk or something, and that often has the same effect.

This works for me too. In addition, going for a coffee helps, as does splashing some water on my face.

In addition to quitting when I'm stuck, I also quit when I achieve. A git commit produces a nice starting point for the next day, and I have the Mortal Kombat guy tell me 'excellent' as a post-commit (yes really). Then I leave the startupcave feeling like a boss.

It's probably really emotionally satisfying to quit on an achievement, but I find it a lot harder to re-enter the next day without an unsolved problem. If I have an unsolved problem to start with I can build enough momentum to go into virgin territory, but starting the day without an unsolved problem can lead to hours of that blank page, writers' block, "what do I do now?" uncertainty.

I have a bunch of paper with checkboxes. Two or three check boxes a day, make that tick, hear the MK guy. But next day I've got another empty check box that I'm going to attack.

It's all about maintaining momentum. Sure, a guy can work till 5:00 AM but the next day is usually shot.

If you stopped at 9:00 PM though, and bottled up that momentum (your progress) till morning - you'll have a brain full of fresh energy for another full day.

I agree. I've had marathon all night coding sessions, but the next day is almost always wasted. I find now its better to step away once I find myself losing my train of thought.

This month I'm traveling and working from cafés. When the battery runs out, it's time to go home. That gives me about five hours of efficient hacking. The time constraint really helps me maintain focus.

I tend to go through periods of overwork and underwork in terms of productivity. Sometimes my most productive periods are actually quite short. Other times they're long.

Earlier this year I used to start work around 7:30am and finish around 9pm monday to friday, and work one day a weekend. At the moment I'm doing 8am starts, a break to travel to the office (in which I use a notebook and pen if I'm working to limit me to ideas and high level stuff, or do stuff on my phone), then I'm normally back for 7pm and work maybe one saturday in 4. I've also started working from home one day a week. I tend to find that's my most productive day, but if I spend more time at home it blurs the lines a little too much for me.

Last week I got in at 10am one day, knocked off at around 4pm. I hadn't had a day like that (with relatively little to do) for two years.

You should feel very lucky for having such struggles. It is a rare and wonderful condition.

I stopped looking for and setting goals, milestones, or anything else that was quantitive in measuring my progress a long time ago. Now I just listen to my gut and stop when I sense I am at a peak or my happiest.

I depends on what kind of work you're doing.

Meaningful programming is a creative endeavor.

If you try to do more than 8-10 hours of creative work a day, you're actually setting yourself back.

I try to average 8 hours of work a day, with bursts up to 12 hours if I'm on a roll or really excited.

Haha. One of my favorites is how Mike Arrington does. According to an interview, he'd keep working until he passed out :)

Even when you think you're tired, you always have more juice in you, especially when you're young.

This was an amazingly inspirational article when I first read it a few weeks ago. It clearly speaks to me since I've developed a similar style of working. Though I know how unhealthy it is so I'm trying to better myself in this way

for the lazy here's a link to said interview: http://www.inc.com/magazine/20101001/the-way-i-work-michael-...

When I think to myself: I could go on a couple hours more ,but I know tomorrow I'll be wasted. The extra hour will become 2 extra sleep hours the next day so , not worth it.

Working until passing out feels like an illusion of accomplishing more. Instead of keeping track of how much hours were spent typing, I just measure how many features/checks on the todo stack I did on X day.

Set some goals for the day. Stop when you've either accomplished them all, or when you don't feel like you can finish them anymore. Then set some more for tomorrow and go do something else.

Alternately, when you've been in front of the computer for 8(ish) hours.

Be in bed by midnight, the circadian clock is your friend

http://en.wikipedia.org/wiki/File:Biological_clock_human.svg