I am not sure whether this is the right place to ask this, but I have seen some extremely helpful comments the job of mathematicians as opposed to mathemetics itself on this site. If you think this question is inappropriate, I shall delete it.
I am leaving my job as an optical engineer to care for my children and hopefully, in the two days a week I have arranged for myself to myself, make some modest contribution to mathematics through a late undergrad / grad exposition on Lie theory - a subject that has interested me for the best part of 20 years. With little children, time will be very precious to me and I shall have to work as smart as I can to get the most out of what mathematics time I give myself.
I would like to gather comments on mathematicians' work habits into a Community Wiki so that people can see both (1) what works for other people or (2) what people believe would be best, even if they don't do it. I think this question is likely a bit different for scientists like mathematicians or theoretical physicists from working in general - such scientists do need to spend time thinking deep and fine details - almost meditating on a problem.
I personally tend to get a bit obsessed by problems and have difficulty throwing them out of my head: I'm sure if I could muster the self discipline to do this regularly, I might not get stuck on dead-end ways of thinking so much. Luckily I have a delightfully pestering little daughter who wrenches my attention away.
To answer this question, I would be interested to hear both what you do, and, if you're like me, what you think might work better for you but which you find yourself not attaining through some personality trait. Also, if people know of research that attempts to compare the effectiveness of different working techniques, that would be really interesting too.
On George Mackey (see some reflexions here)
"George supposedly spent eight hours each day on his research, in addition to his teaching and other obligations. He told me once that his method was to work for forty-five minutes of each hour and then spend fifteen minutes of each hour reading a nonmathematical book or taking some other kind of break. Knowing this has always helped me to pace myself."
On the other hand, if the popular account in Paul Hoffman's "The Man Who Loved Only Numbers" is to be believed, Paul Erdős did little else aside from maths, and it certainly didn't harm his work! However, even in Erdős' case, this has to be taken with a grain of salt - the same popular account showed Erdős to be well versed in world affairs and politics, so he must have taken some recreation outside mathematics and indeed he found joy in playing with small children or even just in the taste of grapefruit.