It's more that software is already so unpredictable. A 90% confidence estimate for a task might be 4 hours to 4 months. That's not very useful. Counting routine work would increase accuracy to something like 2 days to 4 months.
I'm not sure if it's better, but we can make some plans around that. The extra control is a little comforting. And it's less demoralising when it's not done by the end of the day.
I like doing it because I can see how much time it takes to fit in my day, since I usually time block my day. It might be routine so I can estimate the time but if I have a lot of tasks, I might over or underestimate and not get everything done that I'd want to.
Comments
Can you say more about why it's useful to time routine work?
That's what I'm curious about. A consistent set of steps that you're committed to doing on some frequency—do you see value in timing things like that?
It's more that software is already so unpredictable. A 90% confidence estimate for a task might be 4 hours to 4 months. That's not very useful. Counting routine work would increase accuracy to something like 2 days to 4 months.
I'm not sure if it's better, but we can make some plans around that. The extra control is a little comforting. And it's less demoralising when it's not done by the end of the day.
I like doing it because I can see how much time it takes to fit in my day, since I usually time block my day. It might be routine so I can estimate the time but if I have a lot of tasks, I might over or underestimate and not get everything done that I'd want to.