Jacob Kaplan-Moss wrote about his approach to software project estimation which is an important Project Management tool. His writeup is straightforward and clear, and the approach itself is the same: for each task, estimate size and uncertainty (as a multiplier). Then to estimate, you can add them up with and without the uncertainty multiplier to get best and worst case estimates.
Estimating software projects can be hard, but it’s something that people have always had to do and have come up with tricks for it. Project Management often requires having a sense of when we’ll finish. Dan North describes Blink Estimation, which involves getting experienced people in a room and facilitating a quick discussion to get an order of magnitude estimate on the work. At the other end of the spectrum, Erik Bern looks at some statistics behind why software engineers are pretty close on the median size for tasks, but are off on the mean and, therefore, the overall.
I enjoyed Lukas Linhart’s Project Management article there are no bugs, just TODOs. I agree with a lot that’s there (a well-defined ordering for doing things, bugs and features essentially being the same type of task). I’m torn when it comes to “close old issues”. If the UI of the issue tracker does a good job of not shoving those old issues in your face all the time, there’s no real harm to them remaining open. It is sometimes the case that an old issue that didn’t used to be important can become important.