I loved this quote because it points at one of the biggest problems that I think the software development industry has with estimates, and one of the reasons that they are, so often, so wrong.Before asking for an estimate first ask yourself what decisions the estimate will be informing.
Too often we estimate work in a vacuum without adequate knowledge as to why the estimate is being asked for, or how it's going to be used.
There are many different, and valid reasons to estimate work, and different situations call for different techniques. For some of these cases the trivariate estimate method that I suggested in a previous blog is a great answer that will provide low level detail about the estimate and allow good decisions to be made, for other cases it is a massive waste of time. Similar ideas can be drawn for the story points, t-shirt sizes or the no estimates movement.
Each type of estimate is suited to a situation, and you can never know what is needed unless you can understand why you're being asked for the estimate and what decisions that estimate is informing.