Maybe it’s my military background, or maybe I’ve always had this predilection, but I like to use two devices when I need to make a complex decision: A checklist and a decision matrix. I like to use a checklist because it ensures that I remember the big bits of what I need to do, and brings up questions or areas that I didn’t think about when evaluating options for the decision.
And the decision matrix – that’s the thing I use to actually lay out those options. It’s simply a spreadsheet-like grid (I use Excel, but paper and pencil works as well) that lays out the requirements or advantages for the decision across the top, and the options I have on the left-hand side. Then in the “cells” I put whether or not that option on the left will meet the requirement in that column. I then simply “weight” each cell to organize the choices by best-fit. The right answer (or answers) will float right to the top.
I was asked yesterday about options for moving data in SQL Server to another system. There are just dozens of ways to do this, from bcp to Replication, each with certain advantages and costs. But asking the questions for the top row first helped me show the person that it isn’t a particular technology that is important, it’s laying out those requirements and thinking about which elements are more important than the other. For instance, is it more important to have the data moved all the time, or is it OK if that happens once in a while? Does the data have to move in two directions or just one? All of these will help that answer jump right out. Try it sometime – it’s a great learning exercise, since it will force you to focus on filling out the matrix.
The answer is out there, Neo.