Many years ago, I was a manager at a large company. The network was in trouble: there were constant failures due to a poor existing design, as well as political issues galore. I had a group of very smart engineers working for me, and they all had ideas about how to make things better. The engineers all believed their changes should be implemented, solely because they thought their ideas were sound (some were, others weren’t). My problem was how to get the engineers to understand that they could not simply make any changes they wished.
In an effort to get the engineers to understand that not all changes are good changes, and to encourage them to think about the implications of the changes they were proposing, I came up with three rules for them to follow. The rules were simple—each only a word. For me to consider a change, it had to follow one of the three rules. The rules were:
GAD’s Maxim #2: The only valid reasons to change a properly sized production network are simplification, standardization, and stabilization.
If the proposed change did not accomplish one of the three stated goals, I would not even consider the change request. At first, the engineers were constantly frustrated. They wanted to effect change, but they were not being allowed to do what they thought was needed. Over the course of about a month, however, they started to catch on. Within six months, we had so greatly improved the network that it was no longer the source of any ...