It was good to meet Michael and hear from a fellow practitioner on a topic where many of us need to improve. And Michael's primary source, McConnell's Software Estimation, is a good one. I was disappointed, though, that we spent so little time talking about the meat of the book - how to overcome the challenges of estimation (both technical and political) and head down the path of mastery of this important skill. Admittedly there is far too much detail in the Software Estimation book to cover it all in an hour, but Michael didn't even cover such foundational concepts as the idea of making uncertainty explicit by using ranges or confidence levels (so that we learn to do the research to be able to say things like "Based on our historical performance data, we have a 30% likelihood of hitting this goal - is this a level of risk we can accept?"). McConnell's Software Estimation book is all about going beyond gut feelings to numbers, AND learning how to present that effectively to management, but Michael pointed us to very little of that. I am particularly disappointed that in response to a question from the group about how to deal with management that may not appear open to adjusting the schedule or commitment, instead of sharing McConnell's amazingly helpful insights from the last chapter of Software Estimation's last chapter for just such situations, Michael spoke at some length about the perks of being an independent contractor relative to working for a company. Let's not jump to that conclusion, I say.

In case it helps, here's my review of Software Estimation.