How to Fight Complexity in Software (part I)

A couple of weeks ago, Grady Booch gave a lecture at Microsoft. It was a pleasure to hear of my software engineering heroes in person.  Grady talked about “the promise, the limits, and the beauty of software.”  

The main thing that captured my interest during the lecture was a discussion about the cost of complexity in software.

Grady said that there are many factors influencing the cost of software, but two of these factors have disproportionally high impact on the overall cost: a) the processes used to develop and maintain the software, b) the complexity of the codebase.  To fight growing complexity of the codebase, many companies allocate some percentage of their resources for refactoring, cleanup, and other activities that don’t directly result in features for the end user of the software. This really resonated with me. I am a huge fan of this approach. We do some of it here in the developer division (read about the MQ milestones here, here, and here), but I wish we and software industry in general did even more.