Project Lionheart: The Epic Conclusion

Mayor van Damme is pleased. We completed the project only a few weeks behind our original schedule despite the constant stream of changes that occurred nearly from the start. We satisfied all the high-priority quality attributes, though there were a few hiccups during load testing before the official public release. Other than that everything went well.

The team strongly feels that the foundation provided by our architecture made it all possible. Our decision to do less architecture design up front worked out fine. The team had prior experience with most of the frameworks and technologies we used. In the beginning, we focused on the areas of greatest risk and worked to build consensus around the architecture. We had a few lucky breaks, such as stumbling upon a major problem with two web services, but we found the problems early enough that we had time to rewrite them from scratch. The final iterations of the project were stressful but not unbearably so.

Our last official work on Project Lionheart is to create maintenance documentation for the city’s IT department. The team is writing a moth ball document, a user’s guide, and cleaned-up versions of relevant architecture views. The original architecture driver’s specification maybe only 50 percent accurate at this point. We never created a formal architecture description. We’ll retain these documents for historical purposes and create a new teachability viewpoint to walk new developers through the architecture as it exists in the code.

The mayor’s office estimates the application we built will save the city nearly $1 million in the first year. It feels good to see how our design decisions help the city meet its goals in the real world.