For a long time, the software industry followed the notion that architecture was something that ought to be developed and completed before writing the first line of code. Inspired by the construction industry, it was felt that the sign of a successful software architecture was something that didn’t need to change during development, often a reaction to the high costs of scrap and rework that would occur due to a re-architecture event.
This vision of architecture was rudely challenged by the rise of agile software methods. The pre-planned architecture approach was founded on the notion that requirements should also be fixed before coding began, leading to a phased (or waterfall) approach where requirements was followed by architecture which itself was followed by construction (programming). The agile world, however, challenged the very notion of fixed requirements, observing that regular changes in requirements were a business necessity in the modern world, and providing project planning techniques to embrace controlled change.
In this new agile world, many people questioned the role of architecture. And certainly the pre-planned architecture vision couldn’t fit in with modern dynamism. But there is another approach to architecture, one that embraces change in the agile manner. In this view architecture is an constant effort, one that works closely with programming so that architecture can react both to changing requirements but also to feedback from programming. We’ve come to call this evolutionary architecture, to highlight that while the changes are unpredictable, the architecture can still move in a good direction.
At ThoughtWorks, we’ve been immersed in this architectural world-view. Rebecca led many of our most important projects in the early years of this millenium, and developed our technical leadership as our CTO. Neal has been a careful observer of our work, synthesizing and conveying the lessons we’ve learned. Pat has combined his project work with developing our technical leads. We’ve always felt that architecture is vitally important, and can’t be left to idle chance. We’ve made mistakes, but learned from them, growing a better understanding of how to build a code base that can respond gracefully to the many changes in its purpose.
The heart of doing evolutionary architecture is to make small changes, and put in feedback loops that allow everyone to learn from how the system is developing. The rise of Continuous Delivery has been a crucial enabling factor in making evolutionary architecture practical. The authorial trio use the notion of fitness functions to monitor the state of the architecture. They explore different styles of evolvability for architecture, and put emphasis on the issues around long-lived data—often a topic that gets neglected. Conway’s Law towers over much of the discussion, as it should.
While I’m sure we have much to learn about doing software architecture in an evolutionary style, this book marks an essential road map on the current state of understanding. As more people are realizing the central role of software systems in our twenty-first century human world, knowing how best to respond to change while keeping on your feet will be an essential skill for any software leader.