Agile Journal - Resolve to Increase Agility

Home

Agile Journal

Achieving Global Agile Development - December 2007

Resolve to Increase Agility

Written by Kirk Knoernschild

Saturday, 05 January 2008

Last year, I resolved that in 2007 I'd focus on the essential characteristics of technology that help maximize the effectiveness of agile practices. Since about 90 percent of folks don't keep their resolutions throughout the year, I've decided this year not to make any such promises. Instead, I'll let you make your own resolutions this year, and will offer up a few tips that you might want to consider seriously if you're interested in increasing your personal or team's agility.

Team Agility
Increasing team agility is incredibly difficult as it requires a shift in thinking by all members of the team. There is no magic bullet. For those teammates who might be most comfortable with a waterfall lifecycle, transitioning to an agile approach is fraught with doubt. But adopting a few simple practices, while steering clear of others, can help increase your team's agility without ever mentioning the "a" word.

Ignore Traceability. Many teams are quite enamored with traceability. Changes to requirements can be traced through the analysis and design model into the code through the components and onto the executable nodes. But traceability is fundamentally flawed due to synchronization issues across artifacts. Teams do not possess the discipline necessary to maintain consistency across all artifacts, making the source code the only true artifact representing the current state of the system. If you're not on a team that values traceability, then you're off to a great start in 2008. But if you are, admit that traceability is a myth and turn your attention toward the source.

Test Early. I've worked on, observed, and spoken to numerous teams interested in increasing their agility. Some had a robust continuous integration strategy in place. Some strove to achieve 100 percent test coverage. Some refactored mercilessly to ensure a resilient and extensible design. And some developed using timeboxed iterations, engaged the customer throughout the lifecycle, and planned continuously. Yet, not all of these teams felt agile. I've also worked on teams that have felt far from agile, in that their practices weren't consistent with the vision of the agile manifesto, nor were they considered "best," but these teams always seemed to create software that did what their business partners needed. What was the differentiating factor? Teams that create great software test early and test continuously. To test, you must have two critical components in place - a tester and an application to test. The earlier you test your software, the sooner you prove you have software that works. As testing requires a tester, Quality Assurance gets involved very early in the project lifecycle, and the big bang approach to integration and series of handoffs that often take place on traditional projects are avoided.

Allow Emergent Roles. Teams often assign individuals to specific roles. In some situations, this makes sense. Not all developers make a good project manager, and if you know you need a good project manager, then you need to find someone capable of managing your project, whether that individual be found through a new hire, contractor, or from within your team. But in other cases, individuals are prematurely assigned to specific roles. Anointing someone software architect is but one example. While other members of the team might be more qualified, they are not allowed to fill the role due to organizational issues, workload, or resource constraints. Assigning roles where each role has specialized responsibilities doesn't leverage the individual strengths of team members. In The Agile Matrix, I spoke of various ways to organize team structure, and allow roles to emerge naturally from within the software development team results in greater team cohesion. Instead of starting big with individuals allocated to certain roles, start small where each member of the team is a developer. Team members will naturally gravitate toward their individual specialty based on their strengths and interest. Then, as gaps are identified, search for additional developers to fill those gaps.

Build Often. The only true barometer of project progress is a functional software application. If you don't have a functional software product, you cannot accurately assess the current state of the project. A robust continuous integration strategy is the cornerstone of any team getting started with agility. To test you need a software system to test, and if you can't build it, you can't test it.

Avoid Contracts. A typical software lifecycle begins with IT developing estimates that help gauge how long it will take to deliver the software system. Once they've provided acceptable estimates, business and IT draw up a contract and include other important agreements, such as Service Level Agreements (SLA). Once the contract is signed, work commences. Contracts ensure that IT and business will always be on different sides. If you want to create greater team cohesion between IT and your business partners, avoid contracts like the plague. Instead, use functional software, frequent demos, and constant customer collaboration to guide the development effort.

Developer Agility
Increasing your personal agility impacts how you, as an individual developer, go about doing your work every day. If you're interested in increasing the quality of code you create, or delivering more valuable features to your users, a good place to start is by practicing some of the items below.

Run the Tests. If you're not already writing unit tests, you need to start, and that should become your first professional resolution of the year. If you are writing tests, be sure you run them multiple times per day. Ideally, each time you make a change to the code.

Avoid BDUF. If you practice test driven development, chances are you're already an advocate of emergent design. Perhaps, however, you've believed that unit testing is really only about testing. If you're still spending considerable time crafting UML diagrams or other design documents, try foregoing their creation just once and allow the design to emerge. Instead, spend a bit of time sketching on a piece of paper your high-level design, and let the tests drive the design by placing emphasis on testing classes in isolation. It's a natural way to ensure classes are decoupled and methods are cohesive, and with a robust suite of tests in place, you'll be much more effective at the refactoring you'll need to do regardless of the approach you take toward design.

Write Quality Code. It should be every developer's professional code of conduct to write the highest quality code he or she is capable of writing, but writing high quality code is difficult. And in many situations, I doubt that we ever set out to create intentionally crappy code. Instead, it just winds up that way. Aim high to create code that is expressive, well-tested, and highly cohesive. Ensure designs are resilient and extensible. Use tools that can help you.

Avoid Big Tools. For some odd reason, we're led to believe that the larger the software system, the larger the tools we need. We treat large software development efforts like we do large construction projects. The larger the project, the bigger the equipment and tools needed. Yet big tools are often cumbersome and difficult to use, and often involve an extensive learning curve. In last years resolution, I laid out some guidelines for choosing tools that support an agile developers mindset. Keep these in mind when choosing your tools, and only use a tool when you necessarily need to.

Automate. Performing mundane repetitive tasks takes time away from focusing on more important software development activities. In The Value of Executable Artifacts, I presented numerous techniques and tools that can be used to automate processes and increase your productivity. The first thing you might want to automate is execution of those unit tests if you're lacking the discipline to run them regularly.

Building Bridges
If you're able to pull it off and stick to your resolutions longer than a fortnight, start thinking of ways your personal agility can help contribute to greater team agility. For instance, if you're writing tests but know that other members of the team are not, explain to them the benefits you've realized through unit testing and encourage them to experiment. Let your ability to refactor easily and the increased quality of your code speak for itself. Pair up with skeptical developers and let them experience the value of unit testing with you. If you've discovered an aspect of development that benefits greatly from automation, share it with the rest of the development team and show how it's helping increase your productivity or the quality of your code.

In less agile environments, once you've increased your personal agility, work with project management to grow these ideas within and throughout software development teams. For instance, maybe you've realized a great deal of success with improving the quality of your code and design through the use of static source code analysis tools. If that's the case, perhaps there is opportunity to work with project management to leverage the results team-wide and forego the waste that often comes with emphasizing traceability. Regardless, once you've increased your personal agility, spreading your wisdom, however challenging, to other members of the team is especially gratifying for both you and your customer.

The New Year is upon us, and with it comes the opportunity to improve. By making a few simple adjustments and adopting some proven practices, we can create an opportunity to excel. And once we've improved ourselves, possessing the resolve to spread your wisdom throughout the team can be incredibly rewarding.

About the Author
Kirk Knoernschild is Senior Technology Strategist at TeamSoft, where he leads based on his firm belief in the pragmatic use of technology. In addition to his work on large development projects, Kirk shares his experiences through courseware development, teaching, writing, and speaking at seminars and conferences. Kirk has provided training to thousands of software professionals, teaching courses on UML, Java J2EE technology, object-oriented development, component based development, software architecture, and software process.