Mistakes in Agile Development

Three main mistakes

The first is focusing on process ignoring people. The agile declaration enlightened the meaning that "people and communication are better than processes and tools ". But there are still many enterprises trying to implement Agility by creating a perfect process. It is undeniable that a reasonable process plays a certain role in improving the efficiency of the team. However, to truly benefit from agile improvement, enterprises must implement human changes.

The second is to put emphasis on management and contempt for engineering. Many teams equate agility with opening a site meeting, performing iteration, and reviewing. In the end, everything is in form. In the end, agility is a regression to the understanding of software technology. Therefore, continuous integration, automated testing, design, and restructuring cannot be achieved. Without solving problems from these fundamental techniques, various dazzling communication means actually increase the team's cost in vain.

The third is the process of ignoring indicators. Many teams, especially from CMM-type organizations to agile teams, are keen to design the so-called agile measurement system. Measurement should be to help the team enhance confidence and continuous improvement, and indicators should not be the goal. We should not only focus on where we are, but also on where we are going.

There is no panacea for solving these problems, and there is never a perfect process that fits all corners of the world. We have summarized several improvement modes in the process of helping enterprises improve agile processes. Here we will share with you.

Five improvement Modes

SKIP: the entire team is switched from one practice to another.

In some practices, we know the goal and the impact of such a switch on the team is small, or it is not suitable for adopting a gradually implemented solution.

For example, switch the configuration management tool. The configuration management tool used by a team is clearcase. in order to enjoy the advantages of SVN, such as atomic submission and low-cost branch, we often adopt the Skip method, that is, the entire team immediately switches from clearcase to SVN. This is because the switch of configuration management is generally controllable and easy to use.

Parallelism: Some organizations or individuals in the team use the original practice, while the other part switches to another practice method.

Some practices know their goals, but the implementation of the entire team may have a great impact on the way of work, or some organizations or individuals in the team do not have the conditions to switch to the new practical approach, we adopt a parallel approach.

For example, continuous integration of the project team. A project team in the team has a good foundation for continuous integration, while others have a poor foundation. If continuous integration is implemented across the entire team, it may have a great impact on the team. In addition, the specific scheme of continuous integration may need to be tested in a project group. At this time, we will first work in a project group and then promote it to other project groups.

Stepping Stone: to achieve a long-term goal, we must first achieve a closer goal.

Some practices have long-term goals, but you still cannot see the path to achieving the Goals. If you know that doing something will definitely help you reach the goal, you can do it first. For example, unit test. Sometimes you want to implement unit tests in some teams, but you do not need a suitable testing framework, so it is difficult to start the subsequent work before determining the testing framework. In this case, we can fully build this test framework. This method is called a stepping stone.

Simplified: to achieve a long-term goal, we must first achieve a more easily achieved goal, but this goal can still bring us benefits.

Some practices have a long-term goal, but it is still unclear about this long-term goal, or it is still a long way to go to achieve this long-term goal. If some practices can bring us some benefits, although they are not as good as the long-term goals, we can do it first.

System testing and low-cost testing. System Test and low-cost test pull are long-term and arduous tasks. For example, good infrastructure cooperation may be required. But we can do something simple to get similar benefits.

Suspend: a practice is not implemented for the moment.

In some practices, the team does not have the implementation conditions or has a great impact on the team. You can choose not to implement them for the time being. For example, TDD. The implementation of TDD requires high conditions. If the team does not have such a condition, it is difficult to implement it rashly. In this case, it is often paused.

The dimensions of the preceding five modes are different. For example, parallelism is considered from the organizational dimension, while factorial and simplification are considered from the objective dimension. In addition, the same practice may belong to different models in different scopes. The purpose of this model is to help participants clearly understand the nature of the problem, understand the limitations of the solution in terms of time and scope. Therefore, we do not advocate binding practices and models.