Extreme programming (XP) is an agile software development methodology used to implement software projects. This article details the practices used in this methodology. Extreme programming has 12 practices, grouped into four areas, derived from the best practices of software engineering.
Pair programming means that all code is produced by two people programming on one task on one workstation. One programmer has control over the workstation and is thinking mostly about the coding in detail. The other programmer is more focused on the big picture, and is continually reviewing the code that is being produced by the first programmer. Programmers trade roles after minute to hour periods.
The pairs are not fixed; programmers switch partners frequently, so that everyone knows what everyone is doing, and everybody remains familiar with the whole system, even the parts outside their skill set. This way, pair programming also can enhance team-wide communication. (This also goes hand-in-hand with the concept of Collective Ownership).
The main planning process within extreme programming is called the Planning Game. The game is a meeting that occurs once per iteration, typically once a week. The planning process is divided into two parts:
The purpose of the Planning Game is to guide the product into delivery. Instead of predicting the exact dates of when deliverables will be needed and produced, which is difficult to do, it aims to "steer the project" into delivery using a straightforward approach. The Planning Game approach has also been adopted by non-software projects and teams in the context of business agility.
This is an iterative process of gathering requirements and estimating the work impact of each of those requirements.
When business cannot come up with any more requirements, one proceeds to the commitment phase.
This phase involves the determination of costs, benefits, and schedule impact. It has four components:
The business side sorts the user stories by business value. They will arrange them into three piles:
The developers sort the user stories by risk. They also categorize into three piles: low, medium and high risk user stories. The following is an example of an approach to this:
All indexes for a user story are added, assigning the user stories a risk index of low (0–1), medium (2–4), or high (5–6).
Within the steering phase the programmers and business people can "steer" the process. That is to say, they can make changes. Individual user stories, or relative priorities of different user stories, might change; estimates might prove wrong. This is the chance to adjust the plan accordingly.
Considering team velocity storypoints to be planned. Iteration duration can be 1 to 3 weeks.
The exploration phase of the iteration planning is about creating tasks and estimating their implementation time.
Within the commitment phase of the iteration planning programmers are assigned tasks that reference the different user stories.
The implementation of the tasks is done during the steering phase of the iteration.
Unit tests are automated tests that test the functionality of pieces of the code (e.g. classes, methods). Within XP, unit tests are written before the eventual code is coded. This approach is intended to stimulate the programmer to think about conditions in which his or her code could fail. XP says that the programmer is finished with a certain piece of code when he or she cannot come up with any further conditions under which the code may fail.
Test driven development proceeds by quickly cycling through the following steps, with each step taking minutes at most, preferably much less. Since each user story will usually require one to two days of work, a very large number of such cycles will be necessary per story.
For a more intense version of the above process, see Uncle Bob's Three Rules of TDD.
Within XP, the "customer" is not the one who pays the bill, but the one who really uses the system. XP says that the customer should be on hand at all times and available for questions. For instance, the team developing a financial administration system should include a financial administrator.
The development team should always be working on the latest version of the software. Since different team members may have versions saved locally with various changes and improvements, they should try to upload their current version to the code repository every few hours, or when a significant break presents itself. Continuous integration will avoid delays later on in the project cycle, caused by integration problems.
Because XP doctrine advocates programming only what is needed today, and implementing it as simply as possible, at times this may result in a system that is stuck. One of the symptoms of this is the need for dual (or multiple) maintenance: functional changes start requiring changes to multiple copies of the same (or similar) code. Another symptom is that changes in one part of the code affect lots of other parts. XP doctrine says that when this occurs, the system is telling you to refactor your code by changing the architecture, making it simpler and more generic.
The delivery of the software is done via frequent releases of live functionality creating concrete value. The small releases help the customer to gain confidence in the progress of the project. This helps maintain the concept of the whole team as the customer can now come up with his suggestions on the project based on real experience.
Coding standard is an agreed upon set of rules that the entire development team agree to adhere to throughout the project. The standard specifies a consistent style and format for source code, within the chosen programming language, as well as various programming constructs and patterns that should be avoided in order to reduce the probability of defects. The coding standard may be a standard conventions specified by the language vendor (e.g. The Code Conventions for the Java Programming Language, recommended by Sun), or custom defined by the development team.
Collective code ownership (also known as "team code ownership" and "shared code") means that everyone is responsible for all the code; therefore, everybody is allowed to change any part of the code. Collective code ownership is not only an organizational policy but also a feeling. "Developers feel team code ownership more when they understand the system context, have contributed to the code in question, perceive code quality as high, believe the product will satisfy the user needs, and perceive high team cohesion." Pair programming, especially overlapping pair rotation, contributes to this practice: by working in different pairs, programmers better understand the system context and contribute to more areas of the code base.
Collective code ownership may accelerate development because a developer who spots an error can fix it immediately, which can reduce bugs overall. However, programmers may also introduce bugs when changing code that they do not understand well. Sufficiently well-defined unit tests should mitigate this problem: if unforeseen dependencies create errors, then when unit tests are run, they will show failures.
Collective code ownership may lead to better member backup, greater distribution of knowledge and learning, shared responsibility of the code, greater code quality, and reduced rework. But it may as well lead to increased member conflict, increase of bugs, changes of developers mental flow and breaks of their reasoning, increased development time, or less understanding of the code.
Programmers should take a "simple is best" approach to software design. Whenever a new piece of code is written, the author should ask themselves 'is there a simpler way to introduce the same functionality?'. If the answer is yes, the simpler course should be chosen. Refactoring should also be used to make complex code simpler.
The system metaphor is a story that everyone - customers, programmers, and managers - can tell about how the system works. It's a naming concept for classes and methods that should make it easy for a team member to guess the functionality of a particular class/method, from its name only. For example a library system may create
borrowers(class), and if the item were to become overdue it may perform a make_overdue operation on a
catalogue(class). For each class or operation the functionality is obvious to the entire team.
The concept is that programmers or software developers should not work more than 40 hour weeks, and if there is overtime one week, that the next week should not include more overtime. Since the development cycles are short cycles of continuous integration, and full development (release) cycles are more frequent, the projects in XP do not follow the typical crunch time that other projects require (requiring overtime).
Also, included in this concept is that people perform best and most creatively if they are well rested.
A key enabler to achieve sustainable pace is frequent code-merge and always executable & test covered high quality code. The constant refactoring way of working enforces team members with fresh and alert minds. The intense collaborative way of working within the team drives a need to recharge over weekends.
Well-tested, continuously integrated, frequently deployed code and environments also minimize the frequency of unexpected production problems and outages, and the associated after-hours nights and weekends work that is required.
This article needs additional citations for verification. (December 2008) (Learn how and when to remove this template message)