Comparing Extreme Programming and Waterfall Project Results

Only available on StudyMode
  • Download(s) : 87
  • Published : March 28, 2013
Open Document
Text Preview
Comparing Extreme Programming and Waterfall Project Results
Feng Ji
Carnegie Mellon University
Silicon Valley Campus
Mountain View, CA, 94035
jojojifeng@gmail.com

Todd Sedano
Carnegie Mellon University
Silicon Valley Campus
Mountain View, CA, 94035
todd.sedano@sv.cmu.edu

Abstract
Waterfall and Extreme Programming are two software project methods used for project management. Although there are a number of opinions comparing the two methods regarding how they should be applied, none have used project data to clearly conclude which one is better. In this paper, we present the results of a controlled empirical study conducted at Carnegie Mellon University in Silicon Valley to learn about the effective transition from traditional development to agile development. We conducted a comparison research against these two approaches. Multiple teams were assigned a project; some used Waterfall development, others used Extreme Programming. The purpose of this research is to look at advantages and disadvantages based upon the outcomes, generated artifacts, and metrics produced by the teams.

1. Introduction
1.1. Agile vs Traditional
Since the early 1970s, numerous software managers have explored different ways of software development methods (such as Waterfall model, evolutionary model, spiral model etc.) those have been developed to accomplish these goals and have been widely used by the software industry [1]. Methodologists often describe the Waterfall method as a stereotypical traditional method whereas they describe Extreme Programming as the stereotypical agile method. The Waterfall model, as the oldest traditional software development method, was cited by Winston W. Royce in 1970 [2]. He divided the software development lifecycle into seven sequential and linear stages: Conception, Initiation, Analysis, Design, Construction, Testing, and Maintenance. The Waterfall model is especially used for large and complex engineering projects. Waterfall's lasting impression upon software engineering is seen even in the Guide to Software Engineering Body of Knowledge which introduces the first five knowledge areas based upon their sequence in the Waterfall lifecycle even though the Guide does not recommend any particular lifecycle [3].

Although the Waterfall model has been adopted in many large and complex projects, it still has some inherent drawbacks, like inflexibility in the face of changing requirements [1]. If large amounts of project resources have been invested in requirements and design activities, then changes can be very costly later. High ceremony documentation is not necessary in all projects. Agile methods deal well with unstable and volatile requirements by using a number of techniques of which most notable are: low ceremony documents, short iterations, early testing, and customer collaboration. Kent Beck and Cynthia Andres define Extreme Programming 2.0 with many practices [4], like Pair Programming, Test First Programming, and Continuous Integration and so on. These characteristics enable agile methods to obtain the smallest workable piece of functionality to deliver business value early and continually improving it while adding further functionality throughout the life of the project [5].

1.2. PET project background
Carnegie Mellon University Silicon Valley students start their masters program with the Foundations of Software Engineering course. This course is team-based, project-based, and mentored. Each team builds The Process Enactment Tool (PET). The user personas are software developers and managers. The tool helps users plan, estimate, and execute a project plan while analyzing historical data. The tool's domain encourages students to learn about software lifecycles and methods while understanding the benefit of metrics and reflection. 1.2.1. PET 1.0: In 2001, Carnegie Mellon had one of the largest outsourcing firms in the world develop Pet 1.0. Later the student teams were brought in to do the...
tracking img