The Waterfall Process Distilled
Based on the US Department of Defense standard DOD-STD-2167A, we have a well-defined process often referred to as Waterfall. If you are not familiar with the process, here is a short introduction.
A project in the waterfall process goes through four phases before the project is completed.
The first phase is the naïvite phase. This phase should always last 12, 18 or 24 months. 18 months is recommended. There is a detailed plan showing how, at the end of the phase, the system will be done.
After 18 months and 1 day, the project notices two things: 1) They’re not done, and 2) nobody had imagined this possibility.
This is the introduction to the second phase, which usually is the most frustrating of the project: The Panic phase. The Panic phase always lasts 2 or 6 months. 6 months is recommended. During the panic phase, a new plan is made, and everyone works overtime and weekends.
After 6 months and 1 day, the project notices two things: 1) They’re still not done, and 2) nobody cares anymore.
Thus begins the third phase of the waterfall project: The Sliding phase. At this point in time, nobody cares to update the plan and everyone continues to work, although at a more disillusioned pace. The sliding phase lasts for as long as it lasts, but seldom less than 6 months. Cases of up to 2 years have been documented.
The sliding phase ends when somebody finally puts the foot down, and the project enters the final phase of the waterfall cycle: The iterative phase. At this point in time, the project has usually produced quite a bit of software, but they have no idea of how to finalize it. The project comes up with a plan to deliver the project in increments of no more than 3-6 months, and manages to deliver the project within a year in several increments, thus completing yet another successful waterfall project.
Comments:
Christian Rørdam - Feb 19, 2007
This sounds very familiar, all though all projects I have attended were supposed to be iterative from the start…
Johannes Brodwall - Feb 19, 2007
I have been wondering if this process leaves both proponents of waterfall and iterative development feeling vindicated. With waterfall-tinted glasses, it looks like the project was successfull, but the plan “failed a little”. With iterative-tinted glasses, we might not notice that many iterative projects build on the architecture of “failed” waterfalls…