Sei sulla pagina 1di 1

The Crisis in Software:

The Wrong Process Produces the Wrong Results

Software development has till now primarily been driven by a slow, error-prone, rigid and
predictive model - the waterfall method, which heavily relies on complete planning before
execution begins. Its shortcomings were clearly presented in two case studies :
1. FBIs Sentinel project (attempt at digitizing all the case files for easier sharing)
2. Parametric Technology Corporation (which develops product lifecycle management
products)
As seen from these two examples, typical disadvantages of the method include
1. Delayed releases
2. Requirement of a lot of time to plan the project completely
3. Inability to drastically alter plans midway through
4. Difficulty in managing sudden changes
5. Issues in adapting new technology
which ultimately deteriorates the quality of the final software.
These are avoided by Scrum - a newly developed agile process. Scrum, in contrast to the
predictive method works in short sprints and can add features to the software in these
short chunks of time. Instead of following a sequential approach, Scrum completes the
software in an incremental manner. This led to a flexibility in the development process hitherto unseen in the previous approach. There was no fixed first draft and changes could
be made to the software at any point of time. As the Standish Groups 2011 report found
out, on an average, between the years 2002 and 2010, Scrum (42% completion)
outperformed Waterfall (14% completion) by completing three times more projects and
also leaving room for further changes. Needless to say, Scrum managed to solve both the
Sentinel and the PTC averting a huge loss.
Dr. Hank Walker talked about daily meetings when doing our project. These meetings
based on the Scrum agenda(three questions) lets us adapt to the ever changing
requirements of the user. Using this, there is always an incremental progress in the
development. Even if there is a hindrance, it could be solved together as a team thereby not
affecting further progress.
Yes, I agree that Scrum outperforms Waterfall in almost all aspects. The Sentinel project,
for example, saved 90% of the cost when completed using the new method. Since Scrum
compels frequent iterations, meetings, feedback and testing, it does better in performance
and cost as compared to the Waterfall method which has only a very few advantages.
My only query would be this : In, say, a startup company, a small team is trying to manage
multiple tasks and projects. Wont it become difficult to follow scrum in such a situation?
Alternatively, if there are a lot of people working on a single project, will coordination
between the different 2-pizza sized teams become an obstacle in itself?