CS 4448 - Fall 1998
Object-Oriented Programming and Design
Talk 3.3
by
Adam Kaupisch
Pitfalls of Object-Oriented Development
by
Vivek Shah, Marcos Sivitanides, Roy Martin
-
Introduction.
Basically, the authors point to this article was to show what some
of the common pitfalls of OOD are and how to avoid them.
-
They had 7 main points
-
Conceptual Pitfalls
Problem: Confusion about what OOD is and what it entails.
Solution: Education and experience are key to the success of a project.
Therefore, it is key to fully understand how OOD works.
-
Political Pitfalls
Problem: These pitfalls deal with the political infrastructure within
a company. Engineers expect time and thoroughness whereas management
wants it now.
Solution: Educate management and explain how OOD works. Also,
engineers need to gain proper support from key figures within the company
to help better explain what they need.
-
Analysis and Design Pitfalls
Problem: Developers take a traditional design approach and then attempt
to implement it using OO techniques. Or, they take an OO approach
and then try and implement it using traditional methods.
Solution: If the project is going to use OOD, everyone involved should
study OOD before they begin.
-
Environment, Language and Tool Pitfalls
Problem: Incompatibility
between certain languages and tools
Solution: Developers need
to consider what technology is really needed before they begin and validate
that
those technologies will
really work.
Problem: OOAD is easy
therefore the perception is not to spend much time doing it and to rush
the process.
Solution: Insure that you
take your time, follow every aspect of OO through.
-
Class and Object Pitfalls
Problem: Most developers
don't fully understand inheritance and different relationships between
objects
Solution: Developers need
to set guidelines before they begin.
-
Reuse Pitfalls
Problem: Long term vs short term payoff of code reuse
Solution: Developers need to plan ahead for possible code reuse before
they begin.
-
Conclusion
There are many possible pitfalls using OO methodologies, however with
proper training, understanding and education, most of these pitfalls can
be avoided.
Copyright © University of Colorado. All rights reserved.
Revised: September 17, 1998