Myths about object-orientation and its pedagogy

John LewisACM SIGCSE Bulletin 2000 · DOI

TL;DR

This paper debunks common misconceptions surrounding object-oriented programming (OOP) and its pedagogical application, particularly in introductory computer science courses. It argues that many "myths" hinder effective teaching and learning by misrepresenting OOP's history, its relationship with procedural programming, and the practicalities of its implementation.

What Problem Does This Paper Address?

The paper addresses the challenge of effectively teaching object-oriented programming (OOP) in introductory computer science courses (like CS1). As OOP gains traction, a set of "myths" or misunderstandings has emerged, often propagated among newcomers. These myths can lead to confusion, incorrect pedagogical approaches, and a failure to grasp the true essence and benefits of OOP. The authors aim to clarify these misconceptions to promote a more accurate and effective understanding of OOP for both educators and students.

Methods

The paper employs a critical analysis and argumentative approach. It identifies common "myths" about OOP and its pedagogy and then systematically debunks each one by providing historical context, logical reasoning, and practical examples. The authors draw upon established programming language history, fundamental computer science concepts, and pedagogical considerations to support their arguments.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • OOP is not new: Concepts like classes, subclasses, and polymorphism have roots in languages like Simula, dating back to the 1960s, not just recent languages like Java.
  • Procedural and OOP are complementary, not exclusive: OOP augments and strengthens procedural concepts like modularity and encapsulation, rather than replacing them. Objects can be seen as the ultimate Abstract Data Type (ADT).
  • "Objects First" needs clear definition: The phrase is often used loosely. The key is introducing the concept of objects, not necessarily requiring students to write complex classes from day one. Using existing objects is a valid starting point.
  • "Hello World" can be object-oriented: Even a simple "Hello World" program in Java utilizes objects (e.g., System.out) and can serve as an opportunity to reinforce OOP thinking.
  • Control structures are orthogonal to OOP: Concepts like loops and conditionals are low-level tools that define a single function and are independent of whether that function is part of an object or a standalone procedure.
  • Java I/O is complex but necessary: Java's extensive I/O library, while seemingly complex, is designed for versatility and enterprise development. Its complexity stems from its full use of object abstraction and broad interaction types.
  • GUI abstraction needs careful handling: While third-party classes can abstract complexities, they should not distort the fundamental model of GUI development. The goal should be to teach the underlying concepts of user interaction.
  • C++ is hybrid, not purely OOP: C++ supports OOP but doesn't enforce it, allowing for procedural styles that can obscure OOP concepts.
  • Java is nearly pure OOP: While all executable code must be in a class, Java's primitive types mean not everything is an object. More importantly, the design can still lean procedural if not carefully guided.

Key Results

The paper's primary "result" is the debunking of several common myths. It doesn't present quantitative results but rather qualitative arguments that aim to shift pedagogical understanding. The key takeaway is that educators should focus on the fundamental concepts of OOP and avoid rigid interpretations that can hinder learning.

Limitations and Future Work

The authors acknowledge that the definition of "objects first" can vary widely, emphasizing the need for clear communication among educators. They suggest that while third-party classes can abstract complexities, the ultimate goal should be to teach students real GUI development and user interaction concepts. The paper implicitly calls for more consistent and accurate pedagogical approaches to OOP, suggesting that future work could involve developing standardized curricula or best practices that address these myths.

Summary updated

Abstract

Object-orientation continues to find a home in computing curricula, especially in early courses such as CS1. As this trend continues, some ideas seem to take on a life of their own, despite being fundamentally incorrect. Unfortunately they propagate most quickly among those who are relatively new to the ideas of object-oriented development. This paper enumerates and debates the underlying issues of several myths regarding object-orientation and its pedagogy.
BibTeX
@article{lewis2000myths,
  author = {John Lewis},
  title = {{Myths about object-orientation and its pedagogy}},
  journal = {ACM SIGCSE Bulletin},
  year = {2000},
  doi = {10.1145/331795.331863},
}

Loading related content...

Figure