Engr 694-60: Research Methods in Computer Science
Spring 2014
Lecture Notes and Web Resources

Writing (and Rewriting) the Introduction to a Paper

In "How to Write a Great Research Paper" (discussed earlier in the semester), Simon Peyton Jones insisted an introduction should only do two things:

1. Describe the problem (Use an example to introduce.)
2. State the contributions (Make these explicit! They must be refutable statements.)

Peyton Jones advocates leaving out the outline paragraph that describes the "rest of the paper." Instead, he suggests working forward references into the statement of the contributions.

In "How to Write an Introduction: Some Suggestions," Sandro Etalle says an introduction should contain three parts:

1. Background (Establish the context.)
2. The Problem (Why keep reading?)
3. The Proposed Solution (State your contributions.)

Etalle's Background and The Problem parts are similar to what Peyton Jones meant by "describe the problem."

In " 'It was a cold and rainy night': Set the Scene for a Good Introduction," Thomas M. Annesley describes the introduction as a cone pointed downward (into the paper). Going from the larger (broader) to smaller (narrower) end, Annesley's cone-shaped introductions have four parts:

1. Background, known information
2. Knowledge Gap, unknown information
3. Hypothesis, question, purpose statement
4. Approach, plan of attack, proposed solution

Annesley refines Etalle's The Problem part as two new parts: Knowledge Gap and Hypothesis. This is similar to the structure suggested by Kristin Sainani in her Writing in the Sciences course (mentioned earlier in the semester).

Now consider Cunningham's introduction to his paper "A Little Language for Surveys." (He revised the introduction in an assignment in Sainani's course.)

The original paper has a weak introduction: lots of background but no explicit definition of the problem, no statements of the knowledge gap and purpose, and a weak statement of the approach (i.e., no explicit list of refutable contribution statements).

The first revision shortens the background without weakening the explanation of the context, better defines the knowledge gap and purpose, and states the approach and contributions more strongly.

The second revision (a) changes one sentence in the first paragraph and (b) reverses the order of the two sentences in the fourth paragraph to make the flow smoother.

The introduction (and the paper as a whole) should be refined further as suggested by Peyton Jones. The paper's contributions are still weaker than desired. The outline paragraph at the end should be better integrated into the statements of the contributions.

RETURN to Lecture Notes document?