How to read computer science papers

Situation: You have a large pile of computer science papers in front of you. You want to read them all. What to do?

My suggestion is that you read the two guides below. They are really short and helpful. I'm one year into my CS PhD, and I still find reading a large pile of papers to be quite hard. Especially if the papers are exploring problems within a field that I'm not super familiar with.

Tip: Actually the guides can also be used by people who want to get better at writing computer science papers...

Guide #1: Comprehension, Evaluation, Synthesis

How to Read a CS Research Paper? is a guide written by Phillip W. L. Fong at the University of Regina, Canada. It focuses on three sets of questions you should ask yourself while reading a paper.

  1. Comprehension: What does the paper say?
  2. Evaluation: Is something wrong with the paper?
  3. Synthesis: How can you improve the results?

It concludes by a section on how to write a summary, which will sharpen you understanding of the paper. I recommend reading the full guide. It's less than four pages long.

Guide #2: Three-pass

How to Read a Paper is another guide from the country of maple sirup. It's written by S. Keshav at the University of Waterloo. It describes the three-pass method of reading papers.

The three passes are:

  1. Get a general idea about the paper (5-10 minutes)
  2. Grasp the paper's content, but not its details (< hour)
  3. Understand the paper in depth (1-5 hours depending on experience)

It also has a great recipe you can follow for doing a literature survey of a research area. I highly recommend the tips in this section.

Perhaps a detail, but the author prefers a large amount of time between passes (if time permits). If you have a large batch of papers to read (like I currently do), do a first pass over all of them to get an overview. Then later, do a second pass over a subset of the papers that you find interesting. Then even later, a third pass over the subset of papers, where you really want to deepen you understanding.

Slightly off topic

Someone asked a question on cstheory asking what papers should everyone (who's interested in theoretical computer science) read?

There is good discussion about the value of reading original papers versus modern expositions.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.