Writing Paper Summaries for COMP
408
Paper summaries should be one typewritten page in length. The summary
should include the following information:
- The contribution of the paper.
While addressing this, you might ask yourself the following questions:
- If the paper gives an algorithm, what is the algorithm for?
- If the paper proves a result, what is the result in your own words?
- An explanation of why this contribution is useful.
While addressing this, you might ask yourself the following questions:
- Does the algorithm improve dramatically on earlier algorithms for the
same problem?
- In what ways is it an improvement (time complexity,
space complexity, etc)?
- Why is the problem approached by the contribution important?
- What does the result let us do that we couldn't do before?
- For papers giving experimental results, a brief description of
settings or examples for which the algorithm/result has been shown
to be useful and settings or examples for which the algorithm/result
has been shown to be poorly suited.
Do not address these questions individually. Write a single, cohesive
summary that covers each of these points. Make your summary as
concise as possible.
Some general comments from the first set of summaries
In general, you should write your summary for someone literate in
computer science, but who has never read the paper before. Give
enough information for the reader to understand what the paper is
proposing and to be able to evaluate whether this paper would be
useful for him/her to read.
Doing this well may require you to "go beyond" the paper. If you
don't understand some terms, look them up. If some detail is cited as
coming from another paper and it looks important, go to the library
and find the cited paper. Remember, you are presenting an overview of
the technical content of the paper - you may need to read a little
more than the paper itself to digest the content for yourself.
A couple of other comments:
- Don't quote statements from the paper. You should be giving me
your summary of the important details, not a sampling of statements
made by the author.
- Use formulas where appropriate. There is no need to explain
something in English that can easily be stated by a formula (such as a
running time).
- Spell Check!! I'm looking at your ability to communicate
technical details to others through writing. This includes grammar,
spelling, and writing style.
Sample: Summary for Bryant's OBDD paper