Contents (hide)

Performing Research

Reference: The Art of Computer Systems Performance Analysis by Raj Jain

Good researchers pay great attention to detail…

  • when designing and running experiments
  • when analyzing data
  • when creating graphs
  • when writing papers
  • when preparing and giving presentations

You must have a plan

  • State goals and define the system
  • List possible outcomes
  • List possible parameters and variables
  • Select metrics to study
  • Select input model
  • Design experiments
    • justify parameter settings
  • Analyze and interpret data
    • do the results make sense?
    • can you explain them?
  • Present results

Documentation is essential

  • Get a lab notebook and use it!
  • Document experiments
    • why was the experiment run?
    • what were the expected results?
    • what were the experiment parameters?
    • what were the results?
    • write a one-page summary before presenting results to your advisor

Reading Papers

Note: To access papers from ACM Digital Library and IEEE Xplore when off-campus, append 'proxy.lib.odu.edu' to the hostname in the URL. You'll then be prompted for your ODU UIN and password before proceeding.

Reference: How to Read a Research Paper by S. Keshav
First Pass

  • title, abstract, introduction, section headings, conclusions, references
  • answer 5 Cs: category, context, correctness, contributions, clarity

Second Pass

  • entire paper, ignoring details such as proofs
  • look at figures, graphs

Third Pass

  • entire paper, identify and challenge every assumption in every statement

Reference: How to Read an Engineering Research Paper by William Griswold
Answer these questions:

  • What are the motivations for the work?
  • What is the proposed solution?
  • What is the evaluation of the work?
  • What are the contributions?
  • What are the future directions for this research?

Links

Writing Summaries

Reference: Summary of a Scientific Article, George Mason Biology Dept.

  • Turn answers to the questions into a summary
    • summary must be in your own words
  • Don’t “cut and paste” the article
    • If your summary is a "jumble of statements nearly straight from the article", then you haven't really understood what the article was about.

Writing Papers

Reference: Writing for Computer Science by Justin Zobel

Organization

  • Abstract
    • single paragraph
    • readers use it to determine if article is relevant
    • concise summary of aims, scope, conclusions
  • Introduction
    • describe topic, problem/motivation, approach, scope, conclusions
    • clearly tell reader what is novel
  • Related Work
  • Approach and Results
  • Conclusions and Future Work

Style

  • Be clear, simple, correct, interesting, direct
    • delete unneeded words, simplify sentence structure, establish logical flow
  • Be objective and accurate
    • primary objective is to inform, not entertain
  • Don’t use contractions or slang
  • Use examples when needed for clarification
  • Link text together as in a narrative
    • each section should tell a clear story

Style Specifics

  • Every sentence in a paragraph should be related to the paragraph’s topic
  • Don’t italicize words unnecessarily.
  • Don’t use capitalization for emphasis, only for abbreviation

Editing

  • Your first draft is not your final draft!
  • The goal is to make the paper clear and readable
  • There is no excuse for spelling errors!
  • Double-check noun-verb agreement
  • Double-check bibliography - make sure that the citations match your list of references
  • Make sure that you have been consistent throughout the paper
  • If you are unsure of grammar usage, look it up!

Citation Style

  • Don’t use the citation label (e.g., [16]) as a noun
  • et al. (‘and others’) is an abbreviation. It should be italicized because it’s a foreign language phrase
    • et means ‘and’ - no period
    • al. is an abbreviation for alii, meaning ‘others’
  • Provide a complete a citation as possible
    • include page numbers, dates, etc.
    • follow conference/journal guidelines
    • don’t just copy from citeseer, use citation in ACM Digital Library, IEEE Xplore, or author’s webpage

Improving Your Paper Reference: A Referee's Plea by Mark Allman

  • Use the spell checker (and grammar checker)!
  • Sloppy papers take away from the content
  • Don’t rely on color graphs
    • everything should be readable in black & white

Links

Presenting Data

Reference: The Elements of Graphing Data by William S. Cleveland

The Data is the Most Important Part

  • Don’t make the reader work to understand your graph.
    • don’t allow labels to interfere with the data
    • plotting symbols should be easy to distinguish

Graphs in Papers

  • Each figure or graph should be numbered with an informative caption
  • Write descriptive x and y axis labels that include units. Use large fonts.
  • Don’t make readers flip backwards to find your figure
  • If you use a figure from another source, give attribution in the caption

Links

Giving Presentations

Reference: Writing for Computer Science by Justin Zobel

  • Consider the audience
    • don’t bore them with background they already know
  • Think about what you want the audience to walk away knowing
  • Keep in mind your time limit
    • leave time for questions
  • Don’t provide too much detail
  • Start with motivation
  • First slide should always contain the title, your name (and names of your collaborators), and your affiliation
  • Proof-read your slides
  • Check consistency in capitalization and font usage
  • Keep slides clean and simple
  • Make transitions between topics smooth
    • don’t just read the title of each slide as a transition
  • Speak clearly and slowly
  • Face the audience
  • Practice! Practice! Practice!

Links

Professional Communication