From Topic to Question

• This is mostly from The Craft of Research
  • Their aim is to write a research paper.
  • Our aim is to do a research project with supporting documentation.
  • In my mind there is a difference.
• But let's assume you have a Topic, or even a narrow topic.
• Topics might include
  • AI in physical training.
  • Simple computer ISAs
  • Cloud based computing sounds interesting.
  • Profiling MPI programs.
• To succeed in your research project, you need to refine this into some form of a question.
• In our class these might include
  • How can I incorporate AI to help a gym owner improve upper body strength of their clients?
  • Can I design an ISA capable of supporting the following five example programs?
  • Can I build a cloud based system which supports learning to use assembly language across several example architectures?
  • Do sample MPI programs perform like typical production applications?
• We want
  • something we can work with during the course of the semester.
  • That has at least some academic foundation.
• Be careful of being a fact collector.
  • Booth points out that just gathering a ton of information is probably a good high school project.
  • But that is not our purpose.
  • We want to turn this into something that is actionable.
  • We (might) want to answer a question.
• Collecting a set of references is a first start.
• But don't start compiling notes yet.
• Read the sources.
  • What did they do? (Methods)
  • What did they claim they have demonstrated? (Results)
  • Do the results agree with previous work? Other work? Your experiences?
  • Did they make assumptions that might be changed?
  • Did they set limits that are not clearly needed?
  • Did they produce a theoretical result that could now be implemented?
  • Can their results be tested in another way?
  • Can the results be applied in a new way?
• Booth suggests the following steps
  • I want to study topic
    • I want to study profiling in MPI
    • I want to study cloud computing
  • I want to study topic because I want to know question
    • I want to study profiling in MPI because I am interested in becoming a better parallel programmer.
    • I want to study cloud computing because I want to know how it works.
  • He adds an in order to help my readers understand step.
    • But this is more for a research paper, not a project.
• If we were in a real research project, conducting fundamental research we would need to make sure that our work led to something significant.
  • IE can we show through our lit search that no one has written about/done this?
  • Is there a open space or missing value.
  • You don't want to perform years of work and have nothing to show.
• However, you frequently can't find this open question until you become an expert.
  • This is where the reading comes in.
  • And why it is very important to have a good advisor.