View
217
Download
2
Category
Tags:
Preview:
Citation preview
http://www.nottingham.ac.uk/computer-science/
Dr. Tim BrailsfordSchool of Computer Science
Conducting Research in Nottingham
http://www.nottingham.ac.uk/computer-science/
What is Research?
• Human knowledge is incomplete
– There are many unsolved problems, and unanswered questions
• Researchers ask relevant questions, and seek answers to those questions
• Research is an attempt to obtain answers by objectively studying evidence
IMPORTANT
If you can look up an answer to your problem in a book or online, then it isn’t research!
http://www.nottingham.ac.uk/computer-science/
Problems with the Definition
• The word “research” is loosely used to mean several different things:
– Finding an item of information
– Making notes for the writing of an article or talk
– Informing oneself about what one does not know
• These activities are not academic research - they are certainly not scientific research
http://www.nottingham.ac.uk/computer-science/
Types of Research
• Basic Research:
– Aims to advance knowledge for it’s own sake
• No application to existing problems in view
– Common in pure sciences (e.g. physics and chemistry)
– Rare, but not unknown, in Computer Science and IT
• Applied Research:
– Designed to help solve existing problems
– Most Computer Science and IT research is applied
http://www.nottingham.ac.uk/computer-science/5
Step 1: Find a Problem
• Find a problem in the real world (not necessarily an IT problem)
• Focus upon the question(s) to which you are trying to find an answer
• Suggest an IT solution to that problem
• Test your solution
• Where do problems come from?
– The course (subjects that interest you)
– Your reading
– Prior experience
– Outside interests (hobbies etc)
http://www.nottingham.ac.uk/computer-science/
Consider the Need for Research
• There must be some need for your research
– This may be an application, but doesn’t have to be
– The need to understand the nature of some phenomenon for intellectual curiosity could suffice
• The topic must not be trivial
• You must be able to make a convincing case that the research is worthwhile
http://www.nottingham.ac.uk/computer-science/
Define your Research Problem
• You must understand your problem thoroughly and clearly
• You must be able to communicate your problem effectively
– including to non-specialists
http://www.nottingham.ac.uk/computer-science/
Define the Context of the Research Problem
• You need to know what other people have done in the area
– It is very unlikely that no one has ever done any similar work
– Build on previous work - don’t reinvent wheels “shoulders of giants”
• Don’t rely on one resources
– Consult many before starting your work
• You may want to replicate work done previously, but look for some novel twist
http://www.nottingham.ac.uk/computer-science/9
Aims and Objectives
• These are vital - they are related, but quite distinct
• Aims are broad aspirations
• Objectives are the things that you have to do to achieve those aspirations
http://www.nottingham.ac.uk/computer-science/
Contribution to Knowledge
• Knowledge can be increased by:
– New or improved evidence
– New or improved methods
– New or improved data analysis
– New or improved concepts or theories
http://www.nottingham.ac.uk/computer-science/
Documentation of Research
• Research is never complete until you effectively communicate your ideas and discoveries
– Dissertation
– Presentations / Research Talks
– Publication
• Document your daily activities
– Keep good notes
– Never waste time looking for a lost resource
http://www.nottingham.ac.uk/computer-science/12
Types of IT Research
• Software Development– Should be innovative– Should be non-trivial– Writing some software tool to address a problem in
novel way• Investigation
– Must design and implement a programme of original research.
– Collecting and analysing data, and drawing conclusions.
– Consider the practicalities at the outset.• Review
– Must have original and innovative ideas– You should not only describe the work of others
http://www.nottingham.ac.uk/computer-science/13
The General Flow of a Research Project
Flow of research
General Specific General
http://www.nottingham.ac.uk/computer-science/14
The Dissertation
• Your dissertation should be no longer than 20,000 words - be concise - avoid waffle!
• Provide enough words to do intellectual justice to your work (the dissertation should look like 3 months work)
• Tim’s rule of thumb!
So long as you do justice to your work, the shorter the better. Most should be substantially less than 20,000 words.
• Write for the casual IT literate reader
• A distinction level dissertation should come close to being able to be written up as a publishable paper. Software development should be of a professional standard.
http://www.nottingham.ac.uk/computer-science/15
Problems with Dissertations
• Time Management
• Lack of Planning / Poor Idea
• Bad Science
• Amount of Work
• Unforeseen Circumstances
• Plagiarism
http://www.nottingham.ac.uk/computer-science/16
Time Management
• Plan all of your tasks• Background reading
• Practical work
• Writing
• Set yourself milestones• Dates
• Be flexible - everything will take longer than you expect!
• Allow plenty of time for writing• Aim to finish practical work 2-4 weeks before the
final deadline
• Do not rush the dissertation• It must stand alone
http://www.nottingham.ac.uk/computer-science/17
Planning your dissertation
• First think of a problem
– may or may not be IT related
• Suggest an IT-based solution to the problem
• Test your solution
• Evaluate your test
• Be Scientific
http://www.nottingham.ac.uk/computer-science/18
Science
• How do we know the way that things are?
– Argument
– Observation
– Argument is an inherently unreliable method
– Observation should be repeatable by a skeptic
• Opinions, and subjective views have no place in science
• Science is all about objective observations
http://www.nottingham.ac.uk/computer-science/19
Scientific Method - the process
1. Observe something (a fact)
2. Invent a tentative rationale for what you have observed (a hypothesis)
3. Use your hypothesis to make predictions
4. Test the predictions with further observations - modify your hypothesis if necessary
5. Repeat steps 3 and 4 until there are no more discrepancies - the hypothesis is then a theory
http://www.nottingham.ac.uk/computer-science/20
Scientific Method
• Science is unprejudiced
– It should always be objective not subjective
• Scientific results should be repeatable
– If someone else repeats your work they should get exactly the same results
• A scientific theory must be falsifiable
– You must - potentially - be able to prove that it isn’t true
– Science is about attempting to disprove theories - not about proving them.
http://www.nottingham.ac.uk/computer-science/21
Okham’s Razor
• How do you choose between competing theories?
• William of Okhamluralitas non est ponenda sine neccesitateEntities should not be multiplied unnecessarily
• Whenever you are faced by two possible explanations the simplest is the best to use
• This doesn’t guarantee that the simplest is right, it establishes a priority
• Keep it simple!
http://www.nottingham.ac.uk/computer-science/22
How to use Scientific Method
• Always be objective
• Formulate your ideas as falsifiable hypotheses
• When describing our work use evidence (facts) to support your arguments
• If you are designing experiments, make sure that you design them according to scientific principles
http://www.nottingham.ac.uk/computer-science/23
Amount of Work
• This must be a substantial piece of work
• 60 Credit Module
– 1 credit - notional 7.5 hours work
– roughly - 450 hours
– very roughly 37.5 hours per week
• In most cases this is a substantial underestimate
• It must look like a substantial piece of work
• The problem must be non-trivial
http://www.nottingham.ac.uk/computer-science/24
Unforeseen Circumstances
• Extenuating circumstances
• Circumstances that impede your work that were not known when you first came to Nottingham
• Usually this means illness
• Must be documented (e.g. dated medical certificates)
• Tell your supervisor ASAP
http://www.nottingham.ac.uk/computer-science/25
Plagiarism
• This is a MAJOR university offence– Make sure you know what it is!
• Plagiarism is taking credit for someone else’s work– Program code– Written words– Graphics– Experimental results
• NEVER cut & paste from the Internet (or anywhere else) without providing a full citation
• Make quotes very clear– Italicise the text, and provide the full reference at the end of
the quote - this includes graphics in figures
• Excessive quotes are bad style, and may well be penalised, but they are not plagiarism.
http://www.nottingham.ac.uk/computer-science/26
Structuring the Dissertation
• Title page
• Abstract (1 page summary)
• Acknowledgements
• Table of contents
• Introduction
– ... other chapters
• Discussion
– Conclusions
• References
• Appendices
http://www.nottingham.ac.uk/computer-science/27
Structuring the Dissertation (cont.)
• Consider the components of a traditional scientific paper
– Do not use these titles though!
• Introduction
• Materials and Methods
• Results
• Discussion
• Conclusions
• References
http://www.nottingham.ac.uk/computer-science/28
Structuring the Dissertation (cont.)• Introduction
– What are you trying to do and why?
– What have other people done in this area?
– What is original about your work
– What are your aims and objectives
• Technical chapter(s)– What tools and techniques have you used?
• Results chapters– What have you done?
– Go into detail, but stick to facts - don’t discuss them!
• Discussion– What does it all mean?
– Why did you do things the way you did, and how does this fit in with the state of the art?
– Evaluate your aims and objectives
– Conclusions - what have you achieved, and where is the work going
Recommended