Developing a product based MSc Dissertation

1
Developing a product based MSc Dissertation

• Brian Farrimond

2
Outline

• Why create a product
• What a product based dissertation requires
• Examples of products
• Scheduling product development
• Development hints

3
Why create a product

• Depends what you want out of this stage in your
  MSc
• Opportunity to develop software development
  skills
• Analysis
• Design
• Implementation
• Gives a specific focus to your dissertation
• Something to put on your CV

4
What a product based dissertation requires

• Rationale
• An explanation of why the product is worth
  developing
• Development
• A discussion of alternative possibilities and the
  decisions made in specifying, designing and
  (perhaps) implementing the software
• Evaluation
• A critical evaluation of the product
• How well it meets the rationale
• See Tonys presentation

5
Example of a product

• Pascal Trainer

6
Pascal Trainer Demo

• Demo
• It might look nice but
• is it suitable is for an MSc dissertation?
• Questions to be asked
• Does simulation improve learning and teaching?
• Has someone else produced something like this
  already?

7
Pascal Trainer Research Questions

• Can a simulator improve student learning of
  programming?
• Can this simulator improve student learning of
  programming?

8
Answering the research questionRationale

• Literature review of use of simulators in
  teaching and learning programming
• Find out what kind of approaches have been tried
  before and how successful they were.
• Identify reasons for success and failure
• Identify where improvements could be made

9
Answering the research questionDevelopment

• After deciding what your product should do now
  you decide how it is going to do it
• Identify and document alternative strategies
• HCI issues what should it look and feel like
  and why
• Development environment issues (e.g. Visual Basic
  or Java or Visual C). Which one and why.

10
Answering the research questionDevelopment

• Document issues arising during development where
  alternatives had to be assessed and decisions had
  to be made
• E.g. use of data structures
• E.g. use of algorithms
• You need to justify decisions using academically
  credible literature

11
Answering the research questionEvaluation

• Once the product is developed you can now apply
  the techniques discussed by Tony in his
  presentation to find out if it does meet the
  goals set in the rationale.
• H0 this product does not improve the situation
• H1 this product does improve the situation

12
Development issues

• You do not have much time
• You may be over ambitious and so you may end up
  with
• design only
• design plus partial implementation
• You may try to do something in an area you know
  nothing about
• You may try to use languages and software tools
  you are not familiar with

13
Development guidance

• Choose something you are already interested in
• Use tools and languages you are familiar with
  unless
• .. a prime aim of your dissertation is to learn
  a new language or API (e.g. OpenGL)

14
Development guidanceDivide and conquer

• Partition product development into components
• E.g
• the graphical interface
• data entry
• data output
• processing

15
Development guidanceDivide and conquer

• Accept that no matter how many hours you put in,
  it will never be finished
• Identify the order in which you should tackle the
  components

16
Development guidanceDivide and conquer

• Put the component which is key to your product
  (often the novel and perhaps hardest part) first
  in the list. If you cannot get that to work then
  the rest is a waste of time.
• Start in the middle and fight your way out.

17
Development guidance
Fancy input
Processing
Fancy output

• If the processing is the important part of your
  product, use temporary input and output while you
  get that processing right.

18
Development guidance
File input
Processing
File output

• For the input, use a text file that is read at
  run-time by your developing program
• For the output, have your developing program
  generate a text file containing the results of
  processing

19
Development guidance
File input
Processing
Expected file output
File output
Compare

• Create by hand the file you expect your program
  to produce
• Compare this with what your developing program
  produces

20
Development guidance
File input
Processing
Expected file output
File output
Compare

• You are developing a test suite
• Using file I/O forces you to think carefully
  about the structure of your program

21
What part of your product you need to submit

• The design
• The crucial, innovative part of your source code
  in an appendix
• CD containing your product in an installation
  (see Inno web site (http//www.jrsoftware.org/isin
  fo.php) for free installer software to help you
  build an installation).

22
Remember

• The key to your dissertation is the research
  question
• The product is merely a vehicle for a research
  question

23
Other examples of products

• Educational
• Three bit
• Time map
• Builder
• Industrial
• TrcrView

What is the research question in each case?
24
Research questions

• Three bit
• Can a simulation really enable non-technical
  student to understand the inner workings of a
  computer system?
• Time map
• Can software improve the learning and teaching of
  historical developments?
• Can software improve the understanding of events
  such as battles and campaigns?

25
Research questions

• Builder
• Can 3D modelling be done successfully without
  considerable amount of expertise.
• TrcrView
• Can visualisation improve the understanding of
  time series?

26
(No Transcript)