A Master Thesis Project at ICTKTH

1
A Master Thesis Projectat ICT/KTH

• Some practical guidelines
• by Vlad Vlassov and Thomas Sjöland
• IMIT/ICT/KTH

2
Choosing a project to perform

• In which area (topic)?
• 2G1004 Software technology 2G1001 Computer
  Systems 2G1021 Telecommunication Systems, etc.
• Should correspond to your specialization.
• See http//www.imit.kth.se/courses/html/exjobb/Ex
  jobbpage.html
• Where?
• At a company
• The project work is usually paid
• At a department
• A project work is not paid, as its considered as
  a ordinary course
• Examiner and supervisors
• An examiner. Check a list of examiners assigned
  to topics
• See a list of IMIT examiners at
  http//www.imit.kth.se/courses/html/exjobb/examine
  rs.exjobb.html
• An academic supervisor (can be also an examiner)
• An industrial supervisor

3
Types of projects

• Development
• Expected results a prototype, results of
  evaluation (comparison)
• Research-oriented
• Expected results surveys, design choices and
  issues, design (use cases, architecture,
  protocols), a basic prototype, evaluation
  procedure, evaluation
• Evaluation
• Expected results models, evaluation/simulation
  procedure, a simulation environment or/and an
  evaluation test-bed, simulation/evaluation
  results
• In either case, a project includes literature
  study
• Relevant technologies related work (if any)

4
A typical time plan and deliverables
5
1. A project specification

• Should clearly define the amount of work and
  expected results
• Important to agree on the specification in the
  beginning
• Can be written by
• an industrial advisor (together with a student)
• an academic advisor (together with a student)
• a student
• Should include
• Background information
• Motivation for the project (whether it is worth a
  master degree)
• Problem statement. Requirements
• Expected results
• How results must be evaluated

6
2. TOC (Table Of Contents)

• To be delivered by the end of the 1st month
• TOC is a Detailed working plan
• Shows a structure of the thesis
• A short abstract for each chapter
• What is it about
• Expected results
• Should include timing
• TOC will be revised while the project progresses

7
3. Literature study

• Expected that
• you will apply a knowledge you got earlier
• you will get a new knowledge needed to perform
  the project, to make and to motivate design and
  development decisions and solutions
• You show your ability to search, select and study
  relevant literature (papers, books, tutorials,
  manuals, etc.) and related work
• A literature study report should be delivered by
  the end of the 2nd month

8
A literature study report

• Its an introductory part of your thesis
• Should include
• Background
• Motivation
• A detailed problem statement. Requirements
• Expected results
• How to evaluate
• Related work (survey and discussion)
• Existing solutions (systems, etc.)
• Survey of relevant technologies, environments,
  tools, etc.
• You should choose technologies, environments,
  etc., to be used in the project, and motivate
  your choices
• Some conclusions

9
Information sources papers

• In proceedings
• Workshops
• Usually include papers describing work in
  progress, ideas (which might be not yet properly
  validated and evaluated)
• Conferences
• Usually include papers describing rather
  completed work with strong evaluation
• Symposiums
• Usually Include papers describing some completed
  work (project) with strong evaluation
• Different scale international, local
• Sponsored by IEEE and/or ACM
• You should find major workshops, conferences,
  symposiums which are most relevant to your topic.
  Ask you advisors to help.
• In journals
• IEEE, ACM, Elsevier-published journals
  Transactions
• Journals specialized on specific topics special
  issues surveys

10
E-Libraries

• KTH Library Full text e-journals, conference
  proceedings, etc
• http//www.lib.kth.se/kthbeng/full.html
• IEEE digital library (a.k.a. IEEE Xplore)
• http//ieeexplore.ieee.org/Xplore/DynWel.jsp
• Find the link at www.ieee.org
• The ACM digital library
• http//portal.acm.org/dl.cfm
• Find a link at www.acm-.org
• You get free access, if you access from a
  computer with an IP address in the KTH domain

11
Other sources

• Books
• Specifications
• User manuals
• Tutorials
• Technical reports
• Theses
• Courses
• Much information is available on the Web
• Web pages

12
What to read?What can be skimmed or skipped?

• Should be critical to what you are reading and
  selective in what you are reading
• Who are authors? Affiliation?
• Industry (.com) can be just an advertisement.
  However, most of information is trusty when its
  related to research and development
• Which company? IBM, Intel, Sun, Microsoft,
• Academia (.edu) can be a raw idea not properly
  evaluated
• Which university? North America (MIT, Stanford,
  Berkeley, CMU, Caltech,), Europe, Asia,
  Australia, Central or South America
• Which research group? (well established, well
  known in this area, etc.)
• Which project? (scale, competed or in progress,
  etc.)
• Consortium (.org), e.g. OMG, Globus
• Where it has been reported?
• Level of a forum (conference, workshop,
  symposium)
• Level of a journal

13
Typical structure of a thesis

• Abstract
• Introduction
• Background
• Method
• Presents use cases and a system design
  (architecture, protocols, diagrams, etc.)
• Implementation
• Describes implementation
• Analysis
• Validation, Evaluation
• Conclusions and future work
• References
• Appendixes (if any)

14
How to describe

• Design and development
• May follow RUP (Rational Unified Process)
• Vision, use cases, UML diagrams, etc.
• Should describe
• a structure of the system
• how it operates
• typical usage.
• Implementation
• Describe only most essential and important
  classes, interfaces, modules, etc.
• Should give an estimate of the amount of code you
  have developed
• If required, docs, sources and user manuals (if
  any) can be placed in appendixes
• Indicate problems (if any) that you have faced
  when implementing a system prototype

15
Evaluation of results

• Validation
• Functionality tests should show that a system
  prototype works as expected
• Use cases can help
• Evaluation
• Evaluation procedure evaluation flow, input and
  output parameters
• How good is your application
• Define a notion of quality, e.g. performance,
  scalability, reliability, etc.
• What is performance in your case throughput,
  response time, or execution time, etc.?
• Requirements should help to define a quality
  measure
• How and what to measure. Ranges of input
  parameters. Sensitivity analysis.
• Evaluation environment a test-bed, benchmarks,
  test applications

16
Conclusions

• Summary
• What have you done, achieved, solved
• Conclusions
• Future work

17
Final stages

• A first thesis draft should be delivered to
  supervisors 1-1,5 month before the presentation
• May require several revisions
• A final draft should be given to an opponent 2-3
  weeks before the presentation
• The time depends on the opponent how fast he/she
  can read your report and write an opposition
  protocol
• Opponent
• Should come up with an opposition protocol to be
  sent to the examiner a few days before the
  presentation
• The examiner can make a decision whether to
  proceed to the presentation, or to postpone the
  presentation until the thesis is revised (if
  needed)

18
Presentation

• 20-45 minutes (20-30 slides)
• May have more slides (hide some slides) to answer
  questions
• Put all figures (and tables) on slides to avoid
  drawing
• Discussion with an opponent
• Questions

19
More advices

• A text editor
• Select an editor (e.g. MS Word) that provides an
  automatic update of cross-references, spelling
  and grammar checker, changes tracker, convenient
  drawing tool, comments, etc.
• Literature study
• Keep a list of references
• Take notes when reading
• A project web site
• Helps to keep a list related links and show how
  the project is progressing
• Take and keep notes of project meeting
• Diary
• Protect some sensitive data with a password