Master's thesis guidelines

The amount of credit points, 30 ects, corresponds in average 780 hours work. The exact amount may vary according to the student's personal efforts and topic of the thesis. Keep tight contact with the supervisor at all times during the work. The superviser provides necessary literature references to get the work started. For additional searches of litterature, the student must independently look for proper scientific journals, books etc., and use electronic retrievals. When reading the articles, other references therein, and the later work of the author should be browsed.

The main goals in writing master's thesis are to learn (1) basic practices of research work in computer science, and (2) learning how to write in the language of science. The latter includes not only the English grammar or appropriate use of literature references, but learning to write scientific text in a reader-friendly way. Even if your research results would be significant, but you fail to convince your readers (audience), the impact of your findings might be lesser than in deserves. Thus, when writing a scientific paper -- or in your case, master's thesis -- keep always the reader in your mind. Specifically, do not write your thesis for the supervisor only, but for a general computer science reader. Your thesis has to be understandable to your fellow students, even if they are not familiar with your particular topic.

In the following documents are collected guidelines for MSc thesis writing, and a Word template in English:

MSc thesis guidelines (doc)

MSc thesis template (dot)

You can use LaTex to write your thesis, and use a template to have the basic formatting. You can find more information in this link.

1. Collection of information

It is natural that the topic first seems obscure, and little information is found. After an active search, even too much materials can be found. Then it is necessary to return to the assignment in order to obtain a more precise definition of the task. Make notes of the article contents, and your own ideas while browsing the litterature. When enough litterature, notes on the articles and your own ideas is cathered - together with theoretic or practical research - it is time to schetch the paper. Even one page is enough for that. The reason is to find a basis for communication with the superviser, and for the final agreement and definition of the goals.

When the contents have found a form, the writing can begin. When the topic and materials are mastered well enough, the extent of each part of the work can be defined. Now the final text may in some cases be produced in a short period of time.

When writing the thesis, you should bear in mind the points the referees will pay attention to:

  • Mastering the literature. Does the author show adequate mastering of relevant litterature, and is the litterature cited appropriately?
  • Mastering the topic. Does the author show command on the topic and can he/she apply the information correctly? If the subject is not in command by the author, the reader cannot comprehend it either.
  • Presentation. Is the text outline, formatting, and language trimmed? A bad presentation steals the reader's attention from the subject, and makes it harder to comprehend.
  • Originality. Does the work consist of the authors own contribution? Is the presentation original and, perhaps, more clear than in the source materials? Is there pertinent critics on the materials refered? Are there illustrative examples compiled by the author? Does the work contain original observations or scientific results?

The aims of pro gradu work can vary greatly. The work may be e.g. a review on the topic, purely based on litterature. Then a logical and comprehensive presentation is essential. At the other end of the range may be a description of original research. When the final assignment is defined, the goals must be glarified.

Irrespective of the nature of the work, it must always contain a review on the background issues of the topic on the basis of relevant litterature. This proving of scholarship is essential even when the weight of the work is set on presentation of original research. On the other hand, the grading is affected by the amount of original contribution generated by the author. An important issue is also a critical evaluation of the subject.

2. Structure of the document

Writing is a creative process that is not easy in all circumstances. You can learn to write acceptable scientific prose. In scientific writing, the structure and style of the text affect readability and comprehensibility of the message. The structure of a scientific paper depends on the subject at hand, and the rules for formatting are specified by the forum of publication.

The structural elements of MSc thesis are in the order following:

  • cover page
  • abstract
  • table of contents
  • the actual text
  • list of references
  • possible appendixes

The abstract may be followed by a classification, e.g. applying the classes from the journal "Computing Reviews", and by a list of key words. 

When the title of the paper is chosen, one must consider the way how scientific papers are read. Because of the multitude of journals available, the reader must choose. The first criteria is the title; if the title is unintresting the paper will not be considered, even though the contents would be intresting. The summary is a brief description of the paper's contribution. The Introduction puts the work in its context, and what topics are dealt with. On the basis of these information, the reader desides whether is worthwile to continue.

The title should describe the paper contents in a concise and accurate manner. It should, nevertheless, not be verbose.

The objectives, methods, and results are described in a short, compact abstract. The abstract must also reveal, how the results were derived: are they new results generated by the author, or have they been collected from litterature. The abstract must be written for a reader somewhat familiar with the topic. The lenght of the abstract is less than one page and it does not usually contain any references.

The first chapter of the actual text is an introduction with a short review to the background of the work: relevant litterature, general objectives of the field, methods applied, and main results obtained. The reader is oriented in the context of the other chapters. The reader of the introduction may have less comprehence on the subject than the reader of the abstract. Basic concepts can sometimes be defined in the introduction, but also later in the text. The introduction shall not become a mere verbose table of contents. Even though the introduction gives a glance on the main results, it must do so in a general way, not by revealing them in detail. The introduction must underline the main issues of the work, and what the author consideres original or new.

The last chapter of text may consist of a summary and conclusions. The need for such chapter depends on the nature of the work. In no case it should replicate the introduction. Contrary to the introduction, the summary must presume that the reader is already famiuliar to to work, and make sure the reader has picked the main contributions. The conclusions should include critics on the work, show possible alternative approaches, and topics not dealt with here but worth assessing.

The text is to be divided in chapters, the chapters in sections, and possibly in subsections, all with numbered titles. A hierachy too deep makes the outline overly complex. The introduction usually needs no such hierarchy. In the beginning of each chapter the reader is oriented in the contents by showing and explaining the division into sections. Correspondingly, each section is preceded by a similar description. Thus, there never appears sequential titles without any text in between.

3. Style

Readability must be considered while writing. In no situation the author may presume that the reader knows something that is said only later. It is worthwile to use simple sentences, still trying to avoid monotony and repetition.

Before new issues are introduced, the reader must be prepared. Presentation of main concepts and results must follow an opportunity to 'have a breath' and check if everything was correctly understood. This may happen by inspecting some special situations, logical consequences, or showing simple exaples. The reader may also be aided by providing enough redundance: the same issue might be approached from different directions, i.e. by using both mathematical formalism and natural language.

Terminology must be used in a consice manner. A lexicon for IT terminology should be used, but with caution, since terms may have multiple meanings. When a term has been introduced, it must be used without any synonyms later in the text.

Selection of symbols must be consistent. A symbol shall not have multiple meanings, or one concept not denoted by different symbols in different context without a specific reason. This applies even to selection of indexes. If i is used as an index in one place, k should not be used in another without a reason. When multiple sources are refered, the notation must be unified.

Formulas should not listwise follow each other without words making the text readble. The words intended to make reading easier shall not be replased by the respective mathematical symbols. The symbols and formulas must be understood as parts of the sentence, taking this into account in language structures and punctuation. Theoremes and separate definitions must be formatted as independent entitites with no references to other text.

Direct citations or translations from source litterature may not be used in papers of our discipline. Even when an issue is presented according to a source, the text refered must be adapted to one's own writing. The source must be clearly stated.

Use of list notation must be avoided.

  • If that must be applied, texts in a list must have a uniform structure.
  • If e.g. one list element is in imperative, also other elements must follow the form.
  • If a list element is a complete sentence, it starts with a capital letter, and ends with a dot, otherwise it starts lowercase and ends without a dot.

3.1 SWAN


To help practicing to write reader-friendly scientific text, School of Computing provides you a free, semi-automatic SWAN (Scientific Writing AssistaNt)  tool that gives feedback on your text. While this tool is mainly targeted for PhD students and researchers, similar principles apply also to writing master's theses and, especially towards the end of your writing process, you are encouraged to use SWAN to get feedback of your text. The developers of SWAN are also constantly interested in feedback to improve the tool itself.

4. Language

Use of personal pronoun may follow e.g. the rules specified below:

Use "I" when you refer to your own, independent results. Do not address the writing process itself, however, if it has no importance to the end product.

Use "we" when the text implies both to the author and the reader. It is thus natural to write: "In the following we examine ..."

Use of passive voice should be avoided. For example in a descrition of a system's operation, use of passive voice leads to an ambiguous conception of when the text refers to the system, when to the user of the system, and when to the author of the text. Passive voice may replace "we", if a concise manner is maintained.

A fluent course of text must be paid attention to.

A sentence shall never begin with a numeral.

5. Layout

The text is formatted with line space 1.5. Each chapter (but not section) starts a new page. Use clearly different typing in titles, e.g.:

2. CHAPTER TITLE

2.1. Section title

2.1.1. Subsection title
 

A suitable margin width is 3 cm (left) and 2 cm (right), due to binding.

If the paragraphs are separated with blank lines, the text starts at left margin, otherwise a new paragraph starts with an indent (about three spaces wide). Extra space should be left around titles, lists, and separate line formulas. In the same way, theorems and independent definitions are separated by space.

New terms are emphasized by italics. If the term is not especially defined, its first occurrence is emphasized.

Mathematical formulas are written on separate lines if they are too long to be written inline, or if they need reference numbers. The reference numbers in parenthesis are right justified. The formulas must be placed neatly on the page, centered or at constant indent from the left margin.

Figures and tables are numbered and must have captions that specify their content. Figures, tables, formulas, definitions and theorems are numbered independently within their class. In short texts a sequential numbering will do, in larger presentations the numbers may be preceded by the number of the chapter. When the numbers are referenced in text, the number is not followed by a dot (Fig. 2), even when a dot were used in the caption, like:

Fig. 2. A schematic presentation ...

Table of contents occupies its own page after the Abstract. Use indent to illustrate the hierarchy of chapters, sections and subsections. List of references and appendixes are also mentioned in the Table of contents.

The page numbers of the Abstract and Table of contents are either not numbered, or numbered using small Roman numerals (i, ii, ...). Page number of Introduction is 1. Page numbering continues over the List of references. Multi-page appendixes are numbered as follows: Appendix 1(1).

6. Use of literature references

A part of scientific presentation is the use of references. The references give credit to the original presenter of an idea, but on the other hand shows where the reader may find more information. A correct use of references can best be learned by reading scientific articles, and by reasoning why a specific reference exists in the text, and in the location where it is placed. The following paragraphs list some observations on the usage of references, but by no means void the need of an inspection of litterature practices.

If the idea borrowed from litterature is no longer that one sentence, the reference is placed at the end of the idea (not necessarily the end of the sentence). If the idea is longer, the reference should be placed before the idea - still not in the beginning of the sentence. It should be placed in a natural position together with a statement that such an idea is about to be presented. Especially when all information in a paragraph has a single source, the correct place of the reference is not at the end of the paragraph, but somewhere near the beginning. If the contents of more than one paragraph refers to a same source, the reference is not to be repeated in every paragraph, but the entity should start with only one reference, and an explanation of which part of the text comes from the source.

The references are often identified by numbers or Family name(s) of the author(s) and the year of publication. In the latter case, if the refered author has several works in the list of references for the same year, those are distinguished by adding a small letter (e.g. Aho, 1985a).

There are various styles of referencing, depending on the tradition of each of the sub-disciplines of computer science. In algorithm and theoretical CS,  a reference is the text that is indicated by the number in brackets. The same brackets may contain several references, e.g. [7, 17, 19]. If a reference to a specific location in text is needed, the position or page number can be included in the brackets, e.g. [15, pp. 103-113]. Istead of brackets, slashes can be used /7, 17, 19/.

When the references are identified by the name and year of publication, the reference can be written by two ways. If the name is included in text, only the year is enclosen in parenthesis, e.g. "Aho (1985b) has developed the following...". In other cases, the name(s) are also enclosed in parenthesis, e.g. "We can now calculate the result using the algorithm (Aho, 1985b)...". If there is two authors, both names are often mentioned, but if there is more than two, only the first author is mentioned in the reference, followed by text et al., for example (Denning et al., 1989). Consult the APA manual and with your supervisor for more details on referencing.

List of references always starts on a new page. The list contains all source materials refered to in the paper. The list may contain materials that have not been used as actual sources, but have been refered to in the text. The list of references may not contain materials not refered to in the text. The list is sorted alfabetically by the author name(s), or if absent, by name of the publication.

List of references is ment to give the reader all relevant information to identify the original publication. Therefore, a great care must be taken in preparing the list of references. The notation shall be consistent throughout. The syntax of references is domain-dependent, but at least the following information must be available:

Book:
Author: Title. Publisher, place of printing, year of publication.

Journal article:
Author: Title. Journal, volume(number): start page-end page.

Conference abstract:
Author: Title. Conference abstract (Ed. Editor's name), Publisher, year of publication, start page - end page.

Manual:
Name of book. Publisher, year of publication.

Electronic source:
Author: Title of page. Internet WWW-page, URL: internet address, (date of version available).

Oral communication:
Person, nature of communication, year.

The following example shows what a list of references could look like:

REFERENCES

Armstrong Laboratory. Advanced Instructional Design Associate. Internet WWW-page, URL: http://www.brooks.af.mil/AL/HR/HRT/HRTD/aida.htm (18.10.1996).

Cronbach, L. and Snow, R. Aptitudes and Instructional Methods: A Handbook for Research on Interactions. Irvington, New York, 1977.

Denning, P.J., Comer, D.E., Gries, D., Mulder, M.C., Tucker, A., Turner, A.J. and Young, P.R.: Computing as a Discipline. Communications of the ACM, 32(1):9--23, 1989.

Ernst, G. and Newell, A. GPS: A Case Study in Generality and Problem Solving. Academic Press, New York, 1969.

Symphony Reference Manual. Lotus Development Corporation, USA, 1996.

7. In summary

  • Think about the objective of your presentation.
  • Get familiar with the subject by making notes.
  • Plan the outline of the paper.
  • Write using simple, precise prose, following the plan.
  • When you leave your work for examination, make sure it is in the final form throughout.
  • In all phases of the work you may ask advise from your superviser. Keep the superviser informed on the course of the work.

8. Once the thesis has been accepted

After the thesis has been accepted and graded, the student must take a maturity examination. This is an essay of about three hand-written pages written on each row in English on the given subject. The subjects are from the topic area of your thesis, and there are two aims for the maturity exam: Firstly, to ensure that you have produced and written the materials of your thesis by yourself, and secondly, that you are able to express yourself fluently by writing on the substance area of your thesis. The text will be examined by the supervisor of the thesis.