Hints for Writing a Seminar Report, a Papers, or a Thesis

Prof. Philipp Slusallek, translated by Florian Winter

Saarland University
April 2006

Abstract grading them. The following section then pro-

vides hints how to write a report. Of course,
Studies at a university to a large degree involves these hints cannot be complete and are meant
passive, reproductive activities that often re- solely as a coarse framework for your work. You
quire little creative work by the students. Most can find further hints for writing texts in the
study programs contain only few opportunities, literature (e.g. [vW93], in German).
apart from theses and seminars, to practice the
independent writing of longer texts. However,
this particular skill is of significant importance
1.1 Format
for a later scientific or industry carreer, for ex- The optimal length of a report depends to a
ample for writing reports or conference papers. large extent on the type of a course (seminar,
It is therefore in your own interest to make best Bachelor thesis) and on the topic. Please refer
use of the few existing opportunities of training to the information about the course in this re-
this skill at a university. gard and discuss with your advisor. Remember
This document, on one hand, will give you that a short and succinct report is always better
hints for your report and, on the other hand, than a lengthy text with little content.
serves as a meta document that demonstrates In most courses, you can write your work in
what such a text may look like. You can and German. But if you prefer, you can also try to
should also use it as a template for writing your write your text in English. We will factor in
own report. Furthermore, the LaTeX source language deficits accordingly when grading.
code of this report can be used to get started I recomment to write your work in LATEX,
in writing LaTeX. since this is the best :-) and most widely used
system for scientific work. You should find the
sources of this (meta) report on the same web-
1 Introduction site where you got this one and can use it as
template and starting point. It shows most im-
Strengthening the active aspects of university
portant aspects of LATEX.
education in order to better prepare students
for their later carrer (a part of “soft skills”) is
a frequently discussed topic. There, it is not 1.2 Grading
enough to be able to acquire knowledge from Your grade will, on the one hand, depend on
books and other sources, but it is also important technical aspects: Correctness of content, the-
to summarize this acquired knowledge in written matic completeness with respect to available
form and to present it to others in a compact space and understandability. As in any paper,
way. the form of presentation also affects grading:
Seminar reports and theses offer the best Structure of the paper, structural completeness,
opportunity in the context of university edu- language, readability, clarity of the text and the
cation to practice active writing. Yet, first images. In scientific papers, of course, relevance
drafts of Bachelor or Master theses frequently of the topic, novelty of the contribution, techni-
and clearly show that these opportunities are cal solution etc. also count, however we cannot
not sufficiently used by students. and will not grade these aspects in seminars.
The following section presents the require-
ments for reports and introduces the criteria for

1
2 Report to continue reading the paper, but it must also
archieve the appropriate expectations, such that
This section contains some hints about writing the reader will not be disappointed if the content
your report. Most of these hints also apply to doesn’t fulfill the initial promises.
other scientific publications or technical reports A common recpipe for an abstract is the fol-
in a similiar way. lowing:
• 1-2 sentences about the environment of
2.1 General the report: What is it all about? What
In general, you should not assume that a report is the current state of the art?
will be read completely before at least a first
• 1-2 sentences describing the problem:
(sometimes, however, even the last) impression
What kind of problem do we want to solve,
is formed by the reader.
why is it important and relevant to the
At first, someone will typically read the ab-
reader?
stract of the report to find out what the paper
is all about, what to expect, and whether it is • 1-2 sentences about the approach or the
worth readin the text. Here, you must grab the solution strategy: How is the problem ap-
attention of the reader, you must motivate him proached? What is the solution based on?
to read on, and set up the right expectations If appropriate, a very short description of
for what is to come. Next, the reader will likely the solution can follow here.
browse through the paper and skim it by means
of the chapter headings, images, and illustra- • 1-2 sentences about results: Point out can
tions (especially in computer graphics). This be done with what is described in the text.
means that these aspects should outline the con-
tent of the paper and tell the core of the “story” 2.4 Introduction
(e.g. repeat the most important results, e.g. as
captions of good, expressive figures). The introduction is, in principle, a more de-
After this, typically, the results and conclu- tailed abstract and an introduction to the topic
sions are read, to assess, whether the paper is in general. Here is the opportunity to motivate
worth giving it a closer look. For a report, it is the topic more precisely, define the environment,
therefore highly important to get all these as- discuss the problem, previous work, and to give
pects well done for turning a potential reader an overview of the paper. Usually, the own,
into an actual reader. novel contributions are also summarized here.
Once this is has been achieved, you are in a Some of these aspect may form a separate (sub-
good position: from now on is the actual content )section if necessary and important enough.
of your report that counts and decides about
success or failure of a report. 2.5 Previous Work
The following hints for the individual parts
A very important part of a report is the refer-
of a report should point out the structure of a
ence to and discussion of previous work. Here,
report as well as provide tips concerning the con-
all relevant previous work should be mentioned
tent of a text.
(with references to literature, see below). In
particular, it is important to discuss, why these
2.2 Title approaches may not be as good, not relevant,
less efficient etc. and why the approach pre-
The title of your work should describe the topic
in a few keywords as exactly as possible. The sented by your report is in some way better or
at least different from other approaches.
author, his affiliation (i.e. university), and a
Again, this section is sometimes embedded
date of publication should be included below the
into other sections or even placed at the end of
title.
the report. But it may never be missing.

2.3 Abstract
2.6 Main Text
The abstract is the initial contact of the reader
These sections you describe the actual problem,
with the report. It must motivate the reader
solution, etc. The structure therefore depends

2
strongly on the particular topic. In general,
you should pick some core issues and focus on
describing and discussing these. A paper can
rarely treat a topic comprehensively and must
therefore be limited to the essential ideas. It is
often a good idea to first present an overview of
the technique and then discuss different aspects
in separate section. This provides the reader
with a context to interprete the text in.
If appropriate, the paper can refer to an-
other report or to literature for further details.
However, the basic idea of the paper should
be roughly comprehensible by an experienced
readre even without additional literature. If an-
other paper if required for understanding, pro-
vide at least a summary of the relevant aspects.

2.6.1 The Story

Even a scientific text should have a “story”. You
are not trying to write the most boring text pos- Figure 1: This is a detailed caption of a figure,
sible, but you want to captivate the reader by which explains that in the figure you can see the
the possibly boring topic. Build up a tension impressive result of a Monte Carlo simulation of
curve like any other good story. There are dif- global illumination. Besides, it also serves as
ferent ways to archieve this: For example you an example how to embed postscript images in
can begin with the dramatic “murder” (main LATEX.
result), then fill in the details how it came to
happen this way. A more common approach is
the “happy end” approach: present all the re-
quired details of the detective work and finally
put it all together at the end.

2.6.2 No enumerations
It is important, in this regard, that you do not
give in to the temptation to just enumerate the
things you did for solving the problem (“first I
did this, then I did that”). This is a popular
mistake of beginners.
Much more important for the reader is the First column Condition Value
motivation and justification of the decisions Name1 true 123 000.00
made and the necessary background informa- Name2 false 1.00
tion. Why has this approach been taken? What Name3 don’t know 0.25
are the fundamental principles? What is the en- Sum 123 001.25
vironment in which this development has been
made? Which alternatives have been considered Table 1: Again this is a long caption explaing
and why have they not been used? What are the what can be seen in this table, what to look for,
limitations of this method? Any (honest) paper and how to interpret the data. Help the reader
should always contain this information. in making sense of the data you provide.

2.6.3 Figures
At important positions, it is essential to use im-
ages, illustrations, and other graphical repre-
sentations to make the written statements more

3
clear (see Figure 1). These should always have In German papers, this section is often also
a detailed caption which should also be under- called “Zusammenfassung” (summary), like the
standable without reading the text (see the note abstract. For a seminar report, this is also the
above). Every figure should be referenced from right place to have a critical look at the given
the text and should appear as close to the refer- topic and provide your opinion.
ence as possible and in the order they are first
refernces (sometimes exceptions to this rule are 2.9 Bibliography
necessary).
The same is true for tables that present data Here, all used literature should be listed. The
or results as in Table 1. Also note, that in En- easy management of citations is reason enough
glish you write “section”, “figure”, and “table” to use LATEX. You can generate the bibliography
when taking about them in general, but use cap- section automatically by using the “bibtex” pro-
ital letters when you refer to a specific table, gram. Simple \cite commands in the text in-
such as Table 1. sert references to literature that can be managed
in a separate bib-file. If you are using “emacs”,
2.6.4 Math and Formulas this will also help you with entering literature
into the bib-file according to the different clas-
Of course, you can simply integrate formulas sifications (e.g. book, article, in proceedings,
into the text (a2 +b2 = c2 ) or you can also gener- etc.). Please just copy and paste the provided
ate separated display formulas with or without entries or check with other literature regarding
numbering (see Formula 1). Formulas that ap- the best use of bibtex.
pear in the text, like
Z
r
L (x, ωo ) = fr (ωi , x, ωo )Li (x, ωi ) cos θi dωi 3 Results
Ω+
The result of the preceding sections should be,
and
2 that you hopefully got at least a rough idea
A = πr , (1)
about the structure and content of a technical,
are parts of sentences and should therefore also a seminar report, or a thesis.
contain the appropriate interpunctation sym-
bols.
4 Conclusions
2.7 Results This paper contains the most important infor-
In particular in computer graphics, this section mation about written reports in the context of
is often the crucial point. Almost no paper is seminars and theses. It is also a meta docu-
accepted that can’t demonstrate the results at ment that you can use as a template for your
least by a few examples. Countless papers have own work because it contains the most impor-
already been rejected because of the absence of tant LATEX commands. Further details about
convincing results, even if the ideas by them- LATEX can be found e.g. in [Lam94]
selves might have been very nice. But even the
best ideas sometimes do not work out at the
end when you try to realize them. Papers with-
References
out implementation have become rare, which [Lam94] Leslie Lamport. LaTeX: A Document
has significally contributed to the credibility of Preparation System. Addison-Wesley,
computer graphics. Of course, expressive color 1994.
pictures are appreciated here. [vW93] Lutz von Werder. Lehrbuch des wis-
senschaftlichen Schreibens. Schibri-Verlag,
2.8 Conclusions 1993.

Here you should once again briefly describe the

content of the paper and mention the most
important results and your own contributions.
This is also the right place to discuss still unan-
swered questions, extensions, and future work.