If you like the study materials, a GitHub star is always appreciated
:)
1.1.2 License
Uni Scientific Writing Notes (c) 2021 Felicitas Pojtinger and
contributors
SPDX-License-Identifier: AGPL-3.0
1.2 Organization
Primarily based on the inverted classroom principle
Sent files should not contain metadata on person-specific info (make
pseudonymous)
Paper must be sent in by 2022-01-09
Notes must be sent in by 2022-02-27
Paper may be in German or English
1.3 Overview
What is the scientific method?
Formulating scientific questions
Designing experiments
Analyzing experiments
Planing scientific papers
Researching topics and staying up-to-date
Finding papers → Sci-Hub
Analyzing papers
Referencing papers
Writing a scientific paper
LaTeX
1.4 What is the Scientific
Method?
1.4.1 Writing Style
Structure should not follow the timeline research, but the semantic
structure of the discovery
No rhetorical questions
No judgmental formulations
Sentences should be able to stand on their own; reference people and
things by their name, not implicit references
Do not use the present tense when referring to past events, even if
it is popular in journalism
Do not use metaphors which are highly imprecise, even if they are
common among technical people
“I” should not be used in texts
Summaries should be about the effect of the research on the subject,
not the author’s view on the subject
The “motivation” at the start of the paper should not be the
personal factors, but prior pointers
1.4.2 Typical Criteria
Complexity of the theme
Amount of personal research
Quality of the content
Depth of research
Selection of sources
Implementation of prior knowledge
Structure of the paper
Visual style (used fonts, formatting etc.)
Quote style (standardized quotes)
1.5 Formulating Scientific
Questions
1.5.1 Logic and Conclusion
Argumentation
Logical conclusions
Proofs (i.e. mathematical proofs)
Experiments and their design, execution and analysis
New analysis is always based on existing knowledge
There are different levels of formalism: Argumentation, validation,
predicate-based proofs
Referencing ideas can be done in an “informal” way (whitepapers
etc.), but they must not be the base of any claims!
1.5.2 The Purpose of Writing
Communication is the primary purpose of scientific writing
But scientific writing is also a means of analysis
Formulating thesis helps to grasp the connections between
arguments
Clear formulation makes it much harder to avoid critical
questions
Gaps in analysis and open questions become obvious and lead to new
research opportunities
Writing leads to a deeper internal understanding
Even if scientific writing is limited to Uni, research methods are
always required
1.5.3 The Scientific Thought
Model
Outlook
Own research
Discussion
Proofs, research, experiments, studies
Hypothesis, underlying idea
Summary of the current state of research/technology (“related
work”)
Sources (own and external)
1.5.4 Quality Assurance
New ideas should be able to be based on existing works
Peer reviews try to check the quality of scientific works and
ensures that existing work can serve as a solid base
Own share of own work must be made obvious
1.5.5 Scientific Questions
Formulation a concrete question is required in order to reduce the
scope of topics
The question doesn’t have to be clear in the beginning of the
writing process, but must be at the end
The focus is always on the question, not the means: “Does the raft
algorithm work reliably?” for example would not include/require an
implementation of the raft algorithm, so always make the implementation
a requirement of the question!
The scientific question is not the title of the paper
Just like the goals of the research need to be clearly defined, the
“non-goals” need to be too!
1.5.6 Experiments
Gathering of data
Hypothesis
Creating the hypothesis
Designing the experiment
Executing the experiment
Testing the hypothesis with the result
Further, refined hypothesis ideas
The hypothesis is often “my idea/solution/architecture works”
Experiments support the hypothesis
Paper then describes the current technological state, experiments
and results
All dependencies and state required to reproduce the experiment must
be notes
1.5.7 Methods of Experiments
Design
Matches the scientific question
Creativity is required
Viability in time, budget and with available technology
Planning
Prevention of side effects
No convenience samples
No unethical experiments
Execution
With proper process
Proper documentation, including all unexpected incidents
Analysis
Objective analysis
No suppression of “unwanted” results
Interpretation
Objective interpretation
Usage of statistics: Is the result even statistically relevant?
Testing the feedback loop: Has the research question actually been
answered?
Description: Include all information required to reproduce the
experiment
Archiving: Storage of raw data and analysis (“data can only be
preserved if it massively replicated!”)
1.5.8 Hypothesis
Verification using proofs
Validation based on empirical data
Multiple supporting hypothesis can build a theory
1.5.9 Experiment Design
Experiments should produce a result
Testing in a specific set of parameters
Searching for optimal parameter combinations
Checking for valid sets
Sensitivity analysis
Checking the hypothesis with parameters
Checking if parameters influence results
Hypothesis tests: Statistically testing the results of
experiments
1.5.10 Analysis
Be neutral
Always ask question about results, even if they are positive
Search for additional sources
Comment on unknown factors, don’t hide them - they are means of
finding the next topic to research on!
1.5.11 Working with “Outliers”
Don’t remove or ignore them
Test if they are relevant: Do more research - are they statistically
relevant?
If they are not relevant: Classify and document
1.5.12 Comparisons
The new is not automatically better
Comparison with a baseline reference is required
Detailed description of the reference system used is required
Define the used dimensions for the comparison
Differences often occur in different dimensions
Elaborate why dimensions are being used
Fair basis: i.e. not using an under powered server
Also point out that the tool might perform worse under different
dimensions (i.e. memory constrained systems)
Comparison by
Comparison the reference solution and the new solution
Comparison of the new solution with existing literature
1.6 Planning Scientific Papers
1.6.1 Exposé
Might be required
Significant research requires planning
Assessment of feasibility
Usage of time slots
Focus on the most important goals or topics
Short description of the planned research
Which problem is the basis of the planned research?
Prior, existing research and open questions
The main scientific question: Which question is the research going
to answer?
Goal of the research
What theories is the research based on
Methods
Materials
Structure
How much time are the individual slots expected to take
1.6.2 Structure
Based on argumentation or path of discovery
Balanced
Not too much hierarchy
Minimum length of the chapters and sections
Total average ~50-60 pages
Per chapter ~3-10 pages
Typical:
Abstract (no section number, in both English and German)
Introduction (including overview)
Related work
Main investigation (multiple sections)
Results
Summary and conclusions
References
1.6.3 Basic Procedure
Clarification
Which questions should be answered?
What are the non-goals?
Creating the project plan
Getting up to date from a technical perspective
Which state is the research based on?
Search and analyzing papers
Own works
Sometimes simply structuring the comparison
Normally: Experiments!
Definition
Execution
Analysis
Selecting tools (BibTeX, LaTeX)
Sketching
Creating a structure (i.e. mind maps)
Taking note of keywords and images
Writing
Main section
Introduction
Abstract and summary
Last checks
1.6.4 Planning
Every project needs planning
Sketched planning needs to happen early in the project
Literature studies are often underestimated
Own works
Writing (min. four weeks before time is over!)
More fine-tuned research with more knowledge
Current state of research must be checked during own research
Immediate active countermeasures are required
Plan must be changed
Asses severity of changes
1.6.5 Planning the Main
Section
Structure is central
Amount of pages per section is required
Contents per section must be planned: Keywords, sources, images
Writing takes time; start writing meta before actually starting to
write
1.6.6 Planning the Paper for this
Module
Formulating the scientific question
Creating a structure
Searching and analyzing literature
Refining the structure (two layers) including page numbers
Selecting graphics (with sources)
Writing
Checking
Submitting the paper
1.7 Researching Topics and Staying
Up-to-Date
1.7.1 Sources
Web
Wiki
Google
Libraries: Books and articles
Journals and conferences: Finding journals, special issues,
searching for articles
Use catalogs
1.7.2 Research
Starting with research
Internet (Wikipedia, Library Genesis, Sci-Hub, Scholar, CiteSeerX,
arXiv, ResearchGate)
Libraries
Journals
Skimming the first articles
Doing more research on interesting literature
Finding the primary source
Finding papers which have been cited often
Finding related authors and researching their latest papers
1.7.3 Skimming Papers
Don’t start by reading the paper from start to finish
What did the authors do?
New understanding of existing systems
New solutions for the issue
Explanation of a new research question (with or without a
solution)
Reviewing existing solutions or ideas
What is the result of the paper?
Don’t check only the abstract - skim for keywords too!
Analyze included graphics
Checking the title
Checking figure descriptions
Don’t check all math unless necessary (which it mostly isn’t)
1.7.4 Reading Papers
Maintaining a critical view: Many papers over-promise and
under-deliver
Still: Skim the paper first
Extracting main expressions
Only read subjects in detail which are interesting for the research
topic
1.7.5 Critical Reading
Be aware of deceptive terminology
Don’t use “common sense”
Note implicit and explicit assumptions, approximations: Are they
warranted?
1.7.6 Documenting the Reading
Process
Excerpts
In sections or with paraphrasing
What is the topic? What is being published on it?
Creating a summary
Paraphrasing
Adding comments
Visualizations: Mind maps, concepts maps or logical formulas
1.7.7 Critiquing Papers
Scientifc Standards: Scientifc questions, methods,
literature and other sources
Maybe one could argue, that → I’m not sure what to think
There is consensus → Some people think
For obvious reasons → I have no proof
There is no doubt → I am sure
It is likely → I have no proof and don’t have the time to check
It is not necessary to take a closer look → I do not want to take a
closer look
TODO: Add section on referencing other works
1.8 Citation
1.8.1 Bibliography
Contains all read works
Used sources
Current state of research
Support for argmentations
Base for comparisons
In .bib file
Can be used for multiple papers
1.8.2 References
What
Bibliographic references
Own annotations
Excerpts
Comments
Keywords
Opinions
Relation to other references
How
Findable
Extensible
Linkable (in both directions)
Useful in bibliography
Where: List or database
1.8.3 Using References
Before reading: Taking note of bibliographic data
While reading: Excerpts, annotation and links between
references
While writing
Citing directly (including page number)
Automatic creation of references allows automatic import into word
processing
1.8.4 Purpose of Citations
Showing which ideas came from whom and which publication
Often a requirement due to copyright restrictions (attribution)
Shows that relevant literature was consulted
Creates a chain of trust based on trusted sources
Can allow checking the novelty of a work (what is new, what is
referenced?)
1.8.5 Evaluation of Source
Quality
Sources must be verifiable and trusted, so peer-reviewed
publications are the best basis
Wikipedia is a good entry point due to high quality and depth, but
citing original sources is often the better choice
Blogs and popular science publications are useful for citing
opinions and events, but should not be used to give an overview of the
current state of technology
Whitepapers should only be used for research specific to the
publisher’s technology
1.8.6 Primary and Secondary
Sources
Primary source: The first publication of an idea by
its inventor
Secondary source: Recitation or analysis of an
idea
Reading primary sources allows checkng if secondary sources have
maybe misrepresented studies or used out-of-context quotes
1.8.7 Languages
In non-English publications, using both sources in the native
language and English is acceptable
In English publications, non-English publications should only be
cited if no other sources could be found
1.8.8 Quotes
Short quotes must always be marked using "
Longer quotes should be in an own paragraph and have a different
style
Require exact source, including page number
Have to be 1:1 representation
Including punctuation and writing style
Mark exclusions and own additions with []
Should be from primary source
Quotes are not typically used in informatics papers, except for
loosening up the structure or to introduce chapters; in social sciences,
they are used more frequently, as they can be a subject to analysis
(i.e. in literature analysis)
1.8.9 Reference Style
Reference should link to an information source
Using a reference means that the statement of the work inherit the
quality properties of the reference, as it is based on it
Source reference must contain the relevant data to uniquely identify
a source
Different styles are available
Chicago style (EU method)
Harvard style
Legal style (footnotes)
1.8.10 Literature List
Contains properties for each source
Name of authors
Title of publication
Name, volume, year, edition and page number
Publisher, location, date of publication
BibTeX can generate literature lists for most styles
1.9 Writing a Scientific Paper
1.9.1 Diligence
Formally: A general diligence guideline is
mandatory in scientific writing
Practically: Spelling, syntax and layout issues
Readers get more critical and find more errors
Worse marks
1.9.2 Types of Papers
Survey papers: Overview of a subject
Scientific protocol: Documentation and interpretation of a
experiments
Research paper
Thesis (BSc, MSc, PhD)
Certificate
1.9.3 Types of Documents
Protocol
Whitepaper
Specification
Offers
Presentation
Advertisement
Functional descriptions
Manuals
Press releases
Patents
News articles
Blogposts
1.9.4 Choice of Language
German
Easier as a native speaker
Many proofreaders
English
Important for all relevant documents
More readers
1.9.5 Tips on Style
Writing is an exercise
Structure is a hard requirement
No suspense
Use simple and clear styles
Rather try to impress with content than with complex sentences
Clearness is important, because it is required in the job, makes
reading and writing easier and is polite to the reader
Use foreign words with care
Keep the audience in mind
1.9.6 How to Deal with Writer’s
Block?
Deadlocks?
Just start writing anything
Work on structure instead
Creating a mind map
Don’t trash drafts, refactor them instead
1.9.7 Title Style
Scientific question != title (title should not be a question)
Don’t be too general or to precise
Must contain the main theme
1.9.8 Writing the Abstract
Short summary of the subject’s field and the solution
Must include the result (should not build suspense)
Should not contain short abbreviations, references, formulas and
sentences like “In this paper …”
~250 words
Should be in English and German
Current and future relevance of the subject
Contexts in which the subject has been analyzed
1.9.9 Writing the Overview
Last section of the introduction (first section)
Shows the relations and dependencies between the sections
Should not just reiterate the table of contents
1.9.10 Writing the “Related Work”
Section
Overview of prior and similar work
Creates the base/foundation of knowledge
Who researched what?
Where has the result been published?
Which problems have not been solved in prior work?
In which context does the work stand to related work?
Should exist before starting to write!
1.9.11 Writing the Outlook
Section
Was has been researched?
What could be improved?
Short summary of the results
Meaning of the results
Which problems could not be solved?
Judgement of the implementation
Learned experiences
New contexts to other research topics
1.9.12 Writing the Acknowledgements
Section
Not a formal requirement, but a social requirement
Especially relevant if access to internal info or external unis has
been provided
Can be used for other sources or ideas that can’t be formally
sourced
1.9.13 Scientific Grammar and
Style
Third person
Simple past
Never reference self or other groups/people
Short sentences and words
Don’t repeat formulations but do repeat words instead of using
synonyms (server, node, VPS etc. - choose one!)
Use SI units
Use significant figures
Use consistent list style, examples, unit structure (Mbit/s instead
Mbps, Mbit/sec etc.)
The first sentence of each paragraph should be the paragraph’s
introduction
Define acronyms
Simple and reserved
Should leave no space for interpretation
1.9.14 Embedding Figures
Always numerated
Must have an alt text
Referenced in text by figure number
Text must never flow to the left or right of the
figure
Source can be in alt text (i.e. “(…) using data from [3]”)
1.9.15 Infographics
Diagrams (ER, UML etc.)
Code or pseudocode
Sequential numbering of tables and figures
Tables must have their titles on their top
Figures must have their titles below
Use consistent font sizes for descriptions
1.9.16 Common Mistakes
Spelling
Style/Syntax
It’s a project description, not a scientific paper
Separating defects
Broken references
Missing alt texts
Text in description of graphics too small
Inconsistent terminology
1.9.17 Last Checks
Spelling (i.e. LTex for LaTeX)
Check if all diagrams and graphics
Check for broken References
Empty pages
Do all graphics work in black/white?
Have all acronyms been introduced before they have been used?