Writing a Scientific Article, part I
Vladimír Ulman, Igor Peterlík, Jan Obdržálek
FI MU
DUVOD, November, 2015
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 0 / 20
Focus of the Paper
According to the paper type
Report of some result: improvement, new finding etc.
Special-issue paper
Survey paper
Evaluation/comparison paper
Short communication or letter
Opinion paper
→ Determine the structure and content of the paper
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 1 / 20
Focus of the Paper
According to the readers
Learn in advance:
Conference typical attenders: field of study, profession
CFP topics → include appropriate keywords or buzz words
Expected level of detail: technical vs. shallow (popular)
Expected volume of evaluation/experimental data
Expected foundation/justification/strength of ideas
Expected presentation style: formulas vs. textual description
Expected presentation style: color graphics vs. plain style
→ Determine the structure and content of the paper
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 2 / 20
Typical Structure of the Paper
Title, List of authors, Abstract, Keywords
Introduction, sometimes with separater Related Work
Material and Methods
Results, often with separated Discussion
Often with Conclusion (and Future Work)
Acknowledgment, List of references
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 3 / 20
Hints for Preparing the Paper
Title, Abstract and Introduction are key sections of any paper.
Here the reader decides to continue reading or not.
One must pay extremely good attention to these section.
Allocate well enough time for writing.
Revise and rewrite, keep tuning.
Revise also after a few days again.
Request revisions from people in your field who are not familiar
with your intentions about the content of the paper.
Request revisions from people from similar fields.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 4 / 20
The List of Authors
Should contain everyone who has contributed to the result
Not virtually everyone → use also the Acknowledgment section
Include Supervisor? Include head of the lab?
Consider also funding body
Affiliations for every author (one or even more)
Corresponding author
What order?
http://www.phdcomics.com/comics/archive.php?
comicid=562
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 5 / 20
The Title
This is the first contact with your paper.
Should be: describing what the paper is about but still
sexy/attractive enough
Common dilemmas: short vs. long, amusing vs. descriptive
F. Habibzadeh and M. Yadollahie: Are Shorter Article Titles More
Attractive for Citations? Cross-sectional Study of 22 Scientific
Journals
→ Longer is better in high-impact journals
I. Sagi, E. Yechiam: Amusing titles in scientific journals and article
citation
→ Less amusing is better
Here, the reader decides whether to download or not.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 6 / 20
The Abstract
This is the second contact with your paper.
It gives a truthful description of what the paper is presenting
. . . with a reasonable amount of marketing.
It contains:
Context of the task
Task and the motivation to solve it
Main result(s)
Cool facts about the result(s)
Cool facts = why continue reading
Cool facts = e.g., why the result is important and better than SOTA
The publisher often gives a limit on length of the Abstract.
Preferably, a sentence or two should deal with every item.
Here, the reader hesitates whether to start reading or not.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 7 / 20
The Abstract
What is the optimal time for writing the Abstract?
Perhaps, prepare it as the first thing when several authors should
prepare the manuscript.
It is preferable to have it in advance not to loose the focus of the
paper.
Perhaps, rewrite the Abstract once the manuscript is finished such
that it perfectly reflects the content.
It is allowed to repeat a few sentences from the manuscript in later
sections.
→ a possible strategy for creating the Abstract
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 8 / 20
The Keywords
A short list (3–5 items) of specific but broadly known terms
roughly describing the topics studied in the paper.
It is not very standardized, often takes a free form.
Sometimes the submission system forces the author to choose
from pre-selected list of terms.
It is mainly useful for the peer-review process.
Some hints:
ACM Computing Classification system, AMS Mathematics Subject
Classification, IEEE EDICS, arXiv.org Classification
Keywords used in your favorite papers
Full-text search engines usually do not pay much attention to the
list of keywords.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 9 / 20
Hints for Preparing the Paper
Adapted from:
http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1552-4930/
homepage/SEOforAuthorsLINKS.pdf
Based on it, search engines decide where to list your paper.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 10 / 20
The Introduction
This is the third contact with your paper.
Introduces also your writing style, level of English, presentation
skills etc.
Demonstrates your knowledge of the field.
Aim at leaving a good impression.
You should:
Be straight and efficient
Have the text well organized, logical, and fluent
Briefly coin the terminology
Know the recent development of the field, the cornerstones
Identify your contribution w.r.t. the SOTA
Motivate why to deal with the subject presented
It is definitely not only just an extended Abstract.
Here, a reader decides whether to continue reading or not.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 11 / 20
The Introduction
Typical outline, topics covered:
Here is an issue in this specific context in this field of study.
→ Gives the context and raises a question to be solved.
The issue is interesting, motivate why it’s worth dealing with it.
Prove it is not completely solved yet: State of the Art (SOTA).
→ Might be very brief if a Related Work section is included in the
paper explicitly.
Show your approach, suggest some of the presented ideas.
Identify the work position w.r.t. the SOTA.
Even reveal the main achieved result.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 12 / 20
The Introduction
Reporting SOTA
Sometimes a consequent section Related work is required for this.
Introducing the structure of the paper
Often, the last paragraph is “The rest of the paper is organized as
follows. . . ”
This is not necessarily obligatory, structure is given anyway.
It should contain real information (avoiding “Results section
presents results...”)
http://www.cs.cmu.edu/~jrs/sins.html
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 13 / 20
Hints for Reporting the Related Work
Scientist have a big ego.
Never ever point explicitly at someone’s work saying it is bad or
stupid.
Instead, use a polite “sugar wrapping” way.
→ Hedging:
http://en.wikipedia.org/wiki/Hedge_(linguistics)
You never know when the person
will become a reviewer of your paper,
will suddenly become a potential attractive collaborator.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 14 / 20
The Materials and Methods
A precise and complete description of the conditions and
procedures used.
Report using present and past tenses, “we” is used even in
single-author papers.
Again, stick only to the relevant information.
Sadly, this is where you start shortening when you are over the
page limit.
If granting body does not mind, provide the implementation of
solution free of charge.
→ Will take time, should be ready when submitting.
BTW: Some journals require sample implementation and data.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 15 / 20
The Resuts
Quantitative results are preferred over qualitative ones.
Report what and how was measured.
Report what has been achieved.
Stick to the standard procedures, standard measures etc.
It may be considered suspicious to introduce a new measure.
→ Often, explaining (and comparing) why a new one is needed
requires writing another paper.
Graphics and visualization are highly appreciated.
Supplementary material may be Appended (Appendix).
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 16 / 20
The Results and Discussion
Take a step back and look at the obtained data.
Discuss your observations (as objectively as possible).
Are there any apparent general comments, trends, suggestion to
explore in the future?
Are there any unanswered questions? (This can be a problem, but
it is usually appreciated if you are honest).
Figures should be self-explanatory (via their captions).
Discussing results is a good opportunity to reference the figures.
A separated Discussion section is not very usual in
computer-science papers.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 17 / 20
The Conclusions
To some extent, it can be a rewritten Abstract with attenuated
motivation and emphasized achieved results.
Compared to the Abstract, the results can be reported more
precisely.
Unlike in the Abstract, the reader is expected to have already read
the whole text.
Readers, however, often consult Conclusions early to make sure
the paper is worth reading.
Jonathan Shewchuck: "A good conclusion says things that
become significant after the paper has been read. A good
conclusion gives perspective to sights that haven’t yet been seen
at the introduction. A conclusion is about the implications of what
the reader has learned. Of course, a conclusion is also an
excellent place for conjectures, wish lists, and open problems."
http://www.cs.cmu.edu/~jrs/sins.html
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 18 / 20
The Art of Writing Scientific Papers
Don’t get depressed.
It is:
Difficult to properly line-up logical yet comprehensive outline of the
paper
Tempting to deviate from it while writing
Difficult to choose convincing arguments given limited amount of
pages
Difficult to present them concisely/economically/efficiently yet
clearly
No surprise if your early-reviewers (supervisor, colleagues)
complain too much
Sometimes the case that additional experiments has to be
conducted
Nothing unusual to rewrite the paper considerably afterwards
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 19 / 20
The Art of Writing Scientific Papers
Don’t get depressed.
It is:
Nothing unusual if 20 reviewer–author iterations occur
Typically friends will refuse to read it over and over
Typically author becomes a reviewer in many iterations
It may take ages to prepare a good manuscript.
Even experienced authors have an average of one page per day.
Ulman, Peterlík, Obdržálek (FI MU) Writing a Scientific Article I DUVOD, Autumn 2015 20 / 20