See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/342800015

How to write a high-quality paper

Presentation · July 2020

CITATIONS READS

0 1,418

1 author:

Vinh Phu Nguyen

Monash University (Australia)
117 PUBLICATIONS 4,439 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Simulating the wear of ductile materials using particle methods View project

Artificial Neural Networks Applications in Civil Engineering View project

All content following this page was uploaded by Vinh Phu Nguyen on 14 July 2020.

The user has requested enhancement of the downloaded file.

How to write a high-quality paper

Phu Nguyen

Department of Civil Engineering

Monash University

[email protected]
http://nvinhphu.wixsite.com/mysite
I thank you …
Stephane Bordas
Prof at University of Luxembourg

Lambertus Sluys
Prof at DelD Uni. of Technology

MarHjn Stroeven
DelD Uni. of Technology

Alban de Vaucorbeil
Deakin University
2
Agenda

what is the problem?

general guidelines to wri5ng
5tle/abstract/introduc5on/conclusion
references
tables/figures
how to write a paragraph
some mistakes
wriHng work flow
LaTeX

and possibly an applause? 😂🤣🤣

3
 Have you read any papers lately?

you may think you just lack the technical

sophis5ca5on to understand them

In fact, so many papers are poorly wriKen

Judy Swan – Associate Director for Wri5ng in Science and Engineering at Princeton
University – said: ‘scienHfic wriHng is bad wriHng’

https://www.youtube.com/watch?v=1pzjxYCwb08
If you can write clear, accessible papers…

get accepted quicker as reviewers worked on the content

people will enjoy reading them

people will learn something from them

your papers are likely got high citaHons

learning to write well is an essenHal part of becoming a successful researcher

General guidelines

To inform not to impress

Aim for clarity and readability and reproducibility
ContribuHons must be clearly stated
Every unit of discourse (a sentence/sec5on/ar5cle): only a single idea/message
Avoid jargon
Minimize chances for reviewers to raise issues
Clarity is more important than grammaHcal exactness
If you can remove a word/sentence/figure, do it

G. D. Gopen and J. A. Swan. The science of scientific writing. American Scientist, 78(6):550–558, 1990
https://www.nature.com/scitable/ topicpage/scientific-papers-13815490/
Explain everything
choose a method: explain why
choose a test/example: explain why
value for parameters: sources and if you come up with them, explain why
provide all parameters needed to reproduce your paper

Topology opHmizaHon is a method that allows to find an op5mal material layout within a
prescribed design domain so as to maximize or minimize certain objec5ves and sa5sfying
one or mul5ple design constraints [1]. …

Various techniques have been developed for topology opHmizaHon, for example, the solid
isotropic material with penaliza5on (SIMP) method [2] and the evolu5onary structural
op5miza5on (ESO) method [3] and its improved version – the bi-direc5onal evolu5onary
structural op5miza5on (BESO) method [4]. We choose the BESO method for our work for
the following reasons. First, without needing calcula5ng sensi5vi5es, the BESO is much
easier to implement and can be easily adapted to a wide range of problems. Second,
engineers only care whether the design can be improved, but do not care whether the
solu5on is op5mal or has rigorous mathema5cal founda5on. We refer to the review of [5]
for a discussion on the merits of exisHng topology opHmizaHon methods.
Title: why do paper 5tles ma`er?
the 5tle is the part of a paper that is read the most
it is usually read first
papers with short 9tles got more cita0ons
using a ques9on mark in a paper’s 0tle reduces the cita0ons
using a colon tended to improve the cita0ons
for many researchers: never mind

‘The nucleo5de sequence of a 3.2 kb segment of mitochondrial maxicircle DNA from Crithidia
fasciculata containing the gene for cytochrome oxidase subunit III, the N-terminal part of the
apocytochrome b gene and a possible frameshid gene; further evidence for the use of unusual
ini5ator triplets in trypanosome mitochondria’ by P. Sloof, J. van den Burg, A. Voogd, R. Benne
Nucleic Acids Research, Volume 15, 1987

C. E. Paiva, J. P. d. S. N. Lima, and B. S. R. Paiva. Articles with short titles describing the results are
cited more often. Clinics, 67(5):509–513, 2012

https://www.timeshighereducation.com/news/long-research-titles-have-lower-impact
Title: what makes a good 5tle?
Indicate accurately the subject/scope of the study.
Avoid using abbrevia5ons
Do not include "study of," "analysis of" or similar construc5ons.

An efficient and robust staggered algorithm applied to the quasi-sta5c descripHon of

bri`le fracture by a phase-field approach

⇒ An efficient and robust staggered solver for a phase-field model of quasi-sta5c

bri`le fracture

General mesh method: A unified numerical scheme

⇒ General mesh method: A unified numerical scheme for fluid dynamics problems

Data-driven inverse modelling through neural network (deep learning) and

computa5onal heat transfer
⇒ Data-driven inverse modelling for heat transfer problems through neural network
The abstract

the most important secHon of your paper

the first sec5on that is read by journal editors
once published, the first sec5on that is examined by readers
in many cases, it is the only secHon of the manuscript that they will ever read

write the paper first, abstract is the final part

a concise summary of your paper
include: background, gaps, methodology and results
The abstract: an example
Fracture of hyperelasHc materials such as syntheHc rubber, hydrogels, texHle fabrics is an
essenHal problem in many engineering fields. The computa0onal simula0on of such a fracture
is complicated, but the use of phase field models (PFMs) is promising. Indeed, in PFMs, sharp
cracks are not treated as discon5nui5es; instead, they are approximated as thin damage bands.
Thus, PFMs can seamlessly model complex crack pa`erns like branching, merging, and
fragmenta5on. However, previous PFMs for hyperelasHc materials, which are mostly based on
a PFM with a simple quadraHc degradaHon funcHon without any user-defined parameters,
provide soluHons that are sensiHve to a length scale (that controls the width of the damage
band). The current pracHce of considering this length scale as a material parameter suffers
from two issues. First, such a calculated length scale.. Second … This paper presents a length
scale insensiHve PFM for briKle fracture of hyperelasHc materials. This model is an extension of
the model of Wu [JMPS, 103 (2017)] with a ra0onal degrada0on func0on. This func0on has
some user-defined parameters of which one is defined to be inversely propor0onal to the length
scale in such a way that the damage threshold (and thus maximum stress) is independent of the
length scale. Results of mode-I and mixed-mode fracture problems obtained with the method of
finite elements are in good agreement with previous findings and independent of the
discre5za5on resolu5on. Most importantly, they are independent of the incorporated length
scale parameter.
Introduc5on sec5on
What is the problem domain
What is the specific problem that the paper is solving
Demonstra5on the importance of that problem
What are the current approaches to solving this problem
What is wrong about them
What are the contribuHons of the paper
Planning the readers for reading the subsequent sec5ons
Introduc5on sec5on: example
Material hardness values are oden cited in manufacturing specifica5ons and are used for
quality control purposes. Hardness tests are usually convenient and economical to conduct
and can oden be classified as nondestruc5ve. However, despite the fact that various types
of hardness tes5ng have been quan5ta5vely conducted for over 150 years, a fundamental,
theoreHcal understanding of the test is, in many aspects, sHll lacking. This is because the
physical processes that occur during a hardness test are very complex although such tests
are rela5vely easy to conduct. Contemporary computaHonal mechanics techniques and
computer hardware have made it possible and pracHcal to numerically model hardness
tests. The goal of such modeling is to obtain more informa5on from the tes5ng and
thereby make the results more useful for valida5ng the material models used for impact
simula5ons. Numerically modeling Brinell and Rockwell hardness tesHng of metals is the
topic of this paper.

Literature review …
Introduc5on sec5on

a picture can engage the readers to con5nue reading your paper

The conclusion
People oden read the conclusions directly ader the abstract
Do not repeat the abstract
Some journals skip the conclusions sec5on
Avoid restaHng the problem/context
Highlight most significant things
State limita5ons/issues
The conclusion: one example
We have presented a new method for explicit solid dynamics within the framework of the material
point method. Based on previous works developed in the MPM community, par5cularly the Convected
Par5cle Domain Integrator, the total Lagrangian MPM and the finite element material point method, a
Generalized Par5cle in Cell (GPIC) was presented with the following aKributes:
• ︎ Enables seamless enforcement of Dirichlet and Neumann boundary condi5ons; ︎
• Seamless treatment of material interfaces;
• Higher efficiency as fewer par5cles (or elements) can be used;
All of these are achieved with the introduc5on of a finite element mesh in a supposedly meshfree
method. GPIC can be an efficient tool for modeling mulHple contact large deformaHon problem (Fig.
28).

Although some good results were obtained, there are many issues worthy of further invesHgaHon.
First, we have limited our discussion to… Second, only explicit 5me integra5on was considered due to its
computa5onal simplicity. This limits GPIC to fast transient dynamics problems. Third, we did not
consider … We are working on some of these issues.
References and cita5on

Cite originals not deriva5ves

Avoid ci5ng a list of two many papers e.g. ‘See [1-20] for some relevant work’. ︎
If a author-year reference format is used, all references in a single cita5on should
be ordered in chronological orders: (Day, 1998; Ashby, 2000; Plaxco, 2010).’
References manager

Use a ref manager (Bibdesk, Jabref, …)

Put files in cloud (Dropbox/Google Drive), synced across all devices
Whenever you found a good paper, add it to the ref manager, cite it in your
paper with some sentences.
Tables

Tables should be clear & focus on the data

avoid verHcal lines
avoid double horizontal lines
avoid boxing up cells and
leave enough space between rows
Figures
The reader usually starts by reading the abstract,
conclusion and figures

A high quality figure:

a legible font size (font matches the font text)
a high resoluHon (should be PDFs)
color-blindness aware (disHnguish red and green)
all axes are clearly defined
Sketches

Matlab or matplotlib (graphs)

Adobe Illustrator or Inkscape (sketches)

Graphs
Figures: how to cite them

Fig. 1: This is a cat.

A cat is given in Fig.1. As can be seen, it has two eyes.

A cat has two eyes (Fig. 1).

Same thing applies to tables

Global paragraph for long sec5ons

don’t lose your readers

Don’t start a sec5on with a fig/equa5on/table
2. Formula5ons

Fig. 1: This is a cat.

blahblahblahblahblahblahblahblahblahblahblahblahblah
blahblahblahblahblahblahblahblahblahblahblahblahblahblah
blahblahblahblahblahblahblahblahblahblahblahblahblahblah
How to write a paragraph: flow
In the phase-field modeling of fracture in bri`le and quasi- bri`le solids, it is crucial to represent
the asymmetric tensile/compressive material behavior. ExisHng phase-field models generally
adopt an intuiHve split of the free energy density without capturing the crack boundary
condiHons properly or an ad hoc hybrid formulaHon at the loss of variaHonal consistency. To
address this issue, this work presents a varia5onally consistent phase-field anisotropic damage
model.

In the phase-field modeling of fracture in bri`le and quasi- bri`le solids, it is crucial to
represent the asymmetric tensile/compressive material behavior: fracture does not occur in
domains under compression. To capture this asymmetric behavior, previous phase-field
models generally adopt either an intui5ve split of the free energy density without capturing the
crack boundary condi5ons properly ... This work presents a phase-field anisotropic model that is
able to capture the asymmetric behavior, varia9onally consistent and sa9sfy crack boundary
condi9ons.

sentences start with familiar (old) information and end with unfamiliar (new) information
How to write a paragraph: flow

Security proofs of cryptographic protocols are crucial for the security of everyday
electronic communica5on. However, these proofs tend to be complex and difficult to
get right. To make it easier to manage such proofs, Jones et al. have proposed a new
design principle, called the game-playing technique. This technique follows a code-
based approach where the security proper5es are formulated in therms of
probabilis5c programs, called games.
How to write a paragraph: original version
The material point method (MPM) is a par5cle-grid method suitable for solving large
deforma5ons of complicated geometries. A tradi5onal MPM implementa5on treats each
par5cle as a lumped mass. The generalized interpola5on material point (GIMP) method is a
generaliza5on of the MPM that accounts for finite spa5al extent occupied by each par5cle.
MPM and GIMP have been successfully used in simula5on of a range of complicated
engineering problems… In MPM, the material domain is represented by a set of material points
or par5cles and uses a mixed Eulerian–Lagrangian method in which Lagrangian par5cles carry
history-dependent state data and an Eulerian background grid is used for calcula5on of
deriva5ves and solving the momentum equa5on. In summary, the MPM has the advantages of
both Eulerian and Lagrangian formula5ons. Another interes5ng feature of tradi5onal MPM is
that this method enforces no-slip contact between bodies automa5cally without any further
computa5onal costs. However, the MPM suffers from a ‘cell crossing instability’ for problems
involving large displacements. This instability occurs whenever par5cles cross boundaries of any
cell in the computa5onal background grid and is due to the lack of smoothness of the grid basis
func5ons used as the interpola5on and mapping func5ons in the MPM. By introducing
weigh5ng func5ons and gradient …, the GIMP is capable of reducing these errors and improving
accuracy [3]. These func5ons, which can be interpreted as averages of the basis func5on and its
gradient over the par5cle domain, are used for mapping and interpola5ng data between
par5cles and grid nodes as well as calcula5ng internal and external forces.
How to write a paragraph: analyse
The material point method (MPM) is a parHcle-grid method suitable for solving large
deformaHons of complicated geometries. A tradiHonal MPM implementaHon treats each
parHcle as a lumped mass. The generalized interpolaHon material point (GIMP) method is a
generalizaHon of the MPM that accounts for finite spaHal extent occupied by each parHcle.
MPM and GIMP have been successfully used in simula5on of a range of complicated
engineering problems… In MPM, the material domain is represented by a set of material points
or par0cles and uses a mixed Eulerian–Lagrangian method in which Lagrangian par0cles carry
history-dependent state data and an Eulerian background grid is used for calcula0on of
deriva0ves and solving the momentum equa0on. In summary, the MPM has the advantages of
both Eulerian and Lagrangian formula0ons. Another interes0ng feature of tradi0onal MPM is
that this method enforces no-slip contact between bodies automa0cally without any further
computa0onal costs. However, the MPM suffers from a ‘cell crossing instability’ for problems
involving large displacements. This instability occurs whenever par0cles cross boundaries of any
cell in the computa0onal background grid and is due to the lack of smoothness of the grid basis
func0ons used as the interpola0on and mapping func0ons in the MPM. By introducing
weigh5ng func5ons and gradient …, the GIMP is capable of reducing these errors and improving
accuracy [3]. These func5ons, which can be interpreted as averages of the basis func5on and its
gradient over the par5cle domain, are used for mapping and interpola5ng data between
par5cles and grid nodes as well as calcula5ng internal and external forces.
How to write a paragraph: be`er version?
The material point method (MPM) is a parHcle-grid method suitable for solving large
deformaHons of complicated geometries. In MPM, the material domain is represented by a set
of material points or par0cles, which are treated as lumped masses, and uses a mixed …
However, the MPM suffers from a ‘cell crossing instability’ for problems involving large
displacements. This instability occurs … the lack of smoothness of the grid basis func0ons used
as the interpola0on and mapping func0ons in the MPM.

By introducing weigh5ng func5ons and gradient …, the GIMP is capable of reducing these errors
and improving accuracy [3]. These func5ons, which can be interpreted as averages of the basis
func5on and its gradient over the par5cle domain, are used for mapping and interpola5ng data
between par5cles and grid nodes as well as calcula5ng internal and external forces. The
generalized interpolaHon material point (GIMP) method is a generalizaHon of the MPM that
accounts for finite spaHal extent occupied by each parHcle. MPM and GIMP have been
successfully used in simula5on of a range of complicated engineering problems…
Dont’s and dos
Dont’s and dos
end results → results
completely eliminate → eliminate
completely filled → filled
equal halves → halves
as menHoned above: remove it or be more specific, as menHoned in Sec. 2.
very small → Hny (avoid very which is not precise)
always, never: avoid them
When to use and not to use ‘the’

The tensile membrane structures are widely used for lightweight and long-span roofs.
These structures are defined as a combina5on of tensioned fabric membrane and
suppor5ng elements such as rigid structural frame or flexible cables (Gale and Lewis,
2016). The fabric membrane uses the pre-stress of tension and its shape to resist the
external loading, while the suppor5ng elements are compressed and bent by the
tension in membrane. The design process of these structures involves three steps as
form finding, load analysis and cutng pa`ern genera5on. The first step is to find the
shape of structures in which the prescribed pre-stress are in equilibrium with a given
boundary condi5on, while the second step inves5gates the behavior of [the] fabric
membrane and suppor5ng elements under service loads. The shapes obtained by
[the] form finding process are usually in doubly curved surface, so they cannot be
fla`ened into plane without distor5ng. In addi5on, the fabric membrane itself is
manufactured in plane panels of 1-5m widths (Ishii, 1999). As a result, a specialist
design process, third step of cutng pa`ern, have to be conducted.
Passive or ac5ve voice?
A common belief: passive voice makes wri5ng more formal & objecHve

a personal tone can help to engage a reader

And the sentences are shorter and thus easier to understand.

Use both

For many years researchers have a`empted to model hardness tests with finite
element codes. Hardy et al. (1971) first modeled.… Their calcula5ons produced good
agreement with .… Lee et al. (1972) numerically modeled ….

https://www.youtube.com/watch?v=1pzjxYCwb08
 Start wri5ng your paper from the beginning

Write from the beginning;

When stuck, move to other parts, or do something else
The paper is in a good shape, when the last
experiment/simula5on is done
Don’t worry about the length, if too lengthy, save it as
a report, can be used later in a book
Submission

only send it to your supervisors when typos were carefully checked

do not submit your paper un5l you’re happy about it
ask peers to read over your work: e.g. another PhD student
effec5ve to get feedback sequenHally rather than in parallel
When should you submit your paper? The answer is when you think it is ready
Don’t be too picky as Voltaire once said, “The best is the enemy of the good”
Use of LaTeX in scien5fic publica5ons

https://www. the- scientist.com/uncategorized/dont- format- manuscripts- 44040.

it is open source
do not worry about format
high quality pdf
beauHful equaHons
automate the wriHng
If you’re a MS Word guy, why LaTeX?

why not?
In LaTeX, you are not programming, but simply a markup language
if you know another tool, you can use Word be`er
people using non-default browser (Firefox/Chrome), ended up staying at their
jobs about 15% longer than ones with Safari/Internet Explorer. They
performed be`er on the job
increases career opportuniHes

what if your supervisor is a Word guy??? …

https://www.theatlantic.com/business/archive/2015/03/people-who-use-firefox-or-
chrome-are-better-employees/387781/
Final words

mimic the wriHng of your favourite researcher

read a lots: not only content but also the wriHng
re-write your published papers unHl you’re saHsfied
write a lots (besides papers: blog, reports, …)
feel free to deviate from all these wri5ng styles!
if you do not enjoying wri5ng, consider a non-academic job!
One more thing

h`ps://www.youtube.com/watch?v=WP-FkUaOcOM
h`ps://www.youtube.com/watch?v=jLPCdDp_LE0&t=912s
h`ps://www.youtube.com/watch?v=1pzjxYCwb08

View publication stats