How to write a paper

PC/Mikes Word Documents/Writing a paper.doc

How to Write a Paper

Mike Ashby,
Engineering Department, University of Cambridge,
Cambridge CB2 1PZ, UK
Version 5, January, 2000

This brief manual gives guidance in writing a paper about your research. Most of
the advice applies equally to your thesis or to writing a research proposal. The content of
the paper reflects the kind of work you have done: experimental, theoretical,
computational. I have used, as a model, a typical Materials project: one combining
experiment with modeling and computation to explain some aspect of material behavior.
Sections 1 to 8 give guidelines for clear writing with brief examples. The
Appendix contains longer examples of effective and ineffective writing. The manual is
prescriptive – it has to be, if it is to be short. It is designed to help those struggling with
their first paper, or those who have written several but find it difficult. Certain sections
may seem to you to be elementary; they are there because, to others, they are not.
Section 8, on Style, is open-ended, the starting point for more exciting things.

How to write a paper

1

MFA, 13/03/00

Contents

1

The Design

p 3

2

The Market – Who are your readers?

p 5

3

The Concept – Making a Concept-sheet

p 6

4

Embodiment -- The first draft

p 8

5

Detail I – Grammar

p 14

6

Detail II – Spelling

p 18

7

Detail III – Punctuation

p 19

8

Detail IV – Style

p 24

9

Further Reading

p 30

10 Appendix

p 32

11 Check-list for monitoring progress

p 36


2

MFA, 13/03/00

1 THE DESIGN
Well written papers are read, remembered, cited. Poorly written papers are not.
To write well, you need a design. Like any design activity, there are a number of
steps (Figure 1). I’ve used the language of engineering design here – it fits well.
(a) The Market Need. What is the purpose of the document? Who will read it? How
will the reader use it? The answers help you decide the length, the level of detail, the style.
(b) The Concept. Good writing starts with a plan. Writers have different ways of
developing plans. I find the concept-sheet (Section 3, below) is a good way to do it.

Identify your readers

Market need

Concept

Construct the master-plan

Embodiment

Construct the first draft.

Detail

Clarity, readability, style.

edit

The visual presentation

Product

Figure 1. The Design Process. Designing a paper is like designing
anything else: there are five essential steps.


3

MFA, 13/03/00

(c) The Embodiment. The embodiment is the first draft. Get the facts down on paper
without worrying about style; make drafts of each section; develop the calculations;
sketch the figures; assemble references.
(d) Detail. Now comes the crafting: clarity, balance, readability; in a word – style.
(e) The End-Product. Appearance is important: good layout, clear headings, welldesigned figures.
The Sections that follow expand on each of these in turn.


4

MFA, 13/03/00

2 THE MARKET –Who are your readers?
Your market is your readers. Put yourself in their shoes: what, if you were they,
would you wish to find?
The readers of your thesis are your examiners. They expect details of all relevant
parts of your research: why you did it, its background, your thinking, what you did, your
conclusions and your views on where it is going. They don’t want the irrelevant parts –
details of how standard equipment works, for instance. Find out as much as you can
about content and format from your supervisor and other students, and look at some
recent (successful) theses to get a feel for the product this market expects.
A paper is read by one or more skilled referees, and, if accepted, by a
scientifically-informed audience. This manual focuses on writing papers. The pages that
follow explain how this market should be addressed.
A research proposal usually addresses two markets. One is the funding agency:
the EPSRC, the EU, another Government Agencies, or a Charity. They will look for a
match between their priorities and yours. The other is the referees that the funding
agency will use; they are charged with judging quality, promise and relevance.
Hardest to write is a popular article, addressing an audience who is intelligent –
one should always assume that – but who may know nothing of your subject. Here style,
always important, must be fine-tuned to meet their needs. More on style in Section 8.
Make no mistake. Write poorly and you’ll bore, exasperate and ultimately lose
your readers. Write well, and they’ll respond in the way you plan.
Table 1: Markets for technical writing
What writing ?

Who are the readers ?

How will they use it ?

Thesis

Examiners

To judge and rank your work

Referees

To check originality, quality, suitability

Scientifically-literate public

To extract information

Research
proposal

The funding body and …

To judge if your aims match the priorities
of the funding body

Its referees

To judge quality and promise of the work

Popular article

Intelligent but un-informed To be introduced to a new field
public
To be entertained

Paper


5

MFA, 13/03/00

3 CONCEPT -- Making a Concept-Sheet
When you can’t write, it is because you don’t know what you want to say. The first
job is to structure your thinking.
Settle down comfortably with a cup of coffee (or better, beer) and an A3 sheet of
paper in Landscape orientation as in Figure 2. Devise a tentative title for the paper and
write it at the top. Then – in as orderly way as you can, but disorder is OK too -- jot down
what seem like sensible section headings, each in its own box. Sketch in anything that
occurs to you that belongs in a section – paragraph headings, figures, ideas. Think of
things that might be relevant to the section – a reference, a graph you might need, an idea
that requires further development. Put each in a bubble near the box to which it applies,
with an arrow showing where it fits in. This is the time to de-focus, forget the detail and
think both longitudinally and laterally. What should be in the paper? What else might be
relevant? What else might you need to do the job – a copy of X, a figure of Y, the
reference Z? Put it all down. You realise that you need an extra section - squeeze it in.
You see that the order of sections is not good – add arrows indicating the new order.
All this sounds like a child’s game, but it is not. Its value lies in the freedom of
thought it permits. Your first real act of composition (this one) is to allow your thinking
to range over the entire paper, exploring ways in which the pieces might fit together,
recording the resources you will need and capturing ideas. That way, no matter which part
you start drafting, you have an idea of the whole. Don’t yet think of style, neatness or
anything else. Just add, at the appropriate place on the sheet, your thoughts. This can be
the most satisfying step of writing a paper. Later steps can take time, be hard work,
sometimes like squeezing water out of stone. But not this – it is the moment to be creative
in whatever way your ideas may lead. You can add to the sheet at any time It becomes a
road-map of where you are going.
Figure 3 shows, unexpurgated, the concept sheet I made while thinking about this
manual. Some bits were already planned; most developed in the hour I spent making the
sheet; a few were added later, after some sections had been drafted. It is a mess, notes to
oneself, but it guides the subsequent, more tedious, part of the journey. It is possible that
this starting point may not work for you, but try it more than once before you abandon it.
It is the best way I know to break writers-block and launch the real writing of the paper.


6

MFA, 13/03/00


7
MFA, 13/03/00

Figure 2 A model for a concept-sheet


8

MFA, 13/03/00

4 EMBODIMENT – The First Draft
Now the hard work. Break the job down into stages. The usual stages in writing a
paper are set out in the boxes below. Papers are not drafted sequentially; do it in any
order you wish. Get the scientific facts and technical details down, the ideas formulated,
the graphs and figures planned. If good ways of expressing the ideas occur to you now,
use them; but do not deflect effort from the key job of assembling the pieces, in whatever
form them come. Here they are.

4.1 TITLE • Meaningful and brief, in 14 point bold

Fatigue of Metal Foams
is better than

The Mechanical Response of Cymat and Alporas Metallic Foams to Uniaxial Cyclic Loading
even though it is less specific.

4.2 ATTRIBUTION • The names of the authors, with all initials; the Institute or
organization, with full address; the date.
“A.M.Harte and C.Chen,
The Cambridge Centre for Micromechanics,
Cambridge University Engineering Department,
Cambridge CB2 1PZ, UK
January 1999.”

4.3 THE ABSTRACT • Try for one sentence each on motive, method, key results,
conclusions. Don’t exceed 3 sentences on any one.
The reader of an Abstract has been lured by the title. He or she now want to know
whether to read on. Tell them, in as few sentences as possible, what they will find. No
waffle, no spurious details. Try not to exceed 100 words. Imagine that you are paying a
10p a word. See the Appendix for an example.


9

MFA, 13/03/00

4.4 INTRODUCTION • What is the problem and why is it interesting?
• Who are the main contributors?
• What did they do?
• What novel thing will you reveal?
Outline the problem and why it was worth tackling. Review the literature, recording
briefly the main contributors and summarising the status of the field when you started the
research. Provide any specialised information that the reader might need if he is to
understand what follows. State what you will do that has not been done before (new
experimental approach? new data? new model? new interpretation?) Keep it as brief as
you can whilst still doing all this.
Start with a good first sentence – see Section 8 for examples.

4.5 METHOD • Experimental paper: equipment, materials, method
Modelling paper: assumptions, mathematical tools, method
Computational paper: inputs, computational tools, method
• Explain what is especially different about your method.
• Give sufficient detail that the reader can reproduce what you did.
• Don’t mix Method with Results or Discussion – they come next.
This should be an easy section to write: just say what you did, succinctly. Use “we”
but do so sparingly: too many “we’s” sounds like a child’s day out: “first we did this, then
we did that.”
Build up a reference list as you go. See Section 4.10 for the way to deal with
references.
It is one of the principles of science that a paper should contain sufficient detail to
allow the work to be repeated by someone else. Provide this but no more. Keep the
results for the next section.


10

MFA, 13/03/00

4.6 RESULTS • Present the output of the experiments, model or computation.
• Don’t mix Results with Discussion. It belongs – all of it – in 4.7.
This, too, should be an easy section to write. Report your results simply, without
opinion or interpretation at this stage. Define all symbols and units. Present data in a
form other people can use. Give emphasis in the text the most important aspects of the
tables, graphs or figures. Give error-bars or confidence-limits for numerical or graphical
data. Statistics should be meaningful; avoid confidence-eroding statements such as
“33.3% of the samples failed: 33.3% survived;

the third sample was unfortunately

misplaced.”
Aim for a concise, economical style.
Poor: It is clearly shown in Figure 3 that the shear loading had caused the cell-walls to
suffer ductile fracture or possibly brittle failure.
Better: Shear loading fractures cell-walls (Figure 3).

4.7 DISCUSSION • Extract principles, relationships, generalizations.
• Present analysis, model or theory.
• Show relationship between the results and analysis, model or theory.
Here you are seeking to extract principles, relationships, or generalisations from the
results. Sometimes the results speak for themselves.
The novel heat-treatment described in Section 2 gives steels which are 10% stronger and
20% tougher than those heat-treated in the normal way.
could be all you need. Most of the research we do aims at why materials behave as they
do, and this requires ideas about mechanisms, models and associated theory. The function
of the Discussion is to describe the ideas, models and theories and lead the reader through
a comparison of these with the experimental or computational data. Bring out the most
significant conclusions first; develop subsidiary conclusions after that.
Be clear and concise; a Discussion is not a license to waffle. See Appendix for examples
of waffle and what to do about it.


11

MFA, 13/03/00

4.8 CONCLUSION • Draw together the most important results and their consequences.
• List any reservations or limitations.
The reader scanning your paper will read the Abstract and the Conclusions, glance at
the Figures and move on. Do not duplicate the Abstract as the Conclusions or vice versa.
The Abstract is an overview of the entire paper. The Conclusions are a summing up of the
advances in knowledge that have emerged from it. It is acceptable to present conclusions
as a bullet-pointed list.

4.9 ACKNOWLEDGEMENTS • Thank people who have helped you with ideas,
technical assistance, materials or finance.
Keep it simple, give full names and affiliation, and don’t get sentimental. A formula
such as this works well:
I wish to thank Prof. L.M. Brown of the Cavendish Laboratory, Cambridge, for
suggesting this review, and to acknowledge my debt to the books listed below.
or:
The authors wish to thank Professor A. G. Evans of Harvard University for suggesting the
approach developed in section 4.3; Mr A. Heaver for his technical assistence throughout
the project and Mrs Jo Ladbrooke for proof-reading the manuscript. The research was
supported by the EPSRC under grant number EJA S67, by the DARPA-ONR MURI
program under contract number N00014-1-96-1028, and by a Research Fellowship from
the National Research Council of Canada.

4.10 REFERENCES

• Cite significant previous work
• Cite sources of theories, data, or anything else you have taken
from elsewhere
• References must be complete: name, initials, year, title,
journal, volume, start-page and finish-page.

References tell the reader where an idea, prior results and data have come from. It is
important that you reference all such sources. It is a conventional courtesy to reference
the originators of key ideas or theories or models, even if you modify them.
There are almost as many different formats for references as there are journals. If you
have ENDNOTE on your PC it can solve the problem. Best for drafts is the Name/year
system (also called the Harvard system):


12

MFA, 13/03/00

In text : “Lu (1998)”. If there are two names then “Lu & Chen (1998)”. If there are
more than two, then “Lu et al (1998)”.
In reference list, ordered alphabetically: “Lu, T.J and Chen, C. (1998) An Analysis of
Defects in Metal Foams, Acta Mater. 15, 222-226”.
For papers: Name, initials, year, title, journal, volume, start page-end page.
For books: Name, initials, year, title, publisher, city and country of publisher, chapter
number, start page-end page (if relevant).
All are important. Do not be tempted to make a reference list without all of these. It takes
far longer to track down the missing information later than to do it right in the first place.

4.11 FIGURES • Flow charts show methods, procedures.
• Graphs plot data.
• Schematics show how equipment works, or illustrate a mechanism or
model.
• Drawings and photographs illustrate equipment, microstructures etc.
Anyone scanning your paper will look at the figures and their captions, even if they
do not read the text. Make each figure as self-contained as possible, and give it both a title
(on the figure itself) and an informative caption (below it). Make sure that the axes are
properly labelled, that units are defined and that the figure will tolerate reduction is size
without becoming illegible. Label each curve of graphs.
Good figures are reproduced or imitated by others, often without asking -- the
sincerest of compliments.

4.12 APPENDICES. • Essential material that would interrupt the flow of the main text.
An appendix must have purpose; it is not a bottom drawer for the stuff that you
cannot bear to throw away. It is the place for tedious but essential derivations, or for data
tables or descriptions of procedures, that would disrupt the flow of ideas in the main text.
It should be well structured and stand by itself. Give it a title: “Appendix A1: The
Equation for Toughness” The journal may set it in smaller type than the main text.
-----------||---------------------------||----------------------------||--------------------------||------------


13

MFA, 13/03/00

-----------||---------------------------||----------------------------||--------------------------||------------

When you get this far you have got a long way. Put the draft on one side for at least
48 hours.

Get the graphs plotted, the figures drawn up, micrographs printed and

references assembled. Do not tinker with the text yet. It is a good idea to have a checklist like the one on the last page of this manual; it helps you see where you are.

-----------||---------------------------||----------------------------||--------------------------||------------

…………… Time has passed. The draft has matured for 48 hours or more. Now we
must address the details.


14

MFA, 13/03/00

5 DETAIL I: Grammar
Grammar tells the reader the function of words and their relationship. Mess up the
grammar and you confuse the reader. What follows is a brief summary of the simplest
essentials of grammar.

5.1 The Parts of Speech
Parts of speech are descriptors for the functions of words. There are eight.
• Nouns are the names of peoples or thing: Instron, metal, computer, foam.
Nouns can be used as adjectives. When so used, they are generally hyphenated to the
noun they qualify: table-tennis, metal-foam, computer-power.
• Pronouns stand for nouns: he, she, it, they.
• Adjectives qualify nouns: a small Instron, a red metal, a digital computer, an intricate
foam.
• Verbs signify being or action: is, seems, go, interpret, understand.
Transitive verbs have a subject and an object: The load / deforms / the material.
Intransitive verbs have no object: Flowers / bloom. The research / evolved.
“Being” verbs have a complement: The test / was / completed. The theory / seemed /
correct. (“Completed” and “correct” are complements)
Many verbs have both a transitive and an intransitive form: Time / passed. And:
Pass the biscuits.
• Adverbs qualify verbs: today we interpret this differently.
• Conjunctions link words and sentences: and, but, because...
• Prepositions precede nouns, usually having to do with place or time: on the table, after
this procedure, on the graph, from the appendix.
• Interjections are exclamations; the polite ones include: Alas! Great! Cheers! Many
are impolite. They are inappropriate in technical writing.


15

MFA, 13/03/00

5.2 Sentence Structure
A simple sentence has a subject and a predicate.
Subject

Predicate

The sample

failed.

The measurements

fell into two classes.

Fatigue-loading

causes microstructural damage.

The subject identifies what or whom the sentence is about.
The predicate, containing a verb, says something about the subject.

5.3 Phrases and clauses
Phrases and clauses are groups of words that do the jobs of the parts of speech listed on
Section 5.1.
A phrase is a group of words that does not contain a verb.

Type of phrase

Example

Noun phrase

The interpretation of the experiment presents a problem.

Adjective phrase

The red and white striped cable is live.

Adverbial phrase

We examined the results with considerable care.

Conjunctive
phrase

The test ended owing to the fact that the specimen failed.

Avoid the last of these; there is always a simpler, one-word conjunction (here:
“because”).
A clause contains a verb and its subject or object. Sentences are made by linking
clauses. A sentence made with two equal clauses (each a separate sentence but linked
together) is called a compound sentence. A sentence made with a main clause linked to
one or more subordinate clauses, which cannot stand by themselves as separate sentences,
is called a complex sentence.


16

MFA, 13/03/00

Adjective clauses do the work of adjectives; adverb clauses do the work of adverbs.
Type of Clause

Example

Adjective clause

A computation that uses FE methods is appropriate.

Adverb clause

The modem will operate wherever a phone-line is available.

5.3 Compound sentences
A compound sentence has two co-ordinate (“equal”) clauses linked by a conjunction:
We measured the temperature and (we) adjusted the thermostat.
The tooling cost is high but the material cost is low.
The parts of a compound sentence must be of comparable weight. “We analysed the
microstructures using SEM and left for lunch at midday” is unbalanced.

5.4 Complex sentences
A complex sentence has a main clause and a subordinate clause:
What these results signify / is the subject of a paper by Wegst (1998).
Maine (1998) demonstrates / that technical cost modelling is feasible.
It is possible / that the conclusions were mistaken.

5.5 “that” and “which”
“The computations that were performed on a Cray were the more accurate.”
“The computations, which were performed on a Cray, were the more accurate.”
These two sentences appear at first sight to say the same thing, but they do not. The
italicised part of the first sentence is an adjective clause, qualifying the word
“computations”; it has the effect of limiting the computations the sentence is talking
about to the ones done on the a Cray, as distinct (say) from those done on a Silicon


17

MFA, 13/03/00

Graphics work station. Adjective clauses are just like adjectives; they are not separated
from the noun they qualify by commas.
The italicised part of the second sentence, separated by commas from the rest, adds a
new factor of equal importance to that contained in the main sentence. The two
statements are: the computations were performed on a Cray; and they were more accurate.
The emphases of the two sentences differ. The italicised clause in the first sentence is
subordinate, merely qualifying the noun. The italicised clause in the second sentence is
co-ordinate, meaning that it introduces a new fact.


18

MFA, 13/03/00

6 DETAIL II: Spelling
Use the spell-checker on your computer, but remember that it will fail to distinguish
“their” from “there”, “form” from “from”, “its” from “it’s”, and many more. Watch out
particularly for “effect” and “affect”, “principle” and “principal”, “dependent” and
“dependant”, “compliment” and “complement”.
Most word ending in “-ise” can also be spelt “-ize”, but not all. If, like me, spelling is
one of your lesser talents, use “-ise”.
And when in doubt, use a dictionary.


19

MFA, 13/03/00

7 DETAIL III: Punctuation
Punctuation orders prose and sends signals to the reader about how to interpret it.
Good sentence structure and punctuation makes reading flow; it warns of what is to come;
it helps the reader read without having to re-read. Meaning is changed, sometimes
dramatically, by punctuation. It is one of the toolboxes of good writing.

7.1 The full stop, or period .
The full stop is used to mark the end of a declarative sentence, and to signify abbreviation:
Dr. A. M. K. Esawi, Ph.D.

7.2 The comma ,
The comma keeps apart two words or larger parts of a sentence which would confuse if
they touched. Forget any rules you have heard about the comma and simply used it when
it improves the sense of the sentence. Try the sentence with and without the comma;
keep it if, without it, the sentence becomes ambiguous. Thus:
The measurements employed a photo-diode and a laser was used to check adjustment
requires a comma after photo-diode to avoid a momentary misinterpretation, slowing the
reader down.

7.3 The semi-colon ;
The semi-colon is used to separate when the comma is not enough and the full stop is a
more complete break than the sense demands. Most commonly, it is used between closely
related independent clauses:
At one time the optical microscope was the principal tool of metallography; today, it is
the scanning electron microscope.
When conjunctive adverbs accordingly, also, hence, likewise, similarly… link clauses,
they are proceeded by a semi-colon. It is used, too, to separate members of a list when the
comma is not enough:


20

MFA, 13/03/00

The literature includes Gibson (1997), who studied simple compression; Olurin (1998),
who studied the effect of holes and notches; Deshpande (1999), who….

7.4 The colon :
The colon introduces part of a sentence that exemplifies, restates or explains the preceding
parts. It is expectant: it sets the reader up to anticipate elaboration.
This raises the question: is the model right or wrong?
There are two reasons for repeating this experiment: the first, to improve the precision;
the second, to establish reproducibility.

7.5 The exclamation mark !
The exclamation mark signals surprise, excitement, imperative, even contradiction; it
turns up the volume.
Harte reports that metal foams sink in water.
is a simple statement;
Harte reports that metal foams sink in water!
implies that this is startling, perhaps even mistaken. Scientific writing does not need this
sort of emphasis or innuendo. Delete it, and say what you want to say in a direct way.

7.6 The question mark ?
The question mark is used after a direct question:
Why was this work undertaken? The reason…….
It is used to indicate uncertainty: Euclid, 450? –374 BC.
It is optional after a rhetorical question:
Who would trust that model.
So what.


21

MFA, 13/03/00

7.7 The hyphen The hyphen connects part of a compound word:
Well-known; half-expected; curiosity-provoking; a ball-and-stick model.
It is generally required when a noun is used as an adjective: a box-girder; a bar-chart.
Its most engaging property is its capacity to create new words and meanings by
combinations both established and original:
A Fleck-inspired interpretation; a shark-skin-textured surface.
But treat this with caution; it can easily descend into stomach-lurching purple-prosed
absurdity.

7.8 The dash —
The dash sets off parenthetic material that results in a break in continuity in a sentence.
Magnetic materials – carbon steels for instance – contain atoms with unpaired electron
spins.”
This conclusion – and it is a significant one – appears to violate the first law of
thermodynamics.
The remaining specimens – those which had not fractured – were sent for analysis.
A dash can lead to an upshot, a final summary word or statement, and give emphasis:
Cell-wall bending, cell-wall buckling and cell-wall fracture – are all equally probable.

7.9 Quotation marks “...”
Quotation marks enclose direct “word-for-word” quotations and dialogue.
“Uncork the flagon; let the wine-cups flow.” – Horace, Odes, 27BC.
“One small step for a man; one giant leap for mankind.” – Neil Armstrong, US astronaut
(1969).
Quotation marks are sometimes used to enclose an original, ironic or unusual turn-ofphrase:


22

MFA, 13/03/00

This research took a “try-it-and-see approach.”
This colloquial phraseology is uncomfortable in scientific writing; avoid it.

7.10 The apostrophe ‘
The apostrophe shows either possession or contraction; thus, the possessive forms:
Sutcliffe’s theory; everyone’s idea.
There is no apostrophe in the possessive his, her or particularly, its.
In contractions, the apostrophe indicates missing letters: Don’t, isn’t, it’s (meaning “it
is”). Contractions of this sort are inappropriate in scientific writing, but can be acceptable
in informal or popular writing, as here.

7.11 Italics italics
Underline, embolden or italicise? All three attach emphasis and importance to a word or
phrase. In contemporary scientific writing italics are preferred. Bold tends to be reserved
for headings. Underlining can appear over-emphatic and, within a text, bold can seem
authoritarian. Italics allow smooth definitions of terms:
“The critical value of the fatigue limit, or fatigue threshold, is listed…”
allows the italicised words to be used thereafter in place of the longer definition. Book
titles are often italicised “The Theory of Shell Structures” by C.R.Calladine, as are words
in foreign languages. To write more on this would be de trop.

7.12 Parentheses (...)
Parentheses – literally: putting-asides – embrace material of all sorts, and help structure
scientific writing. But do not let them take over, clouding the meaning of the sentence
with too many asides.
Face-centred cubic metals (copper, silver, and gold, for instance) have nine independent
elastic constants.
Shercliff (1998) surveys the status of modeling in Material Sciences .


23

MFA, 13/03/00

It is plausible (although not everyone agrees) that this theory is correct.

7.13 Brackets […]
Brackets are used to indicate editorial comments or words inserted as explanation:
[continued on p. 62], [see footnote].


24

MFA, 13/03/00

8 DETAIL IV: Style
A good style lifts writing from that which is dull and ordinary to that which is
distinguished, memorable, individual. There is no formula for instant style – it is partly a
personal thing – but there are useful guidelines. Style is approached through plainness,
simplicity, good structure and desire to convey information to the reader in the most
accessible way.

8.1 Be Clear
The essence of technical writing is communication. The first quality, with precedence
over all others, is clarity. Use simple language and simple, concise construction; short
words rather than long; familiar words, not obscure. When you’ve said something, make
sure that you’ve really said it. The writers of these headlines (all from newspapers in
1998) hadn’t:
Red tape holds up new bridge.
Something went wrong in jet crash, expert says.
Chef throws heart in to help feed the hungry.
Prostitutes appeal to Pope.
Panda mating fails; vet takes over.
These are funny because the intended meaning can be guessed. More often, it can’t; then
the loss of clarity misleads and confuses.
AND DON’T WAFFLE. Consider this, from a well-known Materials text:
“The selection of the proper material is a key step in the design process because it is the
crucial decision that links computer calculations and the lines on an engineering drawing
with a real or working design”.
What does it say? “Materials selection is important”, and we knew that already. It is
wasting the reader’s time.


25

MFA, 13/03/00

8.2 Write from an appropriate design
Poor writing lacks order, mixes ideas that should develop separately, fails to progress
in a logical sequence. The concept-sheet gives structure: there is a place on it for each
part of your story. In making it, decide where the bits will go, the logical order, the way
they will fit together.
Remember who you are writing for. Tell them what they want to know, not what they
know already or do not want to know.

8.3 Define everything
Define all symbols and abbreviations.
The mass m scales as E/ ρ where E is Youngs’s modulus and ρ is the density…
leaving a double space on either side of a symbol when it appears in the text.
The measurements, made with a scanning electron microscope (SEM), …
allows you to use the abbreviation SEM thereafter.

8.4 Avoid empty words
Avoid clichés (standard formalised phrases): they are corpses devoid of the vitality
which makes meaning spring from the page:
The long and the short of it is that digital methods are the flavour of the month; the
bottom line is that analog computation is old hat – avoid it like the plague.
Avoid weak qualifiers: very, rather, somewhat, quite…
This very important point … makes less impact than: This important point … or, more
simply: This point ….
The agreement with theory is quite good suggests that it is not.
These ideas could rather easily be extended to the non-linear case …. makes the reader
wonder why you didn’t do it.


26

MFA, 13/03/00

8.5 Revise and rewrite
Revising is part of writing. Nobody gets it right first time; some go through 8 or 10
drafts. The most spontaneous-seeming prose is, often, the most rewritten. Do not be
afraid to write the first draft with the simple aim of getting all the facts down on paper.
You can then see what you’ve got, and can extract, revise, and distil a paper, a conference
report or a research proposal from it, as needed.

8.6 Do not overstate, over-emphasise or apologise
All of these undermine the reader’s confidence in your judgment.
This paper questions the basic assumptions of fracture mechanics …(this from a real
manuscript) fills the reader with mistrust; after all, fracture mechanics works.
This very important result…., This significant finding…. are better replaced by the
simpler This result…. and This finding….
Leave the reader to decide on importance and significance.
Never apologise.
Unfortunately, there was insufficient time to complete the last set of tests
suggests bad planning, incompetence.

8.7 Avoid being patronising, condescending or eccentric
Write in a way that draws attention to the sense and substance or the writing, not to
the mood or whimsical humour of the author. If the writing is solid and good, the
character of the author will emerge. To achieve style, start by trying for none.
Don’t patronise: The amazingly perceptive comment by Fleck …..
Don’t be condescending: Readers familiar with my work will know …..
Do not affect a breezy manner, what you might call Web-speak. Hi! me again with some
hot news about engineering at CUED, or Q’Ed as we call it. It’s been a helluva term for


27

MFA, 13/03/00

good stuff – we got more going on here than ever before… The author says nothing and
is showing off, drawing attention to himself.

8.8 Use appropriate language
Use standard symbols and terms. Calling Young’s modulus G will confuse, even
after you’ve defined it.
Minimise the use of acronyms and abbreviations.
The MEM, analysed by FE methods, was photographed by SEM and chemically
characterised by SAM.
is bad writing. Find other ways of saying it, even if it takes more words.
Avoid jargon. Jargon is the secret language of the field. It excludes the intelligent,
otherwise well-informed, reader, and speaks only to the initiated. Some jargon is
unavoidable – new concepts sometimes need new words. But don’t be tempted to use it to
show that you are an insider. The Appendix has examples.
And above all, remember who you are writing for.

8.9 Good first sentence.
Don’t start introductions (or anything else) with platitudes. Tell the reader something
he does not already know. Openings such as:
It is widely accepted that X (your topic) is important …
has the reader yawning before you’ve started. Try to get a new fact, new idea or a
revealing comparison into the first line.
Poor Opening: Metal foams are a new class of material attracting interest world-wide
and with great potential….. X, Y, Z have measured their strength properties … P, Q, and
R have developed theoretical models … Comparison of the experiments with the models
suggests that the measured strength are less than those predicted …..
The first sentences is a platitude; the second and third involve the reader in details, the
relevance of which is not yet clear; only in the fourth does the point start to emerge.


28

MFA, 13/03/00

Better: Metal foams are not as strong as they should be. Models, which describe
polymer foams well, overestimate the strength of metal foams by factor of 2 to 5. This
research explores the reasons.
To be more specific… (details of literature X, Y, Z, P, Q, R here).
The first two sentences now highlight the problem. The third says what the paper is going
to do. The details that follow then have relevance.
Use a quotation only if it is spot-on; inappropriate quotations give the impression that
the writer is trying too hard.
“God created solids, but the Devil created surfaces” -- anon.
could be a good first line for a review-article on friction and wear, but it is pretentious as
an opening to a paper on (say) the wear of bronze journal bearings. If you do use a
quotation, make sure you get it right – see Quotations, in Further Reading.

8.10 Seek helpful examples and analogies
Ferro-magnetic material – steels, for example – can be shock-loaded by pulsed
magnetic fields.
The example of steels makes the generalisation concrete.
One cause of rolling friction is material damping. A rolling ball deforms the surface
on which it rolls. If the work done in this deformation is lost through damping, a
frictional force opposes motion. It is like riding a bicycle through sand: the rubbing sand
particles dissipate energy much as atom or molecular rearrangements do.
The bicycle analogy is appropriate; it relates the scientific problem to one which is
familiar.
There are more examples of analogies in the Appendix.


29

MFA, 13/03/00

8.11 Linking.
Each sentence in a paragraph should lead logically to the next. When you read a
paragraph, where does it jar? Why did you have to pause or re-read? What word-change
will fix it? Edit for readability.
It helps the reader if one paragraph ends with a device that links it to the next: a word
or phrase picked up in the first sentence of the following paragraph, or a statement of what
is coming next (though be sparing with this, it can get tedious).
………..To progress further, we need a way to rank the materials – a material index.
A material index is a …….
The repeated words link the two paragraphs.
………..This behaviour suggests that the process is diffusion-controlled. A model based
on this idea is developed next.
The stresses at grain boundaries can be relaxed by diffusion. …..
The reader knows what the second paragraph is about before reading it.

8.12 Observe good writing
When you read a good opening, an apt analogy, an illuminating example, or an idea
well expressed, re-read it. Don’t try to imitate it directly, but observe how the author did
it. Bit by bit you can absorb the techniques.

8.13

Finally…
Style takes its final shape from an attitude of mind, not from principles of composition. Focus on

clarity. Make sure you’ve said what you think you’ve said. And remember who your readers are; seek to
express your results and ideas in ways they will most easily grasp.


30

MFA, 13/03/00

Acknowledgements
I wish to thank Prof. Yves Brechet of ENSEEG, the University of Grenoble, and Dr. Ann
Silver of the Physiology Department, Downing Street, Cambridge, and Prof. John Carroll
of the Engineering Department, Trumpington Street, Cambridge, for their advice, and to
acknowledge my debt to the books listed below.

FURTHER READING
Texts on how to write technical prose:
“The Complete Plain Words” 3rd edition, by E. Gower, revised by Greenboum, S. and
Whitcut, J. Penguin Books, London, UK (1986)
“A Writers Guide for Engineers and Scientists” By R. R. Rathbone & J.B. Stone,
Prendice-Hall Inc, Englewood Cliffs, NJ, USA (1962)
“The Reader Over Your Shoulder” By R. Graves & A. Hodge, Collier Books, New York,
USA (1943).
“The Elements Of Style” By W.Strunck Jr., & E.B White, Macmillan Co, New York,
USA (1959).
“Communication in Science”, 2nd edition, by Vernon Booth, Cambridge University Press,
Cambridge UK (1993).
“The Little Brown Handbook”, 6th edition, by H.R.Fowler and J.E. Aaron, Harper Collins,
New York, (1995)
Instructions on preparing scientific papers:
“General notes on the Preparation of Scientific Papers”, 3rd edition, (1974), The Royal
Society, London.
Grammar
“Clear English’ by Vivian Summers”, Penguin Books, London, UK (1991)
“Chambers English Grammar” by A. J. Taylor, W & R Chambers Ltd (1990)
Punctuation
“The Well-Tempered Sentence” By K.E Gordon, Horton Mifflin Co Boston, USA (1993)


31

MFA, 13/03/00

Spelling – the friendliest dictionary is
The Chambers Dictionary, Chambers Harrop Publishers, London U.K. (1998)
…….but the ultimate authorities remain
The Concise Oxford Dictionary, 8th edition, Clarendon Press, Oxford UK (1990), or
The Shorter Oxford English Dictionary, 4th edition, Clarendon Press, Oxford UK (1990)
Quotations
The Oxford Dictionary of Quotations, 4th edition, Oxford University Press, Oxford U.K.
(1996)
Synonyms and Antonyms (words that say the same or the opposite)
The Penguin Dictionary English Synonyms and Antonyms, Penguin Books, London, UK
(1992).
If words fascinate you, the following are delightful:
“Troublesome Words” 2nd edition, by Bill Bryson, Penguin Books, London, UK (1987).
“Panati’s Extraordinary Origins of Everyday Things” by C. Panati, Harper and Row, New
York, USA (1987).
“Panati’s Extraordinary Endings of Practically Everything and Everybody” by C. Panati,
Harper and Row, New York, USA (1989)
“Dictionary of Word Origins” By J.T. Shipley, Littlefield, Adams & Co, NJ, USA (1977).
“Word Histories and Mysteries” Edited By K.Ellis, Horton Mifflin Co Boston, USA
(1974).
“The Penguin Dictionary of Curious and Interesting Words’” By G.S. Sausy III, Penguin
Books, UK And Viking Books USA (1984).


32

MFA, 13/03/00

Appendix: Examples of effective and ineffective writing
Good Abstract
Temography of Shear Bands in Metal Foams
Metal foams, when compressed, deform by shear banding; the bands broaden as
deformation progresses. We have studied the nucleation and broadening of shear bands
by laser-speckle strain-mapping. The foams were non-homogeneous, with spatial
variations of density of a factor of 2; the shear bands nucleate in the low-density zones,
and broaden into the high-density regions as strain progresses. The results indicate that
processing to minimise the density fluctuations could increase the initial compressive
yield strength of the foams, when shear bands first form, by a factor of 1.5.
This, in four sentences and 94 words, gives a clear, concise portrait of the paper,
devoid of unnecessary detail and secondary information.
Good opening sentence
From a review article on the elastic properties of materials
“Ut tensio, sic vis”. As it is stretched, so it resists. With these words Robert
Hooke enunciated in 1674 the law that bears his name.
Enzio Manzini, in “The Materials of Invention”, Design Council, London 1989.
The quotation nicely suggest the history and introduces the subject.
Good analogy (1)
Structured Programming
‘Music, poetry, and programming, all three as symbolic constructs of the human
brain, are found to be naturally structured into hierarchies which have many different
nested levels. Sounds form meaningful units, which in turn form words; words group into
phrases, which group into sentences; sentences make paragraphs and these are organised
into higher levels of meaning. Notes form musical phrases, which form themes,
counterpoints, harmonies etc; which form movements, which form concertos symphonies
and so on.
Structure in programs is equally hierarchical, if not so universally recognised….’
‘Numerical Recipes’ by Press, H W. Flannery, B.P. Teukolsky, S.A. and Vetterling W.T.
Cambridge University Press, Cambridge, UK (1986).


33

MFA, 13/03/00

The analogy is a little long-winded, but it achieves the writers’ aims: to convey
the importance of structure in programming, and , by association, to portray programming
as an art-form and to elevate its stature as an intellectual activity.
Good analogy (2)
The Character of a Volvo
‘Volvos have a certain character. Purchasers see them as solid, safe, long-lasting,
reliably masculine, with built-in Scandinavian qualities of good design – it’s what we call
the “Product DNA” …..’
Ford Company spokesman explaining that Ford, who have just bought Volvo, will retain
and develop the Volvo character.
The DNA analogy captures in a word the subtle combination of real and perceived
values which lie at the heart of customer loyalty.
Avoid Waffle
The Role of the Materials Engineer in Design
‘The role of the Materials Engineer in the design and manufacture of today’s
highly sophisticated products is varied, complex, exciting and always changing. Because
it is not always the metallurgical or materials engineer who specifies the materials, this
ASM Handbook on Materials Selection and Design is prepared to benefit all engineers
who are involved with selecting materials with their related processes that lead to a ready
to assemble manufactured component.’
Extract from the Introduction to ASM Metals Handbook vol. 20, ASM International (1998)
Metals Park, Ohio, USA
There is a warning here for us all. What you want to say is:
Engineers need to choose materials, and to find processes to shape and join them. This
volume is designed to help them.
But that sounds too short, too plain, not grand enough. The fear of sounding trivial, of not
being sufficiently heavyweight, haunts all writers when they are asked to write for
audiences with whom they are unfamiliar. The temptation is to use long words, to sound
sophisticated, to get fancy; and the effect is to dilute the message until its true flavour is
lost. I have been just as guilty of it as anyone else. Don’t do it. Say what you mean to
say and say it clearly and simply.


34

MFA, 13/03/00

Remember who your readers are
The Act of Design
‘Designing is a creative activity that calls for a sound grounding in mathematics,
physics, chemistry, mechanics, thermodynamics, hydrodynamics, electro-engineering,
production engineering, materials technology and design theory, together with practical
knowledge and experience in specialist fields. Initiative, resolution, economic insight,
tenacity, optimism, sociability and teamwork are qualities that will stand designers in
good stead and are indispensable to those in responsible positions.’
From a distinguished book on Engineering Design
How many people do you know who could meet that job description? The authors
wish to convey the idea that design is an inter-disciplinary activity, and one that has
technical, managerial, and social facets, but they have done so in a way that intimidates.
They have lost touch with their readers. An alternative with the same message might be:
Designers cannot be expected to know everything, yet there are times when it might seem
that they must. Design involves an exceptionally broad base of technical competence and
practical experience, leadership, teamwork and management skills.
Try not to alienate your readers. Phrase your message with them in mind.

Jargon (1)
A Definition of love
‘… the cognitive-affective state characterised by intrusive and obsessive
fantasising concerning reciprocity of amorant feelings by the object of the amorance.’
At a US Conference of Sociologists (1977), cited by Bryson (1987) – see Further Reading
This sort of stuff is rife in critiques of music and art, and in writing on Psychiatry,
Psychology and Sociology. It surfaces, too, in books on Industrial Design, and, less
frequently, in scientific and technical writing. Don’t let the jargon-bug infect your own
style.


35

MFA, 13/03/00

Jargon (2)
The justification for a travel grant
My mathematical work is in the area of Symplectic Geometry and Differential Equations,
in particular on a geometrical interpretation of the Painlevé equations. I have succeeded in
attacking the Isomonodromical Deformation problem for higher order singularities by
symplectic means. On the one hand, this involves a symplectic structure obtained from
infinite-dimensional considerations and on the other an analysis of the geometry of the
Stokes matrices in the language of Poisson Lie groups.
From a students application for a travel grant, (1999)

There’s nothing wrong with the grammar, punctuation or spelling here – all are
fine. But how much does the statement convey to the panel awarding the travel grant, all
of them scientists or engineers, but none specialists in this sort of mathematics?
Practically nothing. The meaning is hidden in the jargon; the writer has made no attempt
to translate his ideas into a language the rest of the world can understand. It is not always
easy to do so – the subject of Symplectic Geometry may be a difficult one to illustrate
with simple examples or analogies – but it is always worth trying.


36

MFA, 13/03/00

Checklist for monitoring progress

Make concept sheet

1st draft

edited draft

Title and attribution
Abstract
Introduction
Method
Discussion
Conclusions
Acknowledgements
References
Figures and captions
Appendices

Visual presentation


37

MFA, 13/03/00

Figure 2 A model for a concept-sheet

Figure 3 The concept sheet I made when writing this text


38

MFA, 13/03/00

How to write a paper

More Related Content

What's hot (17)

Similar to How to write a paper (20)

Recently uploaded (20)

How to write a paper