Research Methods in CS

30
Research Research Methods in Methods in CS CS Concision and Simplicity Concision and Simplicity in Scientific Writing in Scientific Writing

description

Research Methods in CS. Concision and Simplicity in Scientific Writing. Concision and Simplicity. Principles: Omit needless words (excessive hedging, ineffectual phrases) Prefer simple words Use simple subjects Use adjectives/adverbs frugally Acknowledgement: - PowerPoint PPT Presentation

Transcript of Research Methods in CS

Page 1: Research Methods in CS

Research Research Methods in Methods in

CSCSConcision and Simplicity Concision and Simplicity

in Scientific Writingin Scientific Writing

Page 2: Research Methods in CS

Concision and SimplicityConcision and Simplicity

Principles:Principles:

1.1. Omit needless words (excessive hedging, ineffectual Omit needless words (excessive hedging, ineffectual phrases)phrases)

2.2. Prefer simple wordsPrefer simple words

3.3. Use simple subjectsUse simple subjects

4.4. Use adjectives/adverbs frugallyUse adjectives/adverbs frugally

Acknowledgement: Acknowledgement:

I based these slides mostly on Duke University I based these slides mostly on Duke University Graduate School Scientific Writing Resource at Graduate School Scientific Writing Resource at https://cgi.duke.edu/web/sciwriting (Concision and Simplicity), but I changed several examples to use computing concepts and terminology.

2

Page 3: Research Methods in CS

Principle 1: Omit needless Principle 1: Omit needless wordswords

Ineffectual phrasesIneffectual phrases Intent to make sentences appear more substantial than Intent to make sentences appear more substantial than

they actually arethey actually are But no sentence made more meaningful by their inclusionBut no sentence made more meaningful by their inclusion

✖✖ Examples:Examples: note thatnote that it should be noted thatit should be noted that respectivelyrespectively it is important to realizeit is important to realize so-calledso-called

3

Page 4: Research Methods in CS

Omit needless wordsOmit needless words

Wordy phrases – MWordy phrases – Multiw0rd phrases that mean nothing beyond ultiw0rd phrases that mean nothing beyond a single worda single word

4

Instead of Consider

a large number of many

due to the fact that because

the question as to whether

whether

there is no doubt that doubtless

used for testing purposes

used for testing

in a careful manner carefully

this is a subject that this subject

a large majority of most

has a capacity to can

whether or not whether

Page 5: Research Methods in CS

Omit needless wordsOmit needless words

5

Instead of Consider

are in agreement agree

prior to before

subsequent to after

at this point in time now

in the event that if

a new initiative an initiative

nearly unique unique / rare

plays a key role in is essential to

both the users were equally affected

the users were equally affected

Page 6: Research Methods in CS

Omit needless wordsOmit needless words

6

Instead of Consider

adding together adding

after the end of after

cancel out cancel

let us now consider consider

divide up divide

totally eliminate eliminate

semantic meaning meaning

completely optimized optimized

separate into partitions partition

Page 7: Research Methods in CS

Omit needless wordsOmit needless words

✖✖ [118 words] As discussed, the second reaction is [118 words] As discussed, the second reaction is really the really the end result end result of of a very large number of a very large number of reactions. reactions. It is also worth emphasizingIt is also worth emphasizing that the that the reactions do not represent a closed system, as r reactions do not represent a closed system, as r appears to be produced out of thin air. In reality, it is appears to be produced out of thin air. In reality, it is created from other chemical species within the cell, created from other chemical species within the cell, but we have chosen but we have chosen here here not to model at such a fine not to model at such a fine level of detail. One detail not included here level of detail. One detail not included here that may that may be worth considering be worth considering is the is the reversible nature reversible nature of the of the binding of RNAP to the promoter binding of RNAP to the promoter regionregion. . It is also It is also worth noting worth noting that these two reactions form a simple that these two reactions form a simple linear chain, whereby the product of the first reaction linear chain, whereby the product of the first reaction is the reactant for the second.is the reactant for the second.

7

Page 8: Research Methods in CS

Omit needless wordsOmit needless words

✔✔ [92 words] As discussed, the second reaction is [92 words] As discussed, the second reaction is really the really the result result of of manymany reactions. reactions. The The reactions do not represent a closed system, as r reactions do not represent a closed system, as r appears to be produced out of thin air. In reality, appears to be produced out of thin air. In reality, it is created from other chemical species within it is created from other chemical species within the cell, but we have the cell, but we have chosenchosen not to model at not to model at such a fine level of detail. One detail not such a fine level of detail. One detail not included is the included is the reversibility reversibility of the binding of of the binding of RNAP to the RNAP to the promoterpromoter. . These These two reactions form two reactions form a simple linear chain, whereby the product of a simple linear chain, whereby the product of the first reaction is the reactant for the second.the first reaction is the reactant for the second.

8

Page 9: Research Methods in CS

Omit needless wordsOmit needless words

Revision TechniqueRevision Technique

Search for the phrases like the ones listedSearch for the phrases like the ones listed

Consider removing or replacing thenConsider removing or replacing then

Make sure revision has intended meaningMake sure revision has intended meaning

9

Page 10: Research Methods in CS

Principle 2: Prefer simple Principle 2: Prefer simple wordswords

Never use a complex word when a simple one will Never use a complex word when a simple one will dodo

Not impressive to use long words—just poor Not impressive to use long words—just poor writingwriting

Choose among alternatives for precise meaning Choose among alternatives for precise meaning neededneeded

Methodology vs. methodMethodology vs. method Method — a way of doing somethingMethod — a way of doing something Methodology — a system of methods followed in a Methodology — a system of methods followed in a

particular disciplineparticular discipline

10

Page 11: Research Methods in CS

Prefer simple wordsPrefer simple words

Utilize vs. use, etc.Utilize vs. use, etc.Use — preferred simple word unless nuances meant — Use — preferred simple word unless nuances meant — use a dictionary or thesaurus to distinguishuse a dictionary or thesaurus to distinguish

UseUse a passive object to accomplish a purpose a passive object to accomplish a purpose

UtilizeUtilize something profitably not designed for the something profitably not designed for the purposepurpose

EmployEmploy person or thing currently idle person or thing currently idle

ApplyApply something general to accomplish a specific, something general to accomplish a specific, practical resultpractical result

Take advantage of Take advantage of (or (or exploitexploit) is similar to utilize but ) is similar to utilize but maybe more opportunistic or selfish, maybe abusing maybe more opportunistic or selfish, maybe abusing thing usedthing used

11

Page 12: Research Methods in CS

Prefer simple wordsPrefer simple words

12

Instead of Consider

methodology method

utilize use

elucidate show

putative (nothing?)

etiology cause

systematic (nothing?)

advantageous helpful

deleterious harmful

prescribed required

erroneous wrong

Page 13: Research Methods in CS

Use simple wordsUse simple words

Revision TechniqueRevision Technique

Search for complex words Search for complex words

Replace if simpler word can convey intended Replace if simpler word can convey intended meaningmeaning

13

Page 14: Research Methods in CS

Principle 3: Use simple Principle 3: Use simple subjectssubjects

14

Scientific writing abounds with complex sentence Scientific writing abounds with complex sentence subjectssubjects

These increase distance between subject (actor) These increase distance between subject (actor) and verb (action) – earlier lectureand verb (action) – earlier lecture

Scientific writers try to cram too much in one Scientific writers try to cram too much in one sentencesentence define complex abstract entitydefine complex abstract entity describe something it doesdescribe something it does

Writers should split such Writers should split such ““multitaskingmultitasking”” sentences sentences into multiple sentencesinto multiple sentences

Page 15: Research Methods in CS

Use simple subjectsUse simple subjects

15

The sequences that had passed our filtering, The sequences that had passed our filtering, trimming, and alignment with ClustalX trimming, and alignment with ClustalX were were scanned for conserved elements across mammalsscanned for conserved elements across mammals..

Long distance between noun and verbLong distance between noun and verb

Many actions hidden in concept nouns Many actions hidden in concept nouns (nominalizations)(nominalizations)

Revision: divide sentence, use strong verbs in first, Revision: divide sentence, use strong verbs in first, summarizing concept noun in second to link backsummarizing concept noun in second to link back

The sequences were trimmed, filtered, and aligned The sequences were trimmed, filtered, and aligned with ClustalX. with ClustalX. The resulting alignmentsThe resulting alignments were were scanned for conserved elements across mammalsscanned for conserved elements across mammals. .

Page 16: Research Methods in CS

Use simple subjectsUse simple subjects

Revision TechniqueRevision Technique

Find the subject (actor) and verb in each sentenceFind the subject (actor) and verb in each sentence

If too far apart, they may have complex subjectIf too far apart, they may have complex subject

Try simplifying subject, e.g., by dividing sentence Try simplifying subject, e.g., by dividing sentence in two or eliminating unnecessary modifying in two or eliminating unnecessary modifying clausesclauses

Consider using summarizing nominalizations Consider using summarizing nominalizations (concept nouns) to simplify subject and link back(concept nouns) to simplify subject and link back

16

Page 17: Research Methods in CS

Principle 4: Use Principle 4: Use adjectives/adverbs frugallyadjectives/adverbs frugally

Often adjectives and adverb modifiers add no meaningOften adjectives and adverb modifiers add no meaning

✖✖This method illustrates the frequency of This method illustrates the frequency of veryvery high- high-energy collisions.energy collisions.

Does Does veryvery add anything to sentence? add anything to sentence?

Can likely distinguish Can likely distinguish high-energy high-energy from from low-energylow-energy

But is there a distinction between But is there a distinction between high energyhigh energy and and very-high-energyvery-high-energy in this context? in this context?

If not, then leave out If not, then leave out veryvery

Many other such uses of adverbs and adjectivesMany other such uses of adverbs and adjectives

17

Page 18: Research Methods in CS

Use adjectives/adverbs Use adjectives/adverbs frugallyfrugally

Repetition problem: using two words where one sufficesRepetition problem: using two words where one suffices

Using multiple synonyms togetherUsing multiple synonyms together““completely and utterly alonecompletely and utterly alone””

““completely alonecompletely alone”” ““alonealone””All have same meaning—generally avoid in scientific All have same meaning—generally avoid in scientific writingwriting

(but might be stylistically useful in popular nonfiction, (but might be stylistically useful in popular nonfiction, fiction, poetry)fiction, poetry)

Using word implied by anotherUsing word implied by another““new inventionnew invention”” ““inventioninvention””InventionInvention implies implies newnew, so , so newnew unneeded unneeded

18

Page 19: Research Methods in CS

Use adjectives/adverbs Use adjectives/adverbs frugallyfrugally

Excessive hedgingExcessive hedging

Scientific writers careful about claims, but too Scientific writers careful about claims, but too many hedges erode confidencemany hedges erode confidence

These results These results suggestsuggest that our method that our method maymay possiblypossibly identify identify putativeputative enhancer elements. enhancer elements. [4 [4 hedges]hedges]

✔✔Our method identifies enhancersOur method identifies enhancers. [no hedges]. [no hedges]

Our method identifies possible enhancersOur method identifies possible enhancers. [1 . [1 hedge]hedge]

19

Page 20: Research Methods in CS

Use adjectives/adverbs Use adjectives/adverbs frugallyfrugally

Demeaning adverbs: using adverbs Demeaning adverbs: using adverbs ““obviouslyobviously””, , ““clearlyclearly””, , ““undoubtedlyundoubtedly””

Sometimes point confusing to readers, not Sometimes point confusing to readers, not clearclear

Author should work to bridge conceptual Author should work to bridge conceptual gap with readersgap with readers

Author should not demean or insult readersAuthor should not demean or insult readers

20

Page 21: Research Methods in CS

Use adjectives/adverbs Use adjectives/adverbs frugallyfrugally

Self-aggrandizement: describing merits of Self-aggrandizement: describing merits of your own work more than deserved your own work more than deserved

Be positive, avoid hedging, but do not Be positive, avoid hedging, but do not inflate importance or novelty of your own inflate importance or novelty of your own work in scientific writingwork in scientific writing

Here we describe an Here we describe an exciting new exciting new groundbreakinggroundbreaking method to … method to …

What about advertising copy?What about advertising copy?

21

Page 22: Research Methods in CS

Use adjectives/adverbs Use adjectives/adverbs frugallyfrugally

Revision TechniqueRevision Technique

Highlight all adjectives and adverbsHighlight all adjectives and adverbs

Ask whether each contributes meaningful Ask whether each contributes meaningful idea or is clutteridea or is clutter

Search specifically for overused modifiers Search specifically for overused modifiers like "very", "extremely", or "clearlylike "very", "extremely", or "clearly””

Cut the clutterCut the clutter

22

Page 23: Research Methods in CS

Example 1Example 1

✖✖These approaches use different kinds of These approaches use different kinds of methodology.methodology.

Problems?Problems? may be pompous soundingmay be pompous sounding different kinds of different kinds of different different methodology methodology methods methods

Revision:Revision:

✔✔These approaches use different methods.These approaches use different methods.

23

Page 24: Research Methods in CS

Example 2Example 2

✖✖To identify RNAs associated with each putative RBP, To identify RNAs associated with each putative RBP, C-C-terminal tandem affinity purification (TAP)-tagged proteins, terminal tandem affinity purification (TAP)-tagged proteins, expressed under control of their native promotersexpressed under control of their native promoters, were , were affinity purified from whole-cell extracts of cultures grown affinity purified from whole-cell extracts of cultures grown to mid-log phase in rich medium.to mid-log phase in rich medium.

Problem: Subject (underlined above) seems too complex, Problem: Subject (underlined above) seems too complex, difficult to parsedifficult to parse

Revision—rewrite complex subject as intro sentence, make Revision—rewrite complex subject as intro sentence, make active, fix dangling modifier, omit unneeded wordsactive, fix dangling modifier, omit unneeded words

✔✔To identify RNAs associated with each RBP, we first tagged To identify RNAs associated with each RBP, we first tagged each RBP using C-terminal tandem affinity purification (TAP) each RBP using C-terminal tandem affinity purification (TAP) tags, and expressed these proteins under control of their tags, and expressed these proteins under control of their native promoters. We then affinity purified these proteins native promoters. We then affinity purified these proteins from whole-cell extracts of cultures grown to mid-log phase from whole-cell extracts of cultures grown to mid-log phase in rich medium.in rich medium.

24

Page 25: Research Methods in CS

Example 3Example 3

✖✖We We estimated estimated that that as much asas much as 12-1812-18% % ((depending on the tissuedepending on the tissue) of inter-species ) of inter-species differences in gene expression levels differences in gene expression levels might be might be explained, explained, at least in partat least in part, by changes in DNA , by changes in DNA methylation patterns.methylation patterns.

Excessive hedging – 6 hedges underlinedExcessive hedging – 6 hedges underlined

Revision (still has 2 hedges):Revision (still has 2 hedges):

✔✔Differences in DNA methylation Differences in DNA methylation could could explain explain 12-1812-18% of differences in gene expression.% of differences in gene expression.

25

Page 26: Research Methods in CS

Example 4Example 4

✖✖Epigenetic eventsEpigenetic events contribute to the etiology contribute to the etiology of diabetes; however, the lack of of diabetes; however, the lack of epigenomic analysis has limited the epigenomic analysis has limited the elucidation elucidation of the of the mechanistic mechanistic basis for this basis for this link.link.

Much simpler revision!Much simpler revision!

✔✔Epigenetic problems can cause diabetes, but Epigenetic problems can cause diabetes, but how?how?

26

Page 27: Research Methods in CS

Exercise 1Exercise 1

27

Page 28: Research Methods in CS

Exercise 1Exercise 1

28

Page 29: Research Methods in CS

Exercise 2Exercise 2

29

Page 30: Research Methods in CS

Exercise 2Exercise 2

30