메뉴 건너뛰기

System Security Research Center

etc

영어 논문 영작 Tip

SSRC 2021.09.04 20:47 조회 수 : 74

Opening sentences of research papers

How academics defeat the blinking cursor

Didem Demirag
Jul 1 · 15 min read
 

Didem Demirag & Jeremy Clark

Imagine that you’ve just wrapped up the research for your next academic paper. Now it is time to write it up. Maybe you’ve already sketched in some text or maybe you will write a full draft top-to-botttom. In both cases, you will eventually find yourself contemplating how to start the paper. Specifically, what will you write as the very first sentence of the paper? The opening sentence feels like it needs to be bold, convey the importance of the subject of the paper, and hook the reader. That is a lot to put into a single sentence.

The novelist Stephen King is said to spend “months or even years” writing an opening sentence. The academic Steven Pinker devotes a portion of his stylography book A Sense of Style to the subject, noting that,

“Good writing starts strong. Not with a cliché (‘Since the dawn of time’), not with a banality (‘Recently, scholars have been increasingly concerned with the question of …’), but with a contentful observation that provokes curiosity.”

For fun, we decided to look at how other researchers approach the opening sentence in their work. Since we work in the field of information security, we studied three years of papers (379 papers in total) from USENIX Security, one of the top venues for security research. More pertinently, it is a venue with open access, easy-to-download full proceedings in a variety of friendly formats that made our job easier. We read every opening sentence (from the main body of the paper, as opposed to the abstract). We thought hard about what the authors were trying to convey (maybe in some cases, we put more thought into it than the authors themselves?). If the sentence is a vehicle, what is it carrying? Very quickly, patterns started to emerge. After numerous iterations, we developed and refined a detailed categorization for opening sentences.

In this article, we share some highlights of what we found. We also have a full explanation of each category (or “code”) at the end of this article, with several examples of each. And you can play with our NVIVO files here.

#1 The Big Picture

Many papers start with a straightforward definition or description (Secure two-party computation allows two parties to process their sensitive data in such a way that its privacy is protected), or a historic fact (In 1996, Wagner and Schneier performed an analysis of the SSL 3.0 protocol) about the research subject. The other majory approach is to establish that the subject is worthy of study, most commonly by arguing its prevalence and ubiquity in our lives (Billions of users now depend on online services for sensitive communication;” “Embedded systems are omnipresent in our everyday life). In some cases, the authors state it is important because, well, it is important (Error handling is an important aspect of software development;” “SSL/TLS is, due to its enormous importance, a major target for attacks).

Less commonly, papers might draw a reader into a paper by establishing a narrative (Consider that you are a domain owner, holding a few domain names that you do not have a better use of). We like this and decided to start this article itself with a narrative (“Imagine that you’ve just wrapped up the research for your next academic paper”).

The figure above shows each of the categories, weighted by number of occurances. To see a definition and several examples of each category, check out our full glossary at the end of this article.

#2 Shortest and longest

Short sentences are striking and easy to remember. The shortest ones from our data set were four words long: Video is ineffably compelling and Software bugs are expensive.”

Long sentences can obviously convey more information or make a complex argument, but they might have to be read a few times to fully parse what they are saying. The longest sentence in our dataset is built around a list: Thousands of new domain names are registered daily that at first glance do not have completely legitimate uses: some contain random characters (possibly used by miscreants [23]), are a composite of two completely unrelated words (possibly used in spam [17]), contain keywords of highly-visible recent events (ex. hillaryclingon.com for political phishing in 2008 [28]) or are similar to other, typically well-known, domain names (ex. twtter.com [27, 32]).”

#3 Lots of writers love the arc of time

While starting a paper with a clause like “since the dawn of time” may be a cliche, the idea of appealing to years or decades of work is extremely common (Over the last few years there have been numerous reports…” “The last decade in cryptography…” “Over the past few years, face authentication systems…” “For several decades, car keys…”). These sentences tend to appeal to the notion that a domain is important because of its longevity.

The award for the sentence that reached the furthest into history is: Redaction of sensitive information from documents has been used since ancient times as a means of concealing and removing secrets from texts intended for public release.

#4 Yes, we found typos

Typos can slip into the opening sentence. Even the text editor we are using to write this article objects to the grammar of: To most people, search engine is the entrance to all sorts of web sites on internet.”

This is not necessarily a bad thing. While writers should perhaps edit more carefully and use grammar checking tools, we also feel that typos could be a welcomed sign of the expansion of the top security venues to researchers around the world, many of whom are not native English speakers. In fact, our article is written by one native English speaker and one non-native English speaker (albeit one who loves languages and is learning five of them).

#5 Literary references

Making references to popular culture or literature is quite common in the title of papers, as well as making meta-references to the titles of previous famous academic papers. It can also be done in the opening sentence. The reference is a nice way to catch the attention of the audience and direct it toward the details to be presented in the paper, assuming that the majority of your audience is familiar with the reference that you made. Even if they are not, it probably comes across as a poetic opening.

We only found one example: It is a truth universally acknowledged, that password-based authentication on the web is insecure. This is a reference to Jane Austen’s famous opening to Pride and Prejudice: “It is a truth universally acknowledged, that a single man in possession of a good fortune, must be in want of a wife.”

#6 The biggest threat

In security, you can imagine that researchers want to tackle the biggest threats. Many use their opening sentence to position themselves as such. While some of them directly state that the subject of the paper is the most prevalent threat (In recent years, unwanted software has risen to the forefront of threats facing users;” “Today, runtime attacks remain one of the most prevalent attack vectors against software programs;” “Remote malware downloads currently represent the most common infection vector.), others prefer to make their point in a stronger way with their choice of words (In spite of extensive industrial and academic efforts (e.g., [3, 41, 42]), distributed denial-of-service (DDoS) attacks continue to plague the Internet).

#7 Colourful commentary

Many opening sentences trend on the boring side. But, does serious research always have to imply boring? Is avoiding some colour in your text a way to prove that the problem you are working on is significant and should be taken seriously?

To add some colour, some papers use rhetoric to make a point. What makes these sentences more vivid is the rich choice of words (The defacement and vandalism of websites is an attack that disrupts the operation of companies and organizations, tarnishes their brand, and plagues websites of all sizes, from those of large corporations to the websites of single individuals;” “The dismissal of human memory by the security community reached the point of parody long ago;” “Video is ineffably compelling).

#8 Collisions

With lots of overlapping research areas, it is perhaps inevitable that two different papers will start the same way. We found several such collisions where the exact wording differs, but the sentiment is largely the same. For example, compare due to the growth in online services, many users now manage dozens of password-protected accounts,” and with the proliferation of Web services, ordinary users are setting up authentication credentials with a large number of sites.”

These two are so similar that there is perhaps another explanation: Private set intersection (PSI) allows two parties P1 and P2 holding sets X and Y, respectively, to identify the intersection X ∩ Y without revealing any information about elements that are not in the intersection and Private set intersection (PSI) allows two parties P1 and P2 with respective input sets X and Y to compute the intersection X ∩ Y of their sets without revealing any information but the intersection itself. Indeed these two papers share multiple co-authors.

Finally, we even found a potential three-way collision although if you carefully parse the sentences, they argue slightly different things. Despite years of study, memory corruption vulnerabilities still lead to control-flow hijacking attacks today and Attacking software systems by exploiting memory-corruption vulnerabilities is one of the most common attack methods today according to the list of Common Vulnerabilities and Exposures both argue the ubiquity of attacks today, while the next sentence argues for its importance: Memory corruption vulnerabilities are currently one of the biggest threats to software and information security.” You could be forgiven for thinking they all sound the same though.

#9 Oddest sentence

Before giving you our top 5, let us tell you the weirdest sentence we came across: The battle for the living room is in full swing.” Admittedly, we extracted the opening sentences without reading the broader context of the opening paragraph. We are sure this sentence is made clearer with all of the other sentences that follow it. But as an individual sentence, it raises more questions than it answers. La-Z-Boy vs Ikea? Sony vs Samsung? What does this have to do with security?

#10 Our top 5 favourites

Whether you like an opening sentence is completely subjective. Based on our personal taste, these are our favourite sentences:

5) It is difficult to keep secrets during program execution.”

4) Our phones are always within reach and their location is mostly the same as our location.”

3) Do programmers leave fingerprints in their source code?

2) The dismissal of human memory by the security community reached the point of parody long ago.”

Each of these sentences left us wanting to read more. They draw us into the paper. Some by being simple and direct (#5), while #2 skewers conventional wisdom. However our top pick is:

  1. Peripheral devices are now powerful, portable, and plentiful.”

This sentence stands out among all the other sentences we read for its lyrical quality. Steven Pinker agrees that the alliteration is under-appreciated, “Killjoy style manuals tell writers to avoid alliteration, but good prose is enlivened with moments of poetry…”

The Closing Lines

Admittedly, we concede it is a little frivolous to shine such a big spotlight on the opening sentences of research papers when the results of the paper are obviously paramount. We further admit to sometimes skipping the entire introduction section of papers we read just to get to the good stuff. However, we hope to have inspired you to care about the reader and write more consciously. Finally, if you are stuck, you can use our glossary below to brainstorm ideas of how to open your next paper.

Thanks to Arvind Narayanan for comments.

USENIIX Security 2011

Appendix: Description of Each Category

Facts: Claimed Fact

Another neutral approach to an opening sentence is to provide a fact that is relevant to the subject of the paper. Later we will discuss arguments which are often expressed as if they are facts but are only debatably true. A claimed fact’s correctness should either be apparent or at least provable (i.e., falsifiable).

Users are often advised or required to choose passwords that comply with certain policies.

Mobile apps frequently demand access to private information.

For several decades, car keys have been used to physically secure vehicles.

Some sentences use stronger and more vivid language but are still factually based.

In spite of extensive industrial and academic efforts (e.g., [3, 41, 42]), distributed denial-of-service (DDoS) attacks continue to plague the Internet.

Facts: Technical Advances

Many opening sentences lay out a technical advance in the subject of the sentence. This creates a window of opportunity for the researcher to later identify a novel research problem caused by the changing technology. It is common to see words like: evolve, become, transition, and improve.

Recent advances in cloud computing enable customers to outsource their computing tasks to the cloud service providers (CSPs).

Browsers have evolved over recent years to mediate a wealth of user interactions with sensitive data.

Since its beginning in the early nineties, the Web evolved from a mechanism to publish and link static documents into a sophisticated platform for distributed Web applications

Facts: Historic Events

A final neutral opening sentence will refer to some historic event.

In 1996, Wagner and Schneier performed an analysis of the SSL 3.0 protocol [67].

In February 2011, a new Tor hidden service [16], called “Silk Road,” opened its doors.

The Network Time Protocol (NTP) is one of the Internet’s oldest protocols, dating back to RFC 958 [15] published in 1985.

In some cases, a paper opens with a “compound” sentence that makes reference to a historic event in one clause of the sentence, while having additional clauses of a different category. For example, the following sentence refers to a historic event as well as a technical advance.

Starting from Denning’s seminal work in 1986 [9], intrusion detection has evolved into a number of different approaches.

Arguments: General Argument

Many opening sentences issue a subjective argument that represents the authors’ opinion. Unlike a fact, it isn’t straightforward that the reader will accept it as true. While arguments are less neutral than facts, they can be more interesting and provocative, which can help draw the reader into the paper.

The arguments we categorize under “general arguments” do not fit elsewhere in our categorization system. As we go through more categories, we will see other more specific kinds of arguments.

It is a truth universally acknowledged, that password-based authentication on the web is insecure.

The dismissal of human memory by the security community reached the point of parody long ago.

In recent years, unwanted software has risen to the forefront of threats facing users.

The phenomenal growth of Android devices brings in a vibrant application ecosystem.

Suitability: Importance of subject

A large set of sentences make a special kind of argument: that the subject of the opening sentence is suitable or worthy of research. The exact reasons they are suitable fall into a few sub-categories: the subject is important, ubiquitous, complex, novel, popular with other researchers, or has been around a long time.

Many opening sentences state that their subject is important, with the implication that it is thus suitable for research.

Security has now become an important and real concern to connected and/or automated vehicles.

Error handling is an important aspect of software development.

SSL/TLS is, due to its enormous importance, a major target for attacks.

Some sentences do not explicitly use the word “important” but find other ways to convey the same notion. For example, a concern or component might be described as essential or crucial or serious.

The threat of data theft in public and private clouds from insiders (e.g. curious administrators) is a serious concern.

The same-origin policy (SOP) is a cornerstone of web security, guarding the web content of one domain from the access from another domain.

Suitability: Ubiquity of subject

The most popular kind of opening sentence argues that a subject is suitable for research because it is ubiquitous and widely used.

Billions of users now depend on online services for sensitive communication.

Embedded systems are omnipresent in our everyday life.

Android is the major platform for mobile users and mobile app developers.

Suitability: Longevity of subject

In this category, how long a subject has been around is the key component to why it is a suitable subject for study. In some cases, a specific duration is provided and in others, it is implied that the amount of time is significant.

Redaction of sensitive information from documents has been used since ancient times as a means of concealing and removing secrets from texts intended for public release.

Since its beginning in the early nineties, the Web evolved from a mechanism to publish and link static documents into a sophisticated platform for distributed Web applications.

Narrative

A potentially interesting way to draw a reader into a paper is by establishing a narrative: a scenario that gets the reader thinking about themselves or other people and what they might do.

Consider that you are a domain owner, holding a few domain names that you do not have a better use of.

Consider the setting where a client owns a public input x, a server owns a private input w, and the client wishes to learn z := F(x,w) for a program F known to both parties.

Narratives might also set a scene, like the academic version of an “establishing shot” from films and TV.

Our phones are always within reach and their location is mostly the same as our location.

We live in a “big data” world.

The battle for the living room is in full swing.

Question

Making the reader curious is another good way to begin a paper, and this can be accomplished using a question. In our sample, this was underused with only one example.

Do programmers leave fingerprints in their source code?

위로