Content uploaded by Mohammad Tahaei
Author content
All content in this area was uploaded by Mohammad Tahaei on Oct 29, 2021
Content may be subject to copyright.
“I Don’t Know Too Much About It”: On the
Security Mindsets of Computer Science Students
Mohammad Tahaei, Adam Jenkins, Kami Vaniea, and Maria Wolters
School of Informatics, University of Edinburgh
{mohammad.tahaei, adam.jenkins, kami.vaniea, maria.wolters}@ed.ac.uk
Abstract. The security attitudes and approaches of software developers
have a large impact on the software they produce, yet we know very little
about how and when these views are constructed. This paper investigates
the security and privacy (S&P) perceptions, experiences, and practices
of current Computer Science students at the graduate and undergradu-
ate level using semi-structured interviews. We find that the attitudes of
students already match many of those that have been observed in pro-
fessional level developers. Students have a range of hacker and attack
mindsets, lack of experience with security APIs, a mixed view of who
is in charge of S&P in the software life cycle, and a tendency to trust
other peoples’ code as a convenient approach to rapidly build software.
We discuss the impact of our results on both curriculum development
and support for professional developers.
Keywords: Usable Security ·Secure Programming ·Computer Science
Students ·Software Developers ·Software Development ·Education
1 Introduction
Software developers can impact millions of lives with seemingly small security
decisions that have a large impact on the people using the technologies. One ex-
ample is the case of the dating site Ashley Madison, where a strong cryptographic
algorithm was used to store passwords but was implemented incorrectly [42].
Even for apps where security is not a primary feature, it is a requirement
needed for stability and safety of operation. Therefore, software developers need
to be keenly aware of the security implications of their design decisions. Ideally,
they should have strong support from their tools to avoid security and privacy
issues in their resulting code.
Basic tools such as cryptographic libraries (OpenSSL) and federated authen-
tication (OAuth) exist partially to assist developers in integrating common se-
curity needs into their projects without needing to know all the complex details.
There are also efforts to help raise awareness of common coding and design issues
such as the IEEE top ten security flaws [4, 5].
Yet, security remains a pervasive problem in deployed code. In 2013 alone,
88% of apps (out of 11,748) analysed on Google Play had at least one mistake
in how the developer used a cryptographic API [19]. Code that they write goes
2 M. Tahaei et al.
into security-critical applications such as banking software [27] as well as software
with less obvious security implications such as Internet connected kettles [46].
Non-usable APIs are a key point of failure for most developers [3, 29, 35, 53,
70]. Providing manuals is not enough. A usability evaluation of programming
security in Android found that developers created code with security errors
even when they were provided with official documentation [2]. Perhaps more
importantly, developer understanding of security is also problematic. Interviews
with professional developers show a range of concern about security and privacy
knowledge [11]. The situation is exacerbated when developers make non-obvious
errors when implementing security which results in believing that code is secure
when it is actually not secure [1].
One potential opportunity for changing developers’ security attitudes and
practices is during their training. In this work, we investigate the security and
privacy (S&P) mindsets of a group of twenty graduate and undergraduate com-
puter science (CS) students on a variety of career trajectories, and with a range
of exposure to formal security training. Our research questions are:
–What are students’ comprehension of S&P related concepts?
–To what extent do students consider S&P while coding applications, and
how do they implement it?
Within the context of developer-centred security, our study highlights the
extent to which students already have similar mindsets and practices as have
been found in professional developers, suggesting that these may form and con-
solidate early. We conclude that, while early educational intervention would be
ideal, we also need to provide developers with usable tools, such as APIs, and
easily accessible training, which can be used both by trainees and professionals.
2 Related work
Creating secure software correctly is quite challenging even for professional de-
velopers, often resulting in unintended security vulnerabilities [3, 29, 70]. The
OWASP organisation publishes the top ten most critical web application security
risks every few years. A review of their last three reports covering seven years
terrifyingly that the most common issues are quite stable [51], with common and
highly damaging vulnerabilities such as code injection and broken authentication
continuously remaining in the top ten.
Arce et al. observed that many of the OWASP vulnerabilities represent un-
intentional errors or mistakes rather than planned actions and therefore are
minimally helpful to someone trying to design a secure system [5]. Instead they
propose a set of top ten security design flaws, that is security issues that are
a planned element of the software. Their list is much higher-level and contains
issues such as “earn or give, but never assume, trust” [5, p. 9].
The problem of code vulnerabilities in live software is further exacerbated by
the steady reduction of the barriers to entry for new software creators. While
generally a good thing, the ‘anyone can code’ movement has also led to an
On the Security Mindsets of Computer Science Students 3
increase in the number of software creators with minimal formal training in
software development and even less training in security. Unsurprisingly, this
group also has difficulty creating secure software [50, 53].
Neither of these groups is, or should be, expected to be security experts,
but the decisions they make can still have serious security impacts. In an effort
to better support these software creators, several tools and libraries have been
proposed such as OpenSSL, PyCrypto, and cryptography.io which encapsulate
many of the security decisions, theoretically making development easier.
Unfortunately, many of these tools still suffer from usability issues, such as
confusing API designs [1, 19, 22, 27, 34, 39, 64] or poorly designed documenta-
tion [2, 47]. Official documentations are often not easy to use, hence developers
prefer online resources which may not offer valid and secure solutions. While
Stack Overflow, for example, helps with getting code working quickly, the sug-
gested solutions may also result in less secure code [2, 23].
Security is also challenging for developers because it causes no obvious vi-
sual effect, making it difficult to identify when an unintended state has oc-
curred [21, 22]. A common example of invisible security effects is SSL/TLS.
When used incorrectly, a connection is still formed, but that connection might
not be encrypted, or it might be encrypted, but without certificate validation.
This results in a vulnerability to man-in-the-middle (MITM) attacks during con-
nection setup. Fahl et al. observed how challenging this can be for developers to
spot. One of their developers even used Wireshark to ‘test’ the SSL/TLS con-
nection and, because the data was garbled looking, incorrectly concluded things
were working even though no certificate checking was happening [22].
Georgiev et al. similarly conducted an analysis of SSL certificate validation
in multiple non-browser platforms and showed that many applications on the
market are open to a MITM attack where data can be read and modified in
transit because developers accidentally or intentionally configure their code to
not validate the certificate source [27]. Such problems arise when developers are
expected to understand the implications of the different settings of SSL, which
is exacerbated by APIs that do not offer a helpful level of abstraction [34].
Security is also not a well-established requirement in the software develop-
ment workflow. Without a dedicated developer in charge, security becomes a
hot potato which is passed between groups because no one wants to deal with
it [11, 49, 54, 68]. In interviews with security experts, Thomas et al. found that
security auditing is seen as a separate task from software development. While
security auditing is performed by the rare breed that are security experts, it is
then the developer’s job to fix the security issues found [63].
Many future software developers were once Computer Science (CS) students.
A survey by Stack Overflow in 2019 showed that 62.4% (75,614 responses) of
developers have a degree in CS, computer engineering, or software engineer-
ing [60]. Given the importance of this group, many researchers study them to
either address gaps between academia and industry [16, 37, 55, 56, 61] or to
suggest educational tools to improve their skill and abilities [48, 62, 69]. Re-
search shows that CS students often work under misconceptions which can lead
4 M. Tahaei et al.
to bad practice. For example, when it comes to software engineering processes
and teamwork [61], many think that working alone is a quicker way of working on
a software project, which goes against established industry best practice. Here
we study the S&P mindsets of CS students with a view to identifying what they
know and think about S&P, and what misconceptions exist.
3 Methodology
We used semi-structured interviews to explore how a range of students from
undergraduate to PhD think about S&P. The semi-structured approach allowed
us to probe students’ S&P mindsets in detail and investigate how they relate to
their own practices as developers.
3.1 Interview design
After informed consent, we explicitly invited participants to talk as much as they
wanted on the various topics discussed. The interview began with an open ques-
tion on academic and professional background and general questions about cod-
ing and software development experience. Questions about demographics were
asked at the end of the interview in order to minimise stereotype threat. The
full interview script is included in the Appendix.
We began the S&P discussion by asking participants to consider creating
“a new group discussion app for in-class discussions.” They were then asked to
free-list the app’s features on paper and after they finished they were asked to
circle those that were S&P related.
Next, we examined participants’ understandings around threats and hackers.
We started by asking participants about the hypothetical app: “Who is most
likely to try and attack this system? What are they likely going to try and
do?” We then moved on to talk about hackers, because work on security folk
models has found them to be an important part of how people think about
security [67]. We elicited participants’ definitions of the term hacker, and their
views on hackers’ intentions, goals, and background.
We then moved on to considering who was responsible for S&P in software
development practice. The discussion was grounded in participants’ own experi-
ence of writing software, in particular problems with (security) APIs.
Finally, we asked participants about personal security and privacy practices.
First, participants were asked to list the words and concepts they associated
with ‘computer security’ on paper. We followed up with questions about good
security practices, and their own security practices.
Since prior negative experiences can impact future choices [65], we also asked
about prior experiences with compromise, prompting them with examples such
as “getting a virus on your computer, losing your password, having an email sent
from your account, or loss of data about you” if needed. We explored how the
experience was resolved, and what participants learned.
On the Security Mindsets of Computer Science Students 5
3.2 Recruitment
We recruited participants through mailing lists associated with a large Russell
Group University in the United Kingdom, Facebook groups, and word of mouth.
Advertisements asked for Computer Science students (BSc, MSc and PhD) to
participate in an interview about opinions and attitudes around software de-
velopment, particularly around the handling of requirements prioritisation. All
advertisements avoided S&P related words to limit self-selection and priming.
3.3 Participants
Our sample, shown in Table 1, includes twenty students (6 BSc, 11 MSc, and
3 PhD students), participants who previously took a computer security course
at any University are indicated with ‘PS’ instead of ‘P’. The sample contains
five female, and fifteen male students with an average age of 24 years old (range:
20−37, std: 3.8, median: 23). They come from various countries and have diverse
CS-related educational backgrounds. Interviews were conducted in English. Our
sample reflects both the diversity seen in the tech industry [10, 28], and the
culturally diverse classrooms found in many computer science departments.
The interviews were advertised to be 60 to 90 minutes long with a compensa-
tion of £10 in cash. In practice, interviews took an average of 68 minutes (range:
41 −108, std: 18.4, median: 65.5) and were completed in July 2018. All inter-
views were audio recorded with participant consent. The study was conducted
in accordance with the Ethics procedures of the School where the students were
recruited (cert ID 2870).
We interviewed students over the summer. This meant that the Masters
students were in their dissertation phase, and had completed the course work
part of their 12-month degree. PhD students in the UK have typically completed
a Masters before starting a PhD and are not necessarily required to take courses,
pass a qualifying exam, or be a Teaching Assistant, though many choose to take
additional courses and tutor. Therefore, beyond teaching and thesis work, PhD
students are unlikely to be impacted by security courses taught at the University.
3.4 Pilot
We conducted seven pilot interviews with Masters and PhD students, six of
which were associated with our research lab but unfamiliar with the work. These
interviews were used to iteratively refine the interview script as well as adjust
the number and content of questions to keep interviews at about 60 minutes.
The pilot contained some students with no security background to help ensure
the phrasing of security questions was clear. Feedback was also sought about the
structure, clarity, and accuracy of the interview schedule. Pilot interviewees and
interviews were not used in our final analysis.
6 M. Tahaei et al.
Table 1. Interview study demographics. P = participant without computer security
background; PS = participant who self-describes as having taken a computer security
course in the past.
Participant Gender Nationality Age Expected Degree
PS01 M EU 29 PhD
P02 M EU 28 MSc
PS03 F Asia 22 MSc
PS04 M Asia 24 MSc
PS05 M Asia 25 PhD
P06 F Asia 23 MSc
P07 M Asia 22 BSc
PS08 M UK 21 MSc
PS09 M Asia 25 MSc
P10 M Asia 21 BSc
P11 M EU 22 BSc
PS12 M Asia 23 MSc
PS13 M EU 21 BSc
P14 M EU 20 BSc
PS15 M EU 25 PhD
PS16 M Asia 37 MSc
P17 F EU 25 BSc
P18 F Asia 23 MSc
P19 M UK 24 MSc
P20 F Asia 20 MSc
3.5 Interview analysis
Interview analysis focused on uncovering students’ mindsets of S&P as they re-
late to the software development process. Relevant themes were extracted using
a three stage process. First, two researchers listened to the full audio of four
interviews which had been selected by the interviewer to cover a wide range of
participants, identified relevant parts for more detailed analysis and transcrip-
tion, and outlined an initial topic guide for coding [45, 59]. Audio was used
because it provides a richer record of the original interview than a standard
transcript. In the second stage, the researchers performed open coding of the
transcripts based on the topic guide [45, 59].
In the third stage, the open codes were analysed using an affinity diagram [40]
to yield a set of seven themes, which are discussed in the Results Section 4 be-
low. While some authors suggest reporting how many participants mention each
theme [40], we chose to follow standard qualitative research reporting practice
and focus on describing and contextualising our themes [57, 67].
4 Results
All participants all had some form of prior programming experience ranging
from classroom projects, internships, and prior employment in industry. Since
On the Security Mindsets of Computer Science Students 7
our participants included a large number of Masters students, they also had
classroom experience from prior universities, with several expressing that they
had worked in industry either as interns or full time before coming back for a
Masters or PhD. Half had taken a computer security course at some point in
their education. We did not ask about the details of these courses.
4.1 ‘Computer security’ word association results
Mid-way through the interview participants were asked to free-list words associ-
ated with ‘computer security’. The words were grouped into topics by the lead
researcher with a bottom-up approach. A second researcher then reviewed the
groupings and disagreements were resolved through discussion. Table 2 shows
the resulting eleven topics.
Participants’ understanding of the term ‘computer security’ was broad, with
participants who wrote words providing an average of 9.6 words (range: 2 −19,
std: 4.2). Listed words included standard security topics such as encryption,
attacks, and system security which are readily found in most security text books.
Participants also listed company names that are either associated with security
(Norton) or that had been discussed recently in the news in relation to security
(Facebook [36, 66]). Two participants (P02 & P20) were not able to list any
words, suggesting uncertainty with the term. “It is all very flimsy ” (P02), “To
be honest I do not know too much about it” (P20).
Of the participants who provided words, participants listed words from an
average of 4.2 topics (range: 1 −7, std: 1.8). The topics cover a wide range,
but each individual participant had less range, with at most seven topics men-
tioned by one participant. Most notable is the lack of a single common topic
amongst participants. For example, the most common word ‘privacy’ was men-
tioned by only 40% of participants. Common security topics such as passwords,
authentication, and encryption also appeared. Some of these topics are similar
to what professional developers associate with security, for example, encryption,
user control, and user access [30].
4.2 Interview themes
Security mindsets. Participants varied substantially in their understanding of
S&P. While some participants had a strong up-front understanding of security
which varied minimally during the interview, others had clearly not thought
much about the topic before resulting in them re-thinking their opinions mid-
interview. This is to be somewhat expected as many people have not previously
devoted extensive time to assessing their own understanding of the topic [7].
This theme provides rich additional context to the initial topics identified
through free association. Those with a more sophisticated understanding of S&P
tended to use more definitive language, had more stable descriptions of attacker
motivations, and were more likely to be sure that their statements were accurate,
8 M. Tahaei et al.
Table 2. Topics mentioned during free-listing, number of words participants listed
associated with that topic, number of unique participants listing at least one word
associated with the topic, and a set of sample words representing the range.
Topic #Words #Participants Example words
Encryption 28 11 End-to-end, hash, RSA, public/private key, SSL, symmetric.
Authentication 28 9 Passwords, permissions, 2FA, tokens, access controls, emails.
Privacy 27 10 Anonymity, right to be forgotten, visibility, cookies.
Attacks 25 8 Reconnaissance, phishing, buffer overflows, DoS, MITM.
System security 13 5 Protocols, database, Unix, system calls, TCP/IPs.
Social 13 7 Regulations, roles, responsibilities, public knowledge.
Finance 8 4 PayPal, Apple Pay, Bitcoin, online payments.
Defending 7 5 Anti-virus/malware, penetration testing, logging, bounties.
Security holes 5 4 Failures, physical access, loopholes.
Companies 5 3 Facebook, Google, Norton, Red Hat.
Trade offs 4 3 Usable security, features vs security, easy to use UX.
and to describe less intuitive or extreme scenarios. For example, PS15, a crpy-
tography PhD student, explains that “in crypto, we assume that the attacker is
any code, literally any Turing machine” (PS15).
Those with an initially less sophisticated understanding of S&P showed signs
of forming their opinions as the interview progressed. Often, this would involve
contradictions in thoughts as they finally reached a definition for themselves.
This was most notable for the hacking theme. Participants with less developed
models exhibited less self-assurance around motivations, or definitions of attack
scenarios. “I think [HTTPS] is standard by now, don’t they? The more encryption
the better? [...] Like exchange of data that’s not encrypted at all. I don’t think
that’s happening anymore. I’m not sure but I don’t think it is” (P17).
Similar to non-tech savvy users [67, 72], some of our participants think they
are not a target for attackers. “We are just average people. It is ok to have small
security measures” (P11), “I am also very boring computer user. I just do my
courses and I watch movie on Netflix. So I don’t really do anything that could put
me in front of a virus” (P20). Conversely, some participants had high awareness
of potential attacks, though they still did not perceive themselves as at risk. “I
am running a server at home, which has an SSH access available. There you can
see a lot of stuff going on, there are just bots or so whatever trying to get into.
That is even a bit scary if you see that happen all the time, but I think my pass
has been strong enough to keep them out ” (PS13).
Participants clearly evidenced their own internal struggle over what S&P
actually was and when it was or was not needed, which might partially explain
its lack of inclusion in initial requirements. “[My address] is not so important,
because every website is required. Maybe because I live in a dormitory, if it is in
my home that is different” (P06).
While participants understood that private data should be protected, they
struggled with what ‘private data’ actually meant. Even when talking about S&P
On the Security Mindsets of Computer Science Students 9
in their private lives, participants had mixed opinions about how problematic it
was for data like bank transactions to be leaked. “So the data [leak] was about
the full info about the bank accounts, the transactions, in and out, the current
amount in it. For me it was normal [...] to have these transactions. But for some
people it was an issue, because they receive money from hidden source, so it was
an issue for them” (PS16).
Who are hackers and what do they want? Some participants’ definitions
of hackers were well articulated. “Really theoretical let’s say, the adversary we
say in crypto is literally anyone that has a computer and some access to your
systems” (PS15). Other participants had a more general understanding. “The
images that you have in your head are from Hollywood. Super smart kids sitting
in the corner of a room then CIA calls upon them to solve a problem” (P02).
We found a wide range of imagined intentions for hacking, such as financial,
personal, political, and just for fun. All four types of previously observed models
of hackers from Wash’s work [67] were mentioned by our participants:
Graffiti, which is a mischief causing attacker with technical background:
“Want to try what they learn from the class. They may write some code to hack
some system of the school to show their ability” (PS12); Burglar, who commits
crimes using computers mostly with financial motivations: “There is nothing
but personal interest. Personal gain. Personal satisfaction. And of course they
are who just do it for financial gain. Stealing identities, pictures, personal info.
Just to sell it afterwards, to like black market” (PS01); Big fish, who looks only
for high valued targets: “Political incentive that certain countries fund a lot of
hacking and cracking to gain power depending how important or how famous you
are there might be people who want to get access to your account” (PS13); and
Contractor, a Graffiti hacker with financial/criminal motivations: “Trained peo-
ple who are trained to do this kind of stuff. Either by some governments to hack
other governments. Or to break the encryption or security mechanism ” (PS05).
The role of security when planning software. When participants were
asked about what features they would consider in an in-class discussion app,
they commonly mentioned functional requirements including task management,
calendar, question/answering, recording classes, and assignment management.
Many of these features currently exist in course management software with which
the students are familiar, such as Blackboard LEARN and Piazza.
Only four participants (PS08, PS15, PS16, and P19) mentioned S&P in their
initial design and feature list, a somewhat small number since ten of our par-
ticipants had previously taken a security course. Only two of the participants
proactively brought up privacy issues. “First thing that comes to my mind is pri-
vacy. Definitely in terms of features. Presumably, the School will wish to host it
locally rather than to have some sort of central cloud back service” (P19), while
PS08 noted the connection between privacy and ethics: “There is some ethical
questions involved in the area of student privacy” (PS08).
10 M. Tahaei et al.
Security of the data was also a concern, particularly in terms of information
leaks. “I will make sure of the safety and security because [no one] wants to use
the tool if he feel he is vulnerable his info may leak to any unwanted person”
(PS16). PS15 was also able to pull on prior experience and identify specific
attacks and solutions that needed to be addressed: “For sure I put HTTPS and
TLS around it. So that would be safe. Because still, I would leave a lot of surface
for attacks, because the big applications have more surface for attacks.... All
those places where there is user input we basically talk about security, and we
have to remember SQL injection and stuff like that” (PS15).
Some participants turned to more authoritative sources such as laws, regu-
lations, and public policies as a guide for what should and had to be built into
the system. “You have designed an app I guess you also think about security.
But you also think about engagement. Does a certain security feature if it is op-
tion not legally required, how does it sort of effecting the engagement” (P02).
Some mentioned the EU General Data Protection Regulation (GDPR, enforce-
ment date: 25 May 2018) [20], either as a convenience tool for end-users or from
a regulation perspective for companies. “Do we have to be GDPR compliance?
Probably, I’m guessing” (P11) was mentioned by a participant when answering
a question about what S&P features his hypothetical classroom app might need.
Requirements and responsibilities: playing hot potato. Several partici-
pants recognised security as an explicit requirement. They consider the develop-
ers’ job to be transforming requirements into code. Therefore if security is an
explicit requirement, then they have to take it into account during design and
development. “So as a software engineer, if I am already given a certain require-
ment, I should not care about anything else outside the specs. You are employed
as software engineer, you just write your. You are given a list, you just have to
code it. Right? Unless you can do that. You are still doing your job” (PS01).
On the other hand, other participants see security as an implied requirement
that is always present. “When the requirement is out but [privacy] has to be taken
care of at every single step here. If someone comes to me asking for something
then I assume that I do security for all the requirements. Wherever applicable
security should be” (PS04).
Security was also sometimes seen as a problem or requirement that should
be solved by a designated entity within their workflow. For some participants,
this entity was the operating system “Android, it is responsible. Because Android
restricts my way of developing an application. So it should provide sufficient se-
curity mechanism for me to rely on” (PS05), “Mostly the OS is the one that
should provide security” (PS15). Others considered that a security team in the
software development workflow should be responsible. “There should be a secu-
rity team. Which takes care of that. Just like any other team inside the company.
Like UI, testing team” (PS04).
Many interviewees thought that the company as a legal entity is responsible
for S&P, and some highlighted the role of legislation and government. “We are
[responsible]. Not me personally but the company that I work for as a legal entity”
On the Security Mindsets of Computer Science Students 11
(P17). Moreover, a few saw end-users having some responsibility as part of the
larger S&P ecosystem. “There should be a certain amount of onus on the user,
they should be responsible for like managing their password” (PS08).
General attitudes to APIs. Participants saw APIs as a useful and handy tool,
especially in terms of code re-use “GUI stuff in python, here you can just call
functions without write whole part of code yourself. It’s always handy” (PS03).
APIs also allowed them to lean on the knowledge of others and not need to
understand all the concepts themselves. “It is quite useful and simple to import
the library from platform. Before I used that library I need to learn each algorithm
one by one mathematically. In terms of math the algorithm is quite hard. With
library I just can, I import them from Internet. With one or two lines code I
can use them. I can focus more on main procedure of neural network and data
manipulation, so I can save a lot of time with the library ” (PS12).
Other peoples’ code was a large theme when discussing APIs, particularly
examples posted online or documentation-like guidance from others. “Sometimes
just some posts either forums or some question and answer community like Stack
Overflow. There are people show you how to use in their answers, kind of you
can copy paste and modify that to suit your needs” (PS05). APIs also tended
to be designed in such a way that they were easy to start using. “Maybe it is
just experience, that makes it easier, because I was using APIs for so long so
it is easy now to just come and start” (PS01). APIs also made it easy to get
code running quickly, especially if the documentation was good and contained
examples. “If you pick a certain thing, you read the documentation, hopefully the
documentation is done well, by done well I mean by examples. That you can get
something to run as fast as possible because that keeps you motivated” (PS13).
Security APIs. When asked about a ‘security API’ participants struggled to
understand what that could even be, falling back on areas commonly associated
with security, like finance. “What do you mean by security APIs? Something like
payment gateway?” (PS04). Only one participant had a hands on experience
with a security API which was problematic. “There is no feedback [in Android
certificate validation]. It is a complete nightmare, various very long complicated
classes archaic options that you are supposed to set. All and all was 40-50 lines
of code. This was just a block of imperative commands for doing something basic
like I’d like to validate against certificate file please. Absolutely crazy” (P19).
While only observed from one participant, his comments closely match what
other researchers have observed from professional developers [6, 19, 22, 27].
P19, who has industry experience both as a developer and an intern, was
one of the few people who discussed issues around secure programming, such
as buffer overflow and functions with known security issues. “buffer overflows,
system calls are an issue of languages, actually more that anything else. We still
use C this is an atrocity, we shouldn’t be using C anymore” (P19). He is referring
to common C function calls like gets which are impossible to use in a secure
way, but are still commonly used due to being part of core C [38].
12 M. Tahaei et al.
Trusting other peoples’ code. Using APIs and examples from the Internet
was convenient for our participants, but it also required them to trust people
they had never met. Some were concerned about blindly trusting code from un-
known sources, but many had no problem instead choosing to trust in collective
intelligence. “If I download, I am often downloading source codes myself from
the Internet and then building it. And again I don’t have the time or the skill
to audit say a code base that has millions of lines. I perhaps trust a little bit
too much the crowd of people. If I look at the code base and see something on
Github and it has let’s say 2000 stars. Few hundred people watching it. The code
is all open. I tend to perhaps foolishly I assume that if this many people have
looked at it and if there was something up. Surely someone would do have said
something. Download the code and build it. So it is possible that I have exposed
myself to security issues as a result of that” (PS08). PS08 is referring to the
‘many eyeballs’ idea in open source software which is an indicator of security
and reliability of code for some developers [32].
Trust is an inherent component of open source, that is code is open for
everyone to read. “As one of the reason I really want an open source app to
do this is that this kind of app is allowed to access a lot of info. I don’t trust
any closed source software. I could use them but I don’t trust them. Open source
is the only way I could trust software. Although open source you could still add
malicious code to open source in hope that people wouldn’t discover that. But this
is the only way” (PS05).
Trust in open source reaches to its highest level when people prefer to write
less code and reuse others’ code instead. “So my idea is that the least I code the
better. As long as [hypothetical app] is still maintained and supported regularly
and I do update that regularly. Then I think I will be fine. Because tools that
are widely used are very exposed to criticism so their maintainers usually patch
up and correct their mistakes as fast as they can. So I’d very aware of what of
dangers of the whole thing. And I would be careful to following news. But I’d
avoid writing my own code” (PS15) is a comment on open source software while
the participant was discussing her hypothetical classroom app.
5 Discussion
Security mindsets. Mindsets are likely to influence actions and decision mak-
ing [41, 43, 67]. We found that most students did not have a clearly developed
concept of security. In fact, some participants even struggled to come up with
words that could be associated with the term ‘computer security’. When it comes
to threats such as hackers, what they can do, their intentions and capabilities is
another point which needs improvements, we observe the similar patterns and
folk models in CS students that others have seen in home users [44, 67].
Mental models could be partially rooted in media [25]; participants cited
media plot elements when describing hackers. End user security has seen success
in teaching users to copy existing mental models such as viruses or home safety
to better understand and reason about security and privacy [17]. Our results
On the Security Mindsets of Computer Science Students 13
suggest that similar approaches may work in the educational context to improve
the mental models of students.
APIs. When it comes to APIs, our results closely mirror what related work
has shown for professional developers. They often use a combination of online
resources to learn and use APIs. They prefer to use easier to use resources, and
because official documentation is often not easy to use they tend to go for online
resources like Stack Overflow [2]. Professional developers (like our student sam-
ple) prefer documentation with examples and matching API scenarios [52, 58].
Therefore, API designers are a significant element in secure software develop-
ment ecosystem particularly industry API designers who have a large impact
on developers. By designing usable APIs [29] and easy to understand documen-
tation [58] they can help students and developers learn and use APIs correctly
which could result in building secure software.
Division of labour. Who is in charge of doing security at organisations has
long been a problem point with different units often thinking that security is the
job of another team [5]. A view shared by several of our participants. Though
such a tendency is considered to be a “key inhibitor toward secure software
development practices” [71, p. 164].
In the work place, security auditors are in charge of checking code for issues
which developers are then in charge of fixing [63]. However this system has some
downsides. First, auditing takes time during which developers work on other
projects and loose the working memory they had about particular code segments.
And second, fixing the code requires an understanding of the security issue in
order to properly address it, and as has been previously shown, developers have
difficulty interacting with security technologies like cryptography libraries due
to misunderstandings around how cryptography works [19].
In industry [9] it is necessary to create a security culture where basic security
is everyone’s responsibility and the security team is a component of that culture
rather than the only people who ‘do’ security. In education such a culture might
be facilitated by providing student with code samples that are secure by default
and by having them use code checking tools in IDEs that check for problems,
such as static analysis tools which teach them not only that they should look
for these issues, but also how.
Companies with high security standards make security as a commitment, do
not satisfy security because of complexity, and they follow strict formal devel-
opment and testing processes [31]. Universities can benefit these best practices
and tech CS students how to become developers that care about S&P.
Security as a requirement. There are several similarities between the stu-
dents’ views and general industry practices. Student developers’ treatment of
security as an implied requirement is in line with findings that security is often
treated as a non-functional feature in agile methods [13, 24], and that the re-
quirement is not explicitly stated [12, 15]. When asked to describe the features of
14 M. Tahaei et al.
the classroom discussion app, which had been intentionally chosen as an example
of a task with implicit S&P requirements, many students did not consider S&P
as an initial priority. For some students, this might be an artefact of their class-
room development experience, where they tend to work on well formed projects
that are unlikely to have security as an explicit requirement.
Poor and inconsistent understanding of S&P among CS students is likely
to cause conflicts between real and best practices in the software industry. For
example, when choosing a framework developers do not consider security as a
deciding factor which contradicts secure development best practices [8]. In align-
ment with other aspects of software development, there is a need to synchronise
the development approaches taught in the classroom with those used by indus-
try. That synchronisation needs to occur in both directions such that students
are taught industry best standards which they are then able to apply.
Internships. Internships are a way to engage students in the topic as well
as prepare them for future careers [14, 18]. Although they require investments
from industry [33] we believe that the shortage of S&P professionals [26] cannot
be solved without involving every player. Hence, we encourage industry to offer
more internships to CS students in S&P fields to improve the number of students
graduating with that type of experience.
6 Limitations
Our population includes only students at a single Russell Group university in the
UK. Even though our sample was diverse, it was not balanced for gender or se-
curity experience. Moreover, only two of our participants were native speakers of
English, and we might have obtained more finely differentiated views and opin-
ions if we had been able to interview each participant in their native language.
Since we conducted the study during summer vacation time, this resulted in a
participant pool biased towards Masters and PhD students, since undergraduate
students are not normally present at University in the summer months. Possibly
some of our potential participants were in their hometown and could not take
part in this study.
7 Future work
We plan to expand our study to other universities with a large scale survey to
investigate differences and similarities across curriculum, universities and coun-
tries. Extending outcomes of this research to industry and professional developers
and comparing results is also a path that could lead to valuable insights. Another
interesting avenue for future work is to investigate the impact of open source and
code reuse in system security. It also remains to question how developers trust
in others’ code and import code from different resources without knowing their
source and coder.
On the Security Mindsets of Computer Science Students 15
8 Conclusions
In this work we reported on a qualitative analysis of twenty semi-structured
interviews with CS students. We find that the attitudes of students match many
of those observed by other researchers looking at professional level developers.
Students have a range of hacker/attack mindsets, lack of experience with security
APIs, a mixed view of who is in charge of S&P in the software life cycle, and a
tendency to trust other peoples’ code as a convenient approach to rapidly build
software. We further give recommendations for both industry and academia to
improve software S&P ecosystem.
Acknowledgements
Thanks to all participants for their time and everyone associated with the
TULiPS Lab at the University of Edinburgh for helpful discussions and feedback.
We also thank the anonymous reviewers whose comments helped improve the
paper greatly. This work was sponsored in part by Microsoft Research through
its PhD Scholarship Programme.
References
1. Acar, Y., Backes, M., Fahl, S., Garfinkel, S., Kim, D., Mazurek, M.L., Stransky, C.:
Comparing the Usability of Cryptographic APIs. In: IEEE Symposium on Security
and Privacy. pp. 154–171 (2017)
2. Acar, Y., Backes, M., Fahl, S., Kim, D., Mazurek, M.L., Stransky, C.: You Get
Where You’re Looking for: The Impact of Information Sources on Code Security.
In: 2016 IEEE Symposium on Security and Privacy (SP). pp. 289–305 (2016)
3. Acar, Y., Fahl, S., Mazurek, M.L.: You are Not Your Developer, Either: A Re-
search Agenda for Usable Security and Privacy Research Beyond End Users. In:
Cybersecurity Development (SecDev), IEEE. pp. 3–8 (2016)
4. Acar, Y., Stransky, C., Wermke, D., Weir, C., Mazurek, M.L., Fahl, S.: Develop-
ers Need Support, Too: A Survey of Security Advice for Software Developers. In:
Cybersecurity Development (SecDev), 2017 IEEE. pp. 22–26 (2017)
5. Arce, I., Clark-Fisher, K., Daswani, N., DelGrosso, J., Dhillon, D., Kern, C.,
Kohno, T., Landwehr, C., McGraw, G., Schoenfield, B., et al.: Avoiding the Top 10
Software Security Design Flaws. Technical report, IEEE Computer Societys Center
for Secure Design (CSD) (2014)
6. Arzt, S., Nadi, S., Ali, K., Bodden, E., Erdweg, S., Mezini, M.: Towards Secure
Integration of Cryptographic Software. In: 2015 ACM International Symposium
on New Ideas, New Paradigms, and Reflections on Programming and Software
(Onward!). pp. 1–13 (2015)
7. Asgharpour, F., Liu, D., Camp, L.J.: Mental Models of Security Risks. In: Dietrich,
S., Dhamija, R. (eds.) Financial Cryptography and Data Security. pp. 367–377.
Springer Berlin Heidelberg (2007)
8. Assal, H., Chiasson, S.: Security in the Software Development Lifecycle. In: Four-
teenth Symposium on Usable Privacy and Security (SOUPS) (2018)
16 M. Tahaei et al.
9. Assal, H., Chiasson, S.: ‘Think secure from the beginning’: A Survey with Software
Developers. In: Proceedings of the 2019 CHI Conference on Human Factors in
Computing Systems (2019)
10. Azhar, M., Bhatia, S., Gagne, G., Kari, C., Maguire, J., Mountrouidou, X., Tudor,
L., Vosen, D., Yuen, T.T.: Securing the human: Broadening diversity in cybersecu-
rity. In: Proceedings of the 2019 ACM Conference on Innovation and Technology
in Computer Science Education. pp. 251–252 (2019)
11. Balebako, R., Cranor, L.: Improving App Privacy: Nudging App Developers to
Protect User Privacy. IEEE Security Privacy 12(4), 55–58 (2014)
12. Bartsch, S.: Practitioners’ Perspectives on Security in Agile Development. In: Pro-
ceedings of the 2011 Sixth International Conference on Availability, Reliability and
Security. pp. 479–484 (2011)
13. Bell, L., Brunton-Spall, M., Smith, R., Bird, J.: Agile Application Security: En-
abling Security in a Continuous Delivery Pipeline. O’Reilly Media (2017)
14. Binder, J.F., Baguley, T., Crook, C., Miller, F.: The academic value of internships:
Benefits across disciplines and student backgrounds. Contemporary Educational
Psychology 41, 73–82 (2015)
15. Bowen, J.P., Hinchey, M., Janicke, H., Ward, M.P., Zedan, H.: Formality, Agility,
Security, and Evolution in Software Development. IEEE Computer 47(10) (2014)
16. Cambazoglu, V., Thota, N.: Computer Science Students’ Perception of Computer
Network Security. In: Learning and Teaching in Computing and Engineering (LaT-
iCE). pp. 204–207. IEEE (2013)
17. Camp, L.J.: Mental models of privacy and security. IEEE Technology and Society
Magazine 28(3), 37–46 (2009)
18. Chillas, S., Marks, A., Galloway, L.: Learning to labour: an evaluation of internships
and employability in the ICT sector. New Technology, Work and Employment
30(1), 1–15 (2015)
19. Egele, M., Brumley, D., Fratantonio, Y., Kruegel, C.: An Empirical Study of
Cryptographic Misuse in Android Applications. In: Proceedings of the 2013 ACM
SIGSAC Conference on Computer and Communications Security. pp. 73–84 (2013)
20. The European parliament and the council of the European union: Gen-
eral Data Protection Regulation (GDPR) (2018), https://eur-lex.europa.eu/
legal-content/EN/TXT/PDF/?uri=CELEX:32016R0679 (Accessed August 2019)
21. Fahl, S., Harbach, M., Muders, T., Baumg¨artner, L., Freisleben, B., Smith, M.:
Why Eve and Mallory love Android: An analysis of Android SSL (in) security.
In: Proceedings of the 2012 ACM conference on Computer and communications
security. pp. 50–61 (2012)
22. Fahl, S., Harbach, M., Perl, H., Koetter, M., Smith, M.: Rethinking SSL Develop-
ment in an Appified World. In: Proceedings of the 2013 ACM SIGSAC conference
on Computer & Communications security. pp. 49–60 (2013)
23. Fischer, F., B¨ottinger, K., Xiao, H., Stransky, C., Acar, Y., Backes, M., Fahl,
S.: Stack Overflow Considered Harmful? The Impact of Copy Paste on Android
Application Security. In: 2017 IEEE Symposium on Security and Privacy (SP).
pp. 121–136 (2017)
24. Fitzgerald, B., Stol, K.J.: Continuous software engineering: A roadmap and agenda.
Journal of Systems and Software 123, 176–189 (2017)
25. Fulton, K.R., Gelles, R., McKay, A., Abdi, Y., Roberts, R., Mazurek, M.L.: The Ef-
fect of Entertainment Media on Mental Models of Computer Security. In: Fifteenth
Symposium on Usable Privacy and Security (SOUPS) (2019)
26. Furnell, S., Fischer, P., Finch, A.: Can’t get the staff? the growing need for cyber-
security skills. Computer Fraud & Security 2017(2), 5–10 (2017)
On the Security Mindsets of Computer Science Students 17
27. Georgiev, M., Iyengar, S., Jana, S., Anubhai, R., Boneh, D., Shmatikov, V.: The
Most Dangerous Code in the World: Validating SSL Certificates in Non-browser
Software. In: Proceedings of the 2012 ACM Conference on Computer and Com-
munications Security. pp. 38–49 (2012)
28. Google: Google diversity annual report (2018), http://diversity.google/
annual-report (Accessed August 2019)
29. Green, M., Smith, M.: Developers are Not the Enemy!: The Need for Usable Secu-
rity APIs. IEEE Security & Privacy 14(5), 40–46 (2016)
30. Hadar, I., Hasson, T., Ayalon, O., Toch, E., Birnhack, M., Sherman, S., Balissa,
A.: Privacy by designers: software developers’ privacy mindset. Empirical Software
Engineering 23(1), 259–289 (2018)
31. Haney, J.M., Theofanos, M., Acar, Y., Prettyman, S.S.: “We make it a big deal
in the company”: Security Mindsets in Organizations that Develop Cryptographic
Products. In: Fourteenth Symposium on Usable Privacy and Security (SOUPS)
(2018)
32. Hissam, S.A., Plakosh, D., Weinstock, C.: Trust and vulnerability in open source
software. IEE Proceedings-Software 149(1), 47–51 (2002)
33. Hoffman, L., Burley, D., Toregas, C.: Holistically Building the Cybersecurity Work-
force. IEEE Security & Privacy 10(2), 33–39 (2012)
34. Iacono, L.L., Gorski, P.L.: I Do and I Understand. Not Yet True for Security
APIs. So Sad. In: Proc. of the 2nd European Workshop on Usable Security, ser.
EuroUSEC (2017)
35. Indela, S., Kulkarni, M., Nayak, K., Dumitras, T.: Toward Semantic Cryptography
APIs. In: Cybersecurity Development (SecDev), IEEE. pp. 9–14 (2016)
36. Information Commissioner’s Office: Investigation into the
use of data analytics in political campaigns (2018),
https://ico.org.uk/media/action-weve-taken/2259371/
investigation-into-data-analytics- for-political- purposes-update.pdf
(Accessed August 2019)
37. Jones, K., Siami Namin, A., Armstrong, M.: What Should Cybersecurity Students
Learn in School?: Results from Interviews with Cyber Professionals. In: Proceed-
ings of the 2017 ACM SIGCSE Technical Symposium on Computer Science Edu-
cation. pp. 711–711 (2017)
38. Kernighan, B.W., Ritchie, D.M.: The C Programming Language. Prentice Hall
(2006)
39. Lazar, D., Chen, H., Wang, X., Zeldovich, N.: Why Does Cryptographic Software
Fail?: A Case Study and Open Problems. In: Proceedings of 5th Asia-Pacific Work-
shop on Systems. p. 7. ACM (2014)
40. Lazar, J., Feng, J.H., Hochheiser, H.: Research Methods in Human-Computer In-
teraction. Morgan Kaufmann (2017)
41. Maier, J., Padmos, A., Bargh, M.S., W¨orndl, W.: Influence of Mental Models on
the Design of Cyber Security Dashboards. In: Proceedings of the 12th International
Joint Conference on Computer Vision, Imaging and Computer Graphics Theory
and Applications - Volume 3: IVAPP, (VISIGRAPP). pp. 128–139 (2017)
42. Mansfield-Devine, S.: The Ashley Madison affair. Network Security 2015(9), 8–16
(2015)
43. M¨arki, H., Maas, M., Kauer-Franz, M., Oberle, M.: Increasing Software Security
by Using Mental Models. In: Advances in Human Factors in Cybersecurity, pp.
347–359. Springer (2016)
18 M. Tahaei et al.
44. Mazurek, M.L., Arsenault, J.P., Bresee, J., Gupta, N., Ion, I., Johns, C., Lee, D.,
Liang, Y., Olsen, J., Salmon, B., Shay, R., Vaniea, K., Bauer, L., Cranor, L.F.,
Ganger, G.R., Reiter, M.K.: Access Control for Home Data Sharing: Attitudes,
Needs and Practices. In: Proceedings of the SIGCHI Conference on Human Factors
in Computing Systems. pp. 645–654 (2010)
45. Miles, M., Huberman, M.: Qualitative Data Analysis: A Methods Sourcebook. Sage
(1994)
46. Munro, K.: Hacking kettles & extracting plain text WPA PSKs.
Yes really! (2015), https://www.pentestpartners.com/security-blog/
hacking-kettles-extracting-plain- text-wpa- psks-yes-really (Accessed
August 2019)
47. Nadi, S., Kr¨uger, S., Mezini, M., Bodden, E.: Jumping Through Hoops: Why Do
Java Developers Struggle with Cryptography APIs? In: Proceedings of the 38th
International Conference on Software Engineering. pp. 935–946. ACM (2016)
48. Nielson, S.J.: PLAYGROUND: preparing students for the cyber battleground.
Computer Science Education 26(4), 255–276 (2016)
49. Oliveira, D., Rosenthal, M., Morin, N., Yeh, K.C., Cappos, J., Zhuang, Y.: It’s
the Psychology Stupid: How Heuristics Explain Software Vulnerabilities and How
Priming Can Illuminate Developer’s Blind Spots. In: Proceedings of the 30th An-
nual Computer Security Applications Conference. pp. 296–305 (2014)
50. Oltrogge, M., Derr, E., Stransky, C., Acar, Y., Fahl, S., Rossow, C., Pellegrino, G.,
Bugiel, S., Backes, M.: The Rise of the Citizen Developer: Assessing the Security
Impact of Online App Generators. In: 2018 IEEE Symposium on Security and
Privacy (SP). pp. 634–647 (2018)
51. OWASP: Top 10 Most Critical Web Application Security Risks. Tech. rep., The
OWASP Foundation (2017)
52. Patnaik, N., Hallett, J., Rashid, A.: Usability Smells: An Analysis of Developers’
Struggle With Crypto Libraries. In: Fifteenth Symposium on Usable Privacy and
Security (SOUPS) (2019)
53. Pieczul, O., Foley, S., Zurko, M.E.: Developer-centered Security and the Symmetry
of Ignorance. In: Proceedings of the 2017 New Security Paradigms Workshop. pp.
46–56. ACM (2017)
54. Poller, A., Kocksch, L., T¨urpe, S., Epp, F.A., Kinder-Kurlanda, K.: Can Secu-
rity Become a Routine?: A Study of Organizational Change in an Agile Software
Development Group. In: Proceedings of the 2017 ACM Conference on Computer
Supported Cooperative Work and Social Computing. pp. 2489–2503 (2017)
55. Radermacher, A., Walia, G.: Gaps Between Industry Expectations and the Abilities
of Graduates. In: Proceeding of the 44th ACM Technical Symposium on Computer
Science Education. pp. 525–530 (2013)
56. Radermacher, A., Walia, G., Knudson, D.: Investigating the Skill Gap Between
Graduating Students and Industry Expectations. In: Companion Proceedings of
the 36th International Conference on Software Engineering. pp. 291–300. ACM
(2014)
57. Renaud, K., Volkamer, M., Renkema-Padmos, A.: Why Doesnt Jane Protect Her
Privacy? In: International Symposium on Privacy Enhancing Technologies Sym-
posium. pp. 244–262. Springer (2014)
58. Robillard, M.P., Deline, R.: A field study of API learning obstacles. Empirical
Software Engineering 16(6), 703–732 (2011)
59. Salda˜na, J.: The Coding Manual for Qualitative Researchers. Sage (2015)
60. StackOverflow: Developer Survey Results (2019), https://insights.
stackoverflow.com/survey/2019 (Accessed August 2019)
On the Security Mindsets of Computer Science Students 19
61. Sudol, L.A., Jaspan, C.: Analyzing the Strength of Undergraduate Misconceptions
About Software Engineering. In: Proceedings of the Sixth International Workshop
on Computing Education Research. pp. 31–40 (2010)
62. Tabassum, M., Watson, S., Chu, B., Lipford, H.R.: Evaluating Two Methods for
Integrating Secure Programming Education. In: Proceedings of the 49th ACM
Technical Symposium on Computer Science Education. pp. 390–395 (2018)
63. Thomas, T.W., Tabassum, M., Chu, B., Lipford, H.: Security During Application
Development: an Application Security Expert Perspective. In: Proceedings of the
2018 CHI Conference on Human Factors in Computing Systems. p. 262 (2018)
64. Ukrop, M., Matyas, V.: Why Johnny the Developer Can’t Work with Public Key
Certificates. In: Cryptographers’ Track at the RSA Conference. Springer (2018)
65. Vaniea, K.E., Rader, E., Wash, R.: Betrayed by Updates: How Negative Experi-
ences Affect Future Security. In: Proceedings of the SIGCHI Conference on Human
Factors in Computing Systems (2014)
66. Vox: The Cambridge Analytica Facebook scandal (2018), https://www.vox.com/
2018/4/10/17207394 (Accessed August 2019)
67. Wash, R.: Folk Models of Home Computer Security. In: Proceedings of the Sixth
Symposium on Usable Privacy and Security (SOUPS) (2010)
68. Weir, C., Rashid, A., Noble, J.: How to Improve the Security Skills of Mobile App
Developers: Comparing and Contrasting Expert Views. In: Twelfth Symposium on
Usable Privacy and Security (SOUPS) (2016)
69. Whitney, M., Lipford-Richter, H., Chu, B., Zhu, J.: Embedding Secure Coding
Instruction into the IDE: A Field Study in an Advanced CS Course. In: Proceedings
of the 46th ACM Technical Symposium on Computer Science Education (2015)
70. Wurster, G., van Oorschot, P.C.: The Developer is the Enemy. In: Proceedings of
the 2008 New Security Paradigms Workshop. pp. 89–97. ACM (2009)
71. Xie, J., Lipford, H.R., Chu, B.: Why do programmers make security errors?
In: 2011 IEEE Symposium on Visual Languages and Human-Centric Computing
(VL/HCC). pp. 161–164 (2011)
72. Zou, Y., Mhaidli, A.H., McCall, A., Schaub, F.: “I’ve Got Nothing to Lose”: Con-
sumers Risk Perceptions and Protective Actions after the Equifax Data Breach.
In: Fourteenth Symposium on Usable Privacy and Security (SOUPS) (2018)
20 M. Tahaei et al.
Appendix: interview script
1. Background
•Can you tell me about yourself? Your academic and professional back-
ground? •Can you tell me about your dream job?
2. App scenario
Let’s say you were asked to create a new group discussion app for in-class
discussions. •Free list: what features would you consider in this app? •
Here is a red pen. Can you circle the features that are security and privacy
related? Or where you might have to consider security and privacy when
building them? •Why these ones? •Who is most likely to try and attack
this system? What are they likely going to try and do?
3. Threats and attacks
•Can you tell me who hackers are, in your opinion? •Their intentions? •
What are hackers trying to get? •Their background?
4. Responsibility attribution
•Who is responsible for providing security and privacy to end users?
5. Prior coding experiences
•Tell me about the last piece of software you wrote. •Did you consider
security while building your project? If not this one, any other projects?
•Can you tell me an example of an API/library? Can you give me some
experiences you have had with them? Any experience with security APIs in
particular? •What was good about it? Why did you like it? •What was
confusing about it?
6. Personal security/privacy practices
Now we are going to switch to talking about how you handle security and
privacy personally as an end user. •Free list: What words and concepts do
you associate with computer security? •Can you give me an example of
a good computer security practice? What about something you have done
yourself? •Have you ever experienced a security or privacy compromise such
as getting a virus on your computer, losing your password, having an email
sent from your account, or loss of data about you? •How did you find out
about the issue? •How did you correct it? •What did you learn from the
experience? •Can you tell me some about the experiences you have had
with passwords?
7. Background and demographics
•How old are you? •What is your degree title? •Which year of the program
are you in? •What programming languages do you know? •What program-
ming courses have you taken? •What security courses have you taken? •
What is your nationality? •Where did you study your undergraduate, Mas-
ters, or other degrees?
