Abstract
Today, computer software is a critical part of social and technological infrastructure.
However, computer systems are not always reliable and usable. For organizations and designers,
these flaws increase the chance of system failure, financial losses, or reputational damage. For
users, especially disabled people, unreliable or unusable applications can unfairly restrict their
autonomy, career opportunities, and access to social services. While both papers address this
fundamental problem, the topics of each paper are different. The technical report analyzes a
specific internship project, from the perspective of software engineers and managers, with an
emphasis on financial and operational risks. Meanwhile, the research paper claims that including
disabled people is necessary to create computer systems that are available, reliable, and usable
for those communities.
The technical report recounts my experience as a software engineering intern at Emerson,
where I stabilized a web application that will replace a legacy inventory system. At the outset,
there were several open questions: what issues does the application have, and how can they be
prevented in the future? It quickly became clear that the app was unreliable and hard to use.
Numerous reliability flaws, including poor security and fault-tolerance, created long-term
financial and operational risks for Emerson. Throughout the project, I painstakingly documented
and corrected these flaws. The simplicity and repetition of the errors suggests that the underlying
problem was a lack of communication, which could have been avoided through robust standards,
documentation, and analytics. Once added to the application, these measures eased the workload
of developers, reduced the creation of new errors, and enabled better knowledge transfer between
teams. Although many issues remain, this report concludes that similar errors could be prevented
by early adoption of these tools and techniques.
Meanwhile, the research paper asks: what makes some accessibility initiatives more
successful than others? Currently, most computer systems have accessibility errors, which create
barriers of access that discriminate against and limit opportunities for people with disabilities. To
find ethical and effective strategies to improve accessibility, this paper compares case studies
across four different mediums: assistive technology, software design, computer science
education, and government regulation. For each, the analysis measures the technology’s (or
process’) success based on the interpretations of several complementary frameworks, including
digital divides, social construction of technology, care ethics, cripistemology, fluid technology,
and universal design. This reveals that accessibility initiatives are most successful when disabled
people are involved throughout the design process. Because disabled knowledge and experiences
are unique, they should be held paramount. Furthermore, this paper raises concerns about the use
of artificial intelligence in accessibility design because it might undermine, misrepresent, or
mislead disabled people and their goals.
While these two papers only address small pieces of the much larger problem, they
successfully interrogate current software design practices and make meaningful suggestions to
improve the reliability and usability of computer software, both technologically and socially.
Since neither result is truly groundbreaking, future researchers should use these ideas to inspire
new methods of software design. Furthermore, the research paper relies heavily on examples
about visual impairments and from western countries, so the values of other communities are
underrepresented in this portfolio and should receive more attention in the future. Nevertheless,
this research suggests several promising new research questions: how do disabled people use and
perceive artificial intelligence in their daily lives? Why do some mediums of accessibility have
more disabled involvement than others? Have accessibility design practices changed with the
introduction of artificial intelligence? By calling future researchers to study these questions, this
portfolio hopes that we may one day create computer systems that are truly reliable and usable
for all.
