Embedded Files
Software Accessibility - Contemporary Practices
Software Accessibility - Contemporary Practices
In #A11yDev, an qualitative study of Tweets by demographic affected by and advocating for software accessibility, we explore current practices in the space. We look into the state of software accessibility from 3 viewpoints:
Process: how accessibility is integrated in the different processes of developing a software i.e. requirements, design, development and testing.
Profession: how accessibility is considered in various professional aspects, whether to train staff, organize teams, or legally ensuring its inclusion.
People: how accessibility taught to learners and professionals, and discussed in professional and online spheres.
Software Accessibility - The Issue of Over-accessible Android Elements
Software Accessibility - The Issue of Over-accessible Android Elements
With Oversight, we investigate the phenomena of over-accessible elements on Android apps. While inaccessibility in most research and tools in the field of software accessibility is refered to a lack of accessibility to features and elements to users of Assistive Technologies (AT) like screen readers, magnifiers and so on, we look into instances where, with ATs, one can gain access to features that are intended to be locked, for instance, content behind a pay wall.
The Relation between Developer Sentiment and Developer Performance
The Relation between Developer Sentiment and Developer Performance
Sentiment or emotional state of software developers can be derived from online collaborative artifacts where they contribute to the software project. This research observes whether these emotional patterns can be correlated with their performance, or more specifically, whether they introduce bugs to the system. As part of my Masters thesis, I investigated three collaborative artifacts from GitHub:
Commits - where the developer contributes (SANER '20 paper)
Pull Requests - where reviews and conversation influence the contribution (APSEC '19 paper)
Issues - where contributors converse to solve problems related to the software
The Effect of COVID-19 Pandemic on the Productivity of Professionals in the IT Sector of Bangladesh
The Effect of COVID-19 Pandemic on the Productivity of Professionals in the IT Sector of Bangladesh
Due to the lockdown caused by the COVID-19 pandemic, organizations needed to move their workforce from offices to workers' homes. The IT sector of Bangladesh also had to adopt the Work From Home (WFH) policy to continue their operations. In this empirical study, more than 1000 professionals in the IT sector have been surveyed to understand how their productivity and other aspects of their work have been affected by remote work.
The work can be viewed as:
A visualization of the primary findings
A preprint describing the methods and primary findings
The Characteristics and Applicability of Topic Modeling in Predicting Lifetime of GitHub Issues
The Characteristics and Applicability of Topic Modeling in Predicting Lifetime of GitHub Issues
Software projects continuously generate large quantities of text for collaboration and communication among developers. This research delves in the characteristics of topic modeling - the process of summarizing and categorizing the corpora based on words and tokens that characterize them - in such collaborative artifacts, specifically GitHub Issues. To investigate its applicability, the topics are used for predicting the resolution time of Issues.
The Evolutionary Properties of Fix-Inducing Changes
The Evolutionary Properties of Fix-Inducing Changes
Fix-Inducing Changes (FICs) refer to Commits that introduce bugs to the system. It is used as a historical reference in the lifetime of a project to find instances of developers writing buggy code. This study investigates how FICs evolve during the project's lifetime and observes their relation with contribution intervals and system complexity.
Detection and Refactoring of Message Chains in Java Code
Detection and Refactoring of Message Chains in Java Code
Message Chains are a specific code smell that causes readability and maintenance issues in software code. This study builds a tool - Chain Breaker - that detects Message Chains in Java code and provides refactoring suggestions in the form of modified code. The study also
A Machine Learning-based Inference and Analysis of Crop Production based on Climate Parameters in Bangladesh
A Machine Learning-based Inference and Analysis of Crop Production based on Climate Parameters in Bangladesh
Bangladesh is an agricultural country; the production of basic crops like rice, jute and more defines the economy of millions of households and the nation as a whole. Predicting the production of crops can therefore directly help the population. This study uses climate parameters, an influential aspect for crop production in a tropical country like Bangladesh, to predict yearly yields of crops. Various machine learning methods are used to derive the most suited prediction model and infer the varying impact of different climate parameters.
Report abuse