Vard Antinyan

PhD candidate in the Software Engineering division at the Department of Computer Science and Engineering.

About Me:

My name is Vard Antinyan. I am PhD candidate in Software Center, Chalmers | University of Gothenburg. My research topic includes software quality, software complexity, technical debt, risk management, and project management. I also extensively use the theory of measurement and cognitive psychology in order to investigate the above subjects and solve practical problems in software development. 

My Research Activities:

In 2013 we conducted research in Ericsson and Volvo Trucks and created a method for automated assessment of technical risks in the companies. The method enabled automated identification and assessment of risky code areas on daily basis. This permitted developers to focus on such code areas that are most likely error-prone or difficult-to-maintain. The automation reduced the effort of manual assessment, and the accuracy of assessment permitted concentrating improvement and testing effort on risky areas.
In 2014 We conducted research in Volvo Trucks, Saab, Volvo Cars, and Grundfos, and created a method for automated internal quality assessment of textual requirements. Because in large companies the number of requirements can be several thousands, it is a tiresome task for developers to conduct manual reviews. We created measures for internal quality and developed a method based on these measures, which could automatically assess the quality of thousands of requirements. The method reduced the review time from weeks to seconds, saving hundreds of hours for the companies. 
In 2015 We conducted research in seven companies for better understanding and management of software complexity. The results imply that the current complexity measures are deficient, therefore, they are not used in practice. Moreover, there are such complexity triggers that increase the complexity of code substantially, therefore, they should be measured and managed. If not controlled, however, the code maintenance effort can increase by a factor of 2.5 - 5.
In 2016 we conducted research in Ericsson and found that test coverage is an inadequate strategy for assuring defect-free software. Increasing coverage does not decrease the defects. Therefore, instead of sheer fulfillment of coverage, smarter ways of testing should be applied.
In 2017 we conducted a more profound research on code complexity, and found that complexity should be managed in accordance with human cognitive capabilities. Guidelines for better complexity management was proposed. 

My Research Principles:

To me there are three perspectives in software engineering research: 1) philosophy of research: comprehending and advancing the philosophy of research for establishing a common lexicon for software engineering community, 2) theory of research: creating powerful methods for making more pleasant working environment and more effective working strategies for software engineers, 3) applicability of research: validating and adapting these methods in industry for enhancing the quality of the products and engineers’ working standards. 
I have learned a lot while conducting research in collaborating companies. Listening engineers' problems and designing practical solutions has been my job. As a researcher I am not only interested in finding the root causes of the problems and presenting them to companies, but more, I am interested in proposing solutions and applying them in companies for solving the problems. It is not often that I succeed in designing good solutions, however, I also recognize that a good solution cannot come with a one-step research process. I prefer to conduct stepwise cyclical process, when the developed solution can be defined, refined, and redefined until the end goal is achieved. Because modern software development advances towards continuous software delivery, proactive assessment is an indispensable activity. And because the modern software products can be huge, the automation of such assessment is a key factor. Therefore, I regard the automation and proactiv​eness as two important pillars when intending to design a successful assessment methods.​

​This research is funded by Software Center, Chalmers | University of Gothenburg. The collaborating companies with my research project are: Ericsson, Volvo Group Truck Technology, Volvo Car Group, Saab Defense, and Grundfors. Additional information about my research and Software Center can be found here.

Published: Tue 25 Feb 2014. Modified: Mon 03 Jul 2017