Uma investigação sobre o impacto da dívida técnica em atividades de desenvolvimento de software

Abstract

The term Technical Debt (TD) is used to describe the increase in productivity and the reduction of long-term costs resulting from shortcuts taken in the development and evolution of the software. It is common for a software project to incur debt during its development. The incidence of TD can contribute to the success of the project more quickly, however, if not managed efficiently, the accumulated debt yields associated risks, among others, unexpected delays in the evolution of the system and difficulty in meeting the quality criteria. defined for the project. TD management can implement actions to find a balance between the benefits of incurring it and the effects of its presence. Managing TD involves making decisions regarding whether an item of debt should be repaid and the most appropriate time to do so. It also includes preventive actions such as avoiding debt items in early stages of software development can reduce the chances of these items impacting development activities later on. Research on TD management related to understanding the causes that lead development teams to incur debt items in their projects and their effects were also done. Knowing the causes of TD can support development teams in defining and creating preventive actions for TD. Having information about the effects of TD can help in prioritizing TD items to pay, supporting more accurate analysis, impact analysis, and the identification of corrective actions to minimize potential negative consequences for the project. Until now, most studies on TD have focused on discussions related to TD with coding problems. This is a worrying scenario because other types of debt (documentation, requirements, tests, human factors) can also have impactful, or even worse, consequences on projects. This work addresses the growing need to expand TD research to other areas of software development. For this, six elements related to the management of TD were analyzed: causes, effects, prevention practices and reasons for non-prevention, reimbursement practices and reasons for non-reimbursement. A subset of the data collected by the InsighTD project, a family of globally distributed research into the causes, effects and management of TD, was used. The subset consists of data from six replications, totaling 653 practitioner software responses. Overall, the results indicated that coding issues related to the causes, effects, prevention, non-prevention, payment and non-payment of TD are only a small part of the concerns that professionals face in the presence of TD. In fact, DT is very commonly found in other software development issues