Technical Debt in Practice

Developers of deep learning applications (shortened as application developers) commonly use deep learning frameworks in their projects. However, due to time pressure, market competition, and cost reduction, developers of deep learning frameworks (shortened as framework developers) often have to sacrifice software quality to satisfy a shorter completion time. This practice leads to technical debt in deep learning frameworks, which results in the increasing burden to both the application developers and the framework developers in future development.In this paper, we analyze the comments indicating technical debt (self-admitted technical debt) in 7 of the most popular open-source deep learning frameworks. Although framework developers are aware of such technical debt, typically the application developers are not. We find that: 1) there is a significant number of technical debt in all the studied deep learning frameworks. 2) there is design debt, defect debt, documentation debt, test debt, requirement debt, compatibility debt, and algorithm debt in deep learning frameworks. 3) the majority of the technical debt in deep learning framework is design debt (24.07% - 65.27%), followed by requirement debt (7.09% - 31.48%) and algorithm debt (5.62% - 20.67%). In some projects, compatibility debt accounts for more than 10%. These findings illustrate that technical debt is common in deep learning frameworks, and many types of technical debt also impact the deep learning applications. Based on our findings, we highlight future research directions and provide recommendations for practitioners.

Cross-disciplinary and cross-life-cycle-phase Technical Debt in automated Production Systems: two industrial case studies and a survey

Technical Debt tracking: Current state of practice

JavaScript (JS) is one of the most popular programming languages for developing client-side applications mainly due to allowing the adoption of different programming styles, not having strict syntax rules, and supporting a plethora of frameworks. The flexibility that the language provides may accelerate the development of application, but also pose threats to the quality of the final software product, e.g., introducing Technical Debt (TD). TD reflects the additional cost of software maintenance activities to implement new features, occurring due to poorly developed solutions. Being able to forecast the levels of TD in the future can be extremely valuable in managing TD, since it can contribute to informed decision making when designating future repayments and refactoring budget among a company's projects. Despite the popularity of JS and the undoubtful benefits of accurate TD forecasting, in the literature, there is available only a limited number of tools and methodologies that are able to: (a) forecast TD during software evolution, (b) provide a ground-truth TD quantifications to train forecasting, since TD tools that are available are based on different rulesets and none is recognized as a state-of-the-art solution, (c) take into consideration the language-specific characteristics of JS. As a main contribution for this study, we propose a methodology (along with a supporting tool) that supports the aforementioned goals based on the Backward Stepwise Regression and Auto-Regressive Integrated Moving Average (ARIMA). We evaluate the proposed approach through a case study on 19,636 releases of 105 open-source applications. The results point out that: (a) the proposed model can lead to an accurate prediction of TD, and (b) the Number of appearances of the “new” and “eval” keyword along with the number of “anonymous” and “arrow” functions are among the features of JavaScript language that are related to high levels of TD.

Context: Technical Debt (TD) is a metaphor that refers to short-term solutions in software development that may affect the cost to the software development life cycle. Objective: To explore and understand TDrelated to the software industry as well as an overview on the current state of TD research. Forty-three TD empirical studies were collected for classification and analyzation. Goals: Classify TD types, find the indicators used to detect TD, find the estimators used to quantify the TD, evaluate how researchers investigate TD. Method: By performing a systematic mapping study to identify and analyze the TD empirical studies which published between 2014 and 2017. Results: We present the most common indicators and evaluators to identify and evaluate the TD, and we gathered thirteen types of TD. We showed some ways to investigate the TD, and used tools in the selected studies. Conclusion: The outcome of our systematic mapping study can help researchers to identify interestand future in TD.

Technical debt and system architecture: The impact of coupling on defect-related activity

Chapter 2 - Measuring and Monitoring Technical Debt

Technical Debt is a term coined by Ward Cunningham to denote the amount of rework required to put a piece of software into that state which it should have had from the beginning. The term outdates the agile revolution. Technical debt can occur using any development approach, but since agile development has become wide spread the notion of “technical debt” has gained much more attention. That is because if agile development is not done properly, technical debt accrues very fast and threatens to strangle the project. In this paper the author describes how technical debt comes to being, how it can be measured and what can be done to prevent it in agile development. The emphasis of the paper is on how to prevent it. What measures are necessary in an agile development project to keep technical debt from accruing over time? The paper lists out a number of measures which can be taken – organizational, procedural and technical. Automated tools play an important role in the struggle against technical debt. Samples are given how a tool can be helpful in identifying and removing problems before they get out of hand. Also important is an external auditing agency which monitors projects in progress with the aide of automated tools. In the end a case study is presented which illustrates the monitoring of technical debt within an agile development and what counter measures are required to stop it.KeywordsTechnical debtcode qualityproduct documentationsoftware defectsrefactoringsoftware metricsagile team organizationagile testing

One major problem in using static analyzers to manage, monitor, control, and reason about technical debt is that industrial projects have a huge amount of technical debt which reflects hundreds of quality rule violations (e.g., high complex module or low comment density). Moreover the negative impact of violating quality rules (i.e., technical debt interest) may vary across rules or even across contexts. Thus, without a context-specific validation and prioritization of quality rules, developers cannot effectively manage technical debt. This paper reports on a case study aimed at exploring the interest associated with violating quality rules; i.e., we investigate if and which quality rules are important for software developers. Our empirical method consists of a survey and a quantitative analysis of the historical data of a CMMI Level 5 software company. The main result of the quantitative analysis is that classes violating several quality rules are five times more defect prone than classes not violating any rule. The main result of the survey is that some rules are perceived by developers as more important than others; however, there is no false positive (i.e., incorrect rule or null interest). These results pave the way to a better practical use of quality rules to manage technical debt and describe new research directions for building a scientific foundation to the technical debt metaphor.

Due to cost and time constraints, software quality is often neglected in the evolution and adaptation of software. Thus, maintainability suffers, maintenance costs rise, and the development takes longer. These effects are referred to as “technical debt”. The challenge for project managers is to find a balance when using the given budget and schedule, either by reducing technical debt or by adding technical features. This balance is needed to keep time to market for current product releases short and future maintenance costs at an acceptable level. Method: The project ProDebt aimed at developing an innovative methodology and a software tool to support the strategic planning of technical debt in the context of agile software development. In this project, we created quality models and collected corresponding measurement data for two case studies in two different companies. Altogether, the two case studies contributed 5–6 years of data, from the end of 2011, resp. mid-2012, until today. Using measurement and effort data, we trained a machine-learning model to predict productivity based on measurement data—representing the technical debt of a file at a given point in time. Result: We developed a prototype and a prediction model for forecasting potential savings based on proposed refactorings of key drivers of technical debt identified by the model. In this paper, we present the approach and the experiences made during model development.

Technical debt (TD) is a concept used to describe a sub-optimal solution of a software artifact that negatively affects its comprehensibility, extendability and maintainability. As such, TD adversely affects the costs or quality associated with the artifact, which is also called interest. TD has through research been identified in all types of software artifacts, from architectural design to automated tests (Testware). However, research into testware technical debt (TTD) is limited and primarily focused on testing on lower level of system abstraction, i.e. unit-and integration tests, leaving a need for more TTD research on GUI-based testing. In this study we explore this gap in knowledge through an industrial case study at a Swedish avionics software development company. Four repositories are studied for the presence of TTD using expert interviews, semi-automated document analysis and automatic metric analysis. Results of the study provide initial support that the concept of TTD is applicable to GUI-based testware and show the presence of both TD items unique to GUI-based testware and items common to software. The implications of these results are that engineering best practices must be established for GUI-based testware to minimize TD interest.

Technical debt are those convenience tasks that developers perform to obtain short-term benefits but that in the future can produce activities that are difficult to perform or more costly. The impact of technical debt is perceived in all project resources, becoming a present and future problem in software development, affecting the quality of the software. This study uses a qualitative analysis, through interviews with a software development team, explores the knowledge and perception of technical debt, the practices used to manage it, the factors that influence its appearance and accumulation, the impact that development suffers, as well as the team and the maintainability of the software, as a result, ideas and solutions are proposed to address it effectively. The study was carried out with a qualitative investigative approach, which was based on the study of a single case with which the experiences and practices in the perception and management of technical debt are investigated. The data obtained reveal that technical debt affects the normal development of the development team and that it is present in the activities carried out by the team. Although the development team has a perception of the debt, it does not address it directly, and it is necessary to implement good practices to identify it and integrate it into the software development process. Addressing technical debt with the implementation of a formal process to manage it will ensure that scheduled time is controlled and resources are optimized, which will result in a culture of good practices that will ensure product quality, improving team performance and guaranteeing software maintainability

Costs and obstacles encountered in technical debt management – A case study

Technical debt (TD) is a metaphor that reflects the technical compromises that sacrifice the long-term health of a software product to achieve short term benefit. It is important to manage TD to avoid software degradation. In large-scale distributed projects, technical debt management (TDM) becomes more complex and challenging. There is a lack of empirical studies on the TD accumulation in large-scale distributed projects. Then, to address this gap, we conducted a case study in Ericsson (a European Telecom Company) to identify the relationship between TD accumulation and factors such as task complexity, lead time, total of developers, and task scaling. We used data from 33 projects extracted from managerial documents to conduct a regression analysis. We also conducted interviews with seniors developers of the team to interpret the results. We found out that Task Complexity has a strong relationship (p-value = 5.69 × 10-5) with Technical Debt, while Global Distance was mentioned by the interviewees as a relevant factor for TD accumulation (although not statistically significant in our regression analysis). Practitioners should consider avoiding complex/big tasks, breaking down big tasks into small ones (if possible). We also plan to analyze other projects in this company to confirm our findings further. © 2021 IEEE.

BackgroundSoftware companies need to manage and refactor Technical Debt issues. Therefore, it is necessary to understand if and when refactoring of Technical Debt should be prioritized with respect to developing features or fixing bugs. ObjectiveThe goal of this study is to investigate the existing body of knowledge in software engineering to understand what Technical Debt prioritization approaches have been proposed in research and industry. MethodWe conducted a Systematic Literature Review of 557 unique papers published until 2020, following a consolidated methodology applied in software engineering. We included 44 primary studies. ResultsDifferent approaches have been proposed for Technical Debt prioritization, all having different goals and proposing optimization regarding different criteria. The proposed measures capture only a small part of the plethora of factors used to prioritize Technical Debt qualitatively in practice. We present an impact map of such factors. However, there is a lack of empirical and validated set of tools. ConclusionWe observed that Technical Debt prioritization research is preliminary and there is no consensus on what the important factors are and how to measure them. Consequently, we cannot consider current research conclusive. In this paper, we therefore outline different directions for necessary future investigations.