The increasing pervasiveness of software-intensive systems requires involving domain experts more directly in technological development. Visual models, expressed in semi-formal notations, can act as shared artefacts that support communication and collaboration between developers and domain experts. However, modelling with semi-formal notations can be challenging for novice modellers. This study presents an AI-infused, web-based modelling tool designed to support users in formalising domain knowledge without requiring advanced modelling skills. The tool features a block-based, domain-specific language that automatically transforms user-generated structures into semi-formal diagrams. AI-based functionalities include a diagram reader, contextual hints, natural-language instructions, and interaction logging. We evaluated the tool through a Wizard of Oz experiment with agronomists in digital agriculture, where participants completed an exploratory modelling task while interacting with AI assistance. Results reveal three key design implications: (i) adaptive AI support accommodating diverse modelling strategies, (ii) concise, actionable guidance delivered at moments of difficulty, and (iii) practice-oriented assistance that preserves user agency and supports learning-by-doing.

The engineering of industrial systems, particularly in safety-critical domains such as railways, demands rigorous verification and validation processes to ensure system dependability. Formal methods have emerged as powerful tools to complement traditional software engineering practices. In the railway sector, which increasingly relies on complex, distributed, and cyber-physical control systems, formal methods have demonstrated particular value for many decades now. In this article, we provide a retrospective overview of the application of formal methods and tools in the railway domain, with emphasis on two prominent verification approaches and one frequently verified railway system: modeling and validation with the B method and tools and formal verification of interlocking systems by model checking. We explore their role in the design and development of key railway systems, highlighting both academic research and industrial success stories, as witnessed by international projects and initiatives. We conclude with an outlook on the potential of integrating AI and formal methods to enhance the efficiency of next-generation railway systems.

Developing large-scale industrial systems requires high-quality requirements to avoid costly rework and project delays. However, linguistic ambiguities in natural language (NL) requirements have been a long-standing challenge, often introducing misinterpretations and inconsistencies that propagate throughout the development lifecycle. Such ambiguous NL requirements necessitate early detection and well-reasoned explanations to clarify and prevent further misunderstandings among stakeholders. While solutions have been developed to detect ambiguities in NL requirements, the advent of generative large language models (LLMs) offers new avenues for explanation-augmented requirements ambiguity detection. This paper empirically investigates LLMs for ambiguity detection and explanation in real-world industrial requirements by adopting an in-context learning paradigm. Our results from three industrial datasets show that LLMs achieve a 20.2% average performance increase in classifying ambiguous requirements when prompted with ten relevant in-context demonstrations (10 -shot), compared to no demonstrations (0 -shot). Additionally, we conducted human evaluations of the LLM-generated outputs with eight industry experts along four dimensions-naturalness, adequacy, usefulness and relevance-to gain practical insights. The results show an average rating of 3.84 out of 5 across evaluation criteria, indicating that the approach is effective in providing supporting explanations for requirement ambiguities.

Indoor navigation presents unique challenges due to complex layouts and the unavailability of GNSS signals. Existing solutions often struggle with contextual adaptation, and typically require dedicated hardware. In this work, we explore the potential of a Large Language Model (LLM), i.e., ChatGPT, to generate natural, context-aware navigation instructions from indoor map images. We design and evaluate test cases across different real-world environments, analyzing the effectiveness of LLMs in interpreting spatial layouts, handling user constraints, and planning efficient routes. Our findings demonstrate the potential of LLMs for supporting personalized indoor navigation, with an average of 86.59% correct indications and a maximum of 97.14%. The proposed system achieves high accuracy and reasoning performance. These results have key implications for AI-driven navigation and assistive technologies.

Background: Multi-label requirements classification is an inherently challenging task, especially when dealing with numerous classes at varying levels of abstraction. The task becomes even more difficult when a limited number of requirements is available to train a supervised classifier. Zero-shot learning does not require training data and can potentially address this problem. Objective: This paper investigates the performance of zero-shot classifiers on a multi-label industrial dataset. The study focuses on classifying requirements according to a hierarchical taxonomy designed to support requirements tracing. Method: We compare multiple variants of zero-shot classifiers using different embeddings, including 9 language models (LMs) with a reduced number of parameters (up to 3B), e.g., BERT, and 5 large LMs (LLMs) with a large number of parameters (up to 70B), e.g., Llama. Our ground truth includes 377 requirements and 1968 labels from 6 output spaces. For the evaluation, we adopt traditional metrics, i.e., precision, recall,$F_{1}$, and$F_{\beta}$, as well as a novel label distance metric$D_{n}$. This aims to better capture the classification's hierarchical nature and to provide a more nuanced evaluation of how far the results are from the ground truth. Results: 1) The top-performing model on 5 out of$\mathbf{6}$output spaces is TS-xl, with maximum$F_{\beta}=0.78$and$D_{n}=0.04$, while BERT base outperformed the other models in one case, with maximum$F_{\beta}=0.83$and$D_{n}=0.04.2$) LMs with smaller parameter size produce the best classification results compared to LLMs. Thus, addressing the problem in practice is feasible as limited computing power is needed. 3) The model architecture (auto encoding, autoregression, and sentence-to-sentence) significantly affects the classifier's performance. Contribution: We conclude that using zero-shot learning for multi-label requirements classification offers promising results. We also present a novel metric that can be used to select the top-performing model for this problem.

Context: Attack-Defense Trees (ADTs) are a graphical notation used to model and evaluate security requirements. ADTs are popular because they facilitate communication among different stakeholders involved in system security evaluation and are formal enough to be verified using methods like model checking. The understandability and user-friendliness of ADTs are claimed as key factors in their success, but these aspects, along with user acceptance, have not been evaluated empirically. Objectives: This paper presents an experiment with 25 subjects designed to assess the understandability and user acceptance of the ADT notation, along with an internal replication involving 49 subjects. Methods: The experiments adapt the Method Evaluation Model (MEM) to examine understandability variables (i.e., effectiveness and efficiency in using ADTs) and user acceptance variables (i.e., ease of use, usefulness, and intention to use). The MEM is also used to evaluate the relationships between these dimensions. In addition, a comparative analysis of the results of the two experiments is carried out. Results: With some minor differences, the outcomes of the two experiments are aligned. The results demonstrate that ADTs are well understood by participants, with values of understandability variables significantly above established thresholds. They are also highly appreciated, particularly for their ease of use. The results also show that users who are more effective in using the notation tend to evaluate it better in terms of usefulness. Conclusion: These studies provide empirical evidence supporting both the understandability and perceived acceptance of ADTs, thus encouraging further adoption of the notation in industrial contexts, and development of supporting tools.

Large Language Models (LLMs) have made remarkable advancements in emulating human linguistic capabilities, showing potential also in executing various requirements engineering (RE) tasks. However, despite their generally good performance, the adoption of LLM-generated solutions and artefacts prompts concerns about their correctness, fairness, and trustworthiness. This paper aims to address the concerns associated with the use of LLMs in RE activities. Specifically, it seeks to develop a roadmap that leverages formal methods (FMs) to provide guarantees of correctness, fairness, and trustworthiness when LLMs are utilised in RE. Symmetrically, it aims to explore how LLMs can be employed to make FMs more accessible. We use two sets of examples to show the current limits of FMs when used in software development and of LLMs when used for RE tasks. The highlighted limitations are addressed by proposing two roadmaps grounded in the current literature and technologies. The proposed examples show the potential and limits of FMs in supporting software development and of LLMs when used for RE tasks. The initial investigation into how these limitations can be overcome has been concretised in two detailed roadmaps for the RE and, more largely, the software engineering community. The proposed roadmaps offer a promising approach to address the concerns of correctness, fairness, and trustworthiness associated with the use of LLMs in RE tasks through the use of FMs and to enhance the accessibility of FMs by utilising LLMs. • We exemplify the use of formal methods in software development • We outline a roadmap to increase usability of formal methods with the support of LLMs • We show how LLMs can be support requirements engineers in automating manual tasks • We propose a roadmap for the use of formal techniques to make LLMs reliable

Rigorous software development involves using highly structured methods and processes in software and system engineering to ensure that the developed products are correct, reliable, and robust. In this context, model-driven development (MDD) has emerged as a development paradigm that emphasizes designing software systems by means of graphical or textual models at different levels of abstraction, which capture different aspects or dimensions of the system-to-be. At the core of MDD is model transformation , which is the process of translating one model into another, according to specific rules. Property preservation in MDD refers to maintaining specific properties of the system model during transformations, including structural, behavioral, and domain-specific constraints. Over the past decades, research on model transformation and property preservation has seen several contributions. In this paper, we present a systematic literature review (SLR) to compile information on study demographics, model properties considered, techniques to ensure property preservation, and other aspects. In addition, through thematic analysis, we highlight significant challenges and benefits associated with model transformation and property preservation. We analyze 202 research studies published between 2000 and 2024. Most of the studies concern case studies (62, 31%) and rigorous analysis (49, 24%), while experimental studies using human subjects are limited (1). Formal logic is the most commonly used transformation language, used in 42 studies (21%), while the Unified Modeling Language (UML) is also used for source (58, 29%) and target (25, 12%) modeling. A total of 100 of the studies (50%) performed system testing on models, while 44 of the studies (22%) used transformation rules to verify transformation properties . Among the verified model properties, 66 studies (33%) focused on consistency management, while 4 (2%) are related to model maintainability and reuse. We conclude from our SLR that property preservation could be improved by using model-specific verification methods and strategies based on the considered model artifacts. Our research also provides a relevant contribution by identifying the major challenges in MDD and proposing relevant solutions. Editor’s note: Open Science material was validated by the Journal of Systems and Software Open Science Board .

The International Working Conference on Requirement Engineering: Foundation for Software Quality (REFSQ) is an established international forum. Its goal is to foster the establishment and maintenance of a strong Requirements Engineering (RE) community across industry and academia through contributions that report on novel ideas and techniques to enhance the quality of RE’s products and processes.

Natural language processing (NLP) techniques have been widely applied in the requirements engineering (RE) field to support tasks such as classification and ambiguity detection. Despite its empirical vocation, RE research has given limited attention to replication of NLP for RE studies. Replication is hampered by several factors, including the context specificity of the studies, the heterogeneity of the tasks involving NLP, the tasks’ inherent hairiness , and, in turn, the heterogeneous reporting structure. To address these issues, we propose a new artifact, referred to as ID-Card , whose goal is to provide a structured summary of research papers emphasizing replication-relevant information. We construct the ID-Card through a structured, iterative process based on design science. In this article: (i) we report on hands-on experiences of replication; (ii) we review the state-of-the-art and extract replication-relevant information: (iii) we identify, through focus groups, challenges across two typical dimensions of replication: data annotation and tool reconstruction; and (iv) we present the concept and structure of the ID-Card to mitigate the identified challenges. This study aims to create awareness of replication in NLP for RE. We propose an ID-Card that is intended to foster study replication but can also be used in other contexts, e.g., for educational purposes.

Capturing users’ engagement is crucial for gathering feedback about the features of a software product. In a market-driven context, current approaches to collecting and analyzing users’ feedback are based on techniques leveraging information extracted from product reviews and social media. These approaches are hardly applicable in contexts where online feedback is limited, as for the majority of apps, and software in general. In such cases, companies need to resort to face-to-face interviews to get feedback on their products. In this article, we propose to utilize biometric data, in terms of physiological and voice features, to complement product feedback interviews with information about the engagement of the user on product-relevant topics. We evaluate our approach by interviewing users while gathering their physiological data (i.e., biofeedback ) using an Empatica E4 wristband, and capturing their voice through the default audio-recorder of a common laptop. Our results show that we can predict users’ engagement by training supervised machine learning algorithms on biofeedback and voice data, and that voice features alone can be sufficiently effective. The best configurations evaluated achieve an average F1 ∼ 70% in terms of classification performance, and use voice features only. This work is one of the first studies in requirements engineering in which biometrics are used to identify emotions. Furthermore, this is one of the first studies in software engineering that considers voice analysis. The usage of voice features can be particularly helpful for emotion-aware feedback collection in remote communication, either performed by human analysts or voice-based chatbots, and can also be exploited to support the analysis of meetings in software engineering research.

App store-inspired elicitation is the practice of exploring competitors' apps, to get inspiration for requirements. This activity is common among developers, but little insight is available on its practical use, advantages and possible issues. This paper aims to empirically analyse this technique in a realistic scenario, in which it is used to extend the requirements of a product that were initially captured by means of more traditional requirements elicitation interviews. Considering this scenario, we conduct an experimental simulation with 58 analysts and collect qualitative data. We perform thematic analysis of the data to identify strategies, benefits, and challenges of app store-inspired elicitation, as well as differences with respect to interviews in the considered elicitation setting. Our results show that: (1) specific guidelines and procedures are required to better conduct app store-inspired elicitation; (2) current search features made available by app stores are not suitable for this practice, and more tool support is required to help analysts in the retrieval and evaluation of competing products; (3) while interviews focus on the why dimension of requirements engineering (i.e., goals), app store-inspired elicitation focuses on how (i.e., solutions), offering indications for implementation and improved usability. Our study provides a framework for researchers to address existing challenges and suggests possible benefits to fostering app store-inspired elicitation among practitioners.

Language workbenches are tools that enable the definition, reuse and composition of programming languages and their ecosystem. This breed of frameworks aims to make the development of new languages easier and more affordable. Consequently, the comprehensibility of the language used in a language workbench (i.e., the meta-language) should be an important aspect to consider and evaluate. To the best of our knowledge, although the quantitative aspects of language workbenches are often discussed in the literature, the evaluation of comprehensibility is typically neglected.Neverlang is a language workbench that enables the definition of languages with a modular approach. This paper presents a preliminary study that intends to assess the comprehensibility of Neverlang programs, evaluated in terms of users’ effectiveness and efficiency in a code comprehension task. The study also investigates the relationship between Neverlang comprehensibility and the users’ working memory capacity. Furthermore, we intend to capture the relationship between Neverlang comprehensibility and users’ acceptance, in terms of perceived ease of use, perceived usefulness, and intention to use. Our preliminary results on 10 subjects suggest that the users’ working memory capacity may be related to the ability to comprehend Neverlang programs. On the other hand, effectiveness and efficiency do not appear to be associated with an increase in users’ acceptance variables.

Allocation of requirements to different teams is a typical preliminary task in large-scale system development projects. This critical activity is often performed manually and can benefit from automated requirements classification techniques. To date, limited evidence is available about the effectiveness of existing machine learning (ML) approaches for requirements classification in industrial cases. This paper aims to fill this gap by evaluating state-of-the-art language models and ML algorithms for classification in the railway industry. Since the interpretation of the results of ML systems is particularly relevant in the studied context, we also provide an information augmentation approach to complement the output of the ML-based classification. Our results show that the BERT uncased language model with the softmax classifier can allocate the requirements to different teams with a 76% F1 score when considering requirements allocation to the most frequent teams. Information augmentation provides potentially useful indications in 76% of the cases. The results confirm that currently available techniques can be applied to real-world cases, thus enabling the first step for technology transfer of automated requirements classification. The study can be useful to practitioners operating in requirements-centered contexts such as railways, where accurate requirements classification becomes crucial for better allocation of requirements to various teams.

Designing systems that account for sustainability concerns demands for a better understanding of the impact that digital technology interventions can have on a certain socio-technical context. However, limited studies are available about the elicitation of impact-related information from stakeholders, and strategies are particularly needed to elicit possible longterm effects, including negative ones, that go beyond the planned system goals. This paper reports a case study about the impact of digitalisation in remote mountain areas, in the context of a system for ordinary land management and hydro-geological risk control. The elicitation process was based on interviews and workshops. In the initial phase, past and present impacts were identified. In a second phase, future impacts were forecasted through the discussion of two alternative scenarios: a dystopic, technology-intensive one, and a technology-balanced one. The approach was particularly effective in identifying negative impacts. Among them, we highlight the higher stress due to the excess of connectivity, the partial reduction of decision-making abilities, and the risk of marginalisation for certain types of stakeholders. The study posits that before the elicitation of system goals, requirements engineers need to identify the socio-economic impacts of ICT technologies included in the system, as negative effects need to be properly mitigated. Our study contributes to the literature with: a set of impacts specific to the case, which can apply to similar contexts; an effective approach for impact elicitation; and a list of lessons learned from the experience.

This contribution presents ModeLLer, a prototype of a web tool for system modelling based on a block-based visual editor. The aim of ModeLLer is to enable collaborative environments in requirements elicitation, allowing end-users to create UML class diagrams without any knowledge of the (semi-)formal UML notation.

Requirements engineering (RE) researchers have been experimenting with machine learning (ML) and deep learning (DL) approaches for a range of RE tasks, such as requirements classification, requirements tracing, ambiguity detection, and modelling. However, most of today’s ML/DL approaches are based on supervised learning techniques, meaning that they need to be trained using a large amount of task-specific labelled training data. This constraint poses an enormous challenge to RE researchers, as the lack of labelled data makes it difficult for them to fully exploit the benefit of advanced ML/DL technologies. This paper addresses this problem by showing how a zero-shot learning (ZSL) approach can be used for requirements classification without using any labelled training data. We focus on the classification task because many RE tasks can be framed as classification problems. The ZSL approach used in our study employs contextual word-embeddings and transformer-based language models (LMs). We demonstrate this approach through a series of experiments to perform three classification tasks: (1) FR/NFR — classification functional requirements vs non-functional requirements; (2) NFR — identification of NFR classes; (3) Security — classification of security vs non-security requirements. The study shows that the ZSL approach achieves an F1 score of 0.66 for the FR/NFR task. For the NFR task, the approach yields F1∼0.72−0.80, considering the most frequent classes. For the Security task, F1 ∼0.66. All of the aforementioned F1 scores are achieved with zero-training efforts. This study demonstrates the potential of ZSL for requirements classification. An important implication is that it is possible to have very little or no training data to perform classification tasks. The proposed approach thus contributes to the solution of the long-standing problem of data shortage in RE.

Abstract Recommender systems for requirements are typically built on the assumption that similar requirements can be used as proxies to retrieve similar software. When a stakeholder proposes a new requirement, natural language processing (NLP)-based similarity metrics can be exploited to retrieve existing requirements, and in turn, identify previously developed code. Several NLP approaches for similarity computation between requirements are available. However, there is little empirical evidence on their effectiveness for code retrieval. This study compares different NLP approaches, from lexical ones to semantic, deep-learning techniques, and correlates the similarity among requirements with the similarity of their associated software. The evaluation is conducted on real-world requirements from two industrial projects from a railway company. Specifically, the most similar pairs of requirements across two industrial projects are automatically identified using six language models. Then, the trace links between requirements and software are used to identify the software pairs associated with each requirements pair. The software similarity between pairs is then automatically computed with JPLag. Finally, the correlation between requirements similarity and software similarity is evaluated to see which language model shows the highest correlation and is thus more appropriate for code retrieval. In addition, we perform a focus group with members of the company to collect qualitative data. Results show a moderately positive correlation between requirements similarity and software similarity, with the pre-trained deep learning-based BERT language model with preprocessing outperforming the other models. Practitioners confirm that requirements similarity is generally regarded as a proxy for software similarity. However, they also highlight that additional aspect comes into play when deciding software reuse, e.g., domain/project knowledge, information coming from test cases, and trace links. Our work is among the first ones to explore the relationship between requirements and software similarity from a quantitative and qualitative standpoint. This can be useful not only in recommender systems but also in other requirements engineering tasks in which similarity computation is relevant, such as tracing and change impact analysis.

The domain of rural areas, including rural communities, agriculture, and forestry, is going through a process of deep digital transformation. Digitalisation can have positive impacts on sustainability in terms of greater environmental control, and community prosperity. At the same time, it can also have disruptive effects, with the marginalisation of actors that cannot cope with the change. When developing a novel system for rural areas, requirements engineers should carefully consider the specific socio-economic characteristics of the domain, so that potential positive effects can be maximised, while mitigating negative impacts. The goal of this paper is to support requirements engineers with a reference catalogue of drivers, barriers and potential impacts associated to the introduction of novel ICT solutions in rural areas. To this end, we interview 30 cross-disciplinary experts in digitalisation of rural areas, and we analyse the transcripts to identify common themes. According to the experts, main drivers are economic, with the possibility of reducing costs, and regulatory, as institutions push for more precise tracing and monitoring of production; barriers are the limited connectivity, but also distrust towards technology and other socio-cultural aspects; positive impacts are socio-economic (e.g., reduction of manual labour, greater productivity), while negative ones include potential dependency from technology, with loss of hands-on expertise, and marginalisation of certain actors (e.g., small farms, subjects with limited education). This paper contributes to the literature with a domain-specific catalogue that characterises digitalisation in rural areas. The catalogue can be used as a reference baseline for requirements elicitation endeavours in rural areas, to support domain analysis prior to the development of novel solutions, as well as fit-gap analysis for the adaptation of existing technologies.

Abstract Requirements are elicited from the customer and other stakeholders through an iterative process of interviews, prototyping, and other interactive sessions. Then, requirements can be further extended, based on the analysis of the features of competing products available on the market. Understanding how this process takes place can help to identify the contribution of the different elicitation phases, thereby allowing requirements analysts to better distribute their resources. In this work, we empirically study in which way requirements get transformed from initial ideas into documented needs, and then evolve based on the inspiration coming from similar products. To this end, we select 30 subjects that act as requirements analysts, and we perform interview-based elicitation sessions with a fictional customer. After the sessions, the analysts produce a first set of requirements for the system. Then, they are required to search similar products in the app stores and extend the requirements, inspired by the identified apps. The requirements documented at each step are evaluated, to assess to which extent and in which way the initial idea evolved throughout the process. Our results show that only between 30% and 38% of the requirements produced after the interviews include content that can be fully traced to initial customer’s ideas. The rest of the content is dedicated to new requirements, and up to 21% of it belongs to completely novel topics. Furthermore, up to 42% of the requirements inspired by the app stores cover additional features compared to the ones identified after the interviews. The results empirically show that requirements are not elicited in strict sense, but actually co-created through interviews, with analysts playing a crucial role in the process. In addition, we show evidence that app store-inspired elicitation can be particularly beneficial to complete the requirements.

Abstract Shift2Rail is a joint undertaking funded by the EU via its Horizon 2020 program and by main railway stakeholders. Several Shift2Rail projects aim to investigate the application of formal methods to new ERTMS/ETCS railway signalling systems that promise to move European railway forward by guaranteeing high capacity, low cost and improved reliability. We explore the ERTMS/ETCS level 3 full moving block specifications stemming from different Shift2Rail projects using Uppaal and statistical model checking. The results range from novel rigorously formalised requirements to an operational model formally verified against scenarios with multiple trains on a single railway line. From the gained experience, we have distilled future research goals to improve the formal specification and verification of real-time systems, and we discuss some barriers concerning a possible uptake of formal methods and tools in the railway industry.

Formal methods and supporting tools have a long record of success in the development of safety-critical systems. However, no single tool has emerged as the dominant solution for system design. Each tool differs from the others in terms of the modeling language used, its verification capabilities and other complementary features, and each development context has peculiar needs that require different tools. This is particularly problematic for the railway industry, in which formal methods are highly recommended by the norms, but no actual guidance is provided for the selection of tools. To guide companies in the selection of the most appropriate formal methods tools to adopt in their contexts, a clear assessment of the features of the currently available tools is required. To address this goal, this paper considers a set of 13 formal methods tools that have been used for the early design of railway systems, and it presents a systematic evaluation of such tools and a preliminary usability analysis of a subset of 7 tools, involving railway practitioners. The results are discussed considering the most desired aspects by industry and earlier related studies. While the focus is on the railway signaling domain, the overall methodology can be applied to similar contexts. Our study thus contributes with a systematic evaluation of formal methods tools and it shows that despite the poor graphical interfaces,usabilityandmaturityof the tools are not major problems, as claimed by contributions from the literature. Instead, support forprocess integrationis the most relevant obstacle for the adoption of most of the tools. Our contribution can be useful to R&D engineers from railway signaling companies and infrastructure managers, but also to tool developers and academic researchers alike.

To complement the manual contact-tracing methods, a flood of coronavirus-related apps was launched in the first half of 2020. Despite the incredible promises made by the governments, contact-tracing apps did not live up to expectations. We provide a contextual perspective of the government commissioned contact-tracing apps from four countries to understand the non-functional requirements (NFRs) and socio-technical factors that hindered the success of these apps. We collected the user reviews from the app stores for iOS and Android versions and identified top news articles related to each app. Our analysis revealed that the dominant factors behind the negligible success of these apps are complex and entangled with the cultural and political dimensions rather than being just technical. The multilayer diversity of the target users also impacted the design and development of contact-tracing apps in an extremely challenging situation. This perspective paper brings into light important elements, such as politics and socio-cultural aspects that should be studied in the design of contact-tracing apps, and public apps in general.

Requirements are elicited from the customer and other stakeholders through an iterative process of interviews, prototyping, and other interactive sessions. Many communication phenomena may emerge in these early iterations, that lead initial ideas to be transformed, renegotiated, or reframed. Understanding how this process takes place can help in solving possible communication issues as well as their consequences. In this work, we perform an exploratory study of descriptive nature to understand in which way requirements get transformed from initial ideas into documented needs. To this end, we select 30 subjects that act as requirements analysts, and we perform a set of elicitation sessions with a fictional customer. The customer is required to study a sample requirements document for a system beforehand and to answer the questions of the analysts about the system. After the elicitation sessions, the analysts produce user stories for the system. These are compared with the original ones by two researchers to assess to which extent and in which way the initial requirements evolved throughout the interactive sessions. Our results show that between 30% and 38% of the produced user stories include content that can be fully traced to the initial ones, while the rest of the content is dedicated to new requirements. We also show what types of requirements are introduced through the elicitation process, and how they vary depending on the analyst. Our work contributes to theory in requirements engineering, with empirically grounded, quantitative data, concerning the impact of elicitation activities with respect to initial ideas.

Formal methods and tools have a long history of successful applications in the design of safety-critical railway products. However, most of the experiences focused on the application of a single method at once, and little work has been performed to compare the applicability of the different available frameworks to the railway context. As a result, companies willing to introduce formal methods in their development process have little guidance on the selection of tools that could fit their needs. To address this goal, this paper presents a comparison between 9 different formal tools, namely Atelier B, CADP, FDR4, NuSMV, ProB, Simulink, SPIN, UMC, and UPPAAL SMC. We performed a judgment study, involving 17 experts with experience in formal methods applied to railways. In the study, part of the experts were required to model a railway signaling problem (a moving-block train distancing system) with the different tools, and to provide feedback on their experience. The information produced was then synthesized, and the results were validated by the remaining experts. Based on the outcome of this process, we provide a synthesis that describes when to use a certain tool, and what are the problems that may be faced by modelers. Our experience shows that the different tools serve different purposes, and multiple formal methods are required to fully cover the needs of the railway system design process.

The core aim of requirements inspection is to ensure the high quality of already elicited requirements in the Software Requirements Specification. Teaching requirements inspection to novices is challenging, as inspecting requirements needs several skills as well as knowledge of the product and process that is hard to achieve in a classroom environment. Published studies about pedagogical design specifically for teaching requirements inspection are scarce. Our objective is to present the design and evaluation of a postgraduate course for requirements inspection training. We conducted an empirical study with 138 postgraduate students, teamed up in 34 groups to conduct requirements inspection. We performed qualitative analysis on the data collected from students' reflection reports to assess the effects of the pedagogical design in terms of benefits and challenges. We also quantitatively analyze the correlation between the students' performance in conducting inspections and their ability of writing specifications. From the analysis of students' reflections, several themes emerged such as their difficulty of working with limited information, but also revealed the benefits of learning teamwork and writing good requirements. This qualitative analysis also provides recommendations for improving the related activities. The results revealed a moderate positive correlation between the performance in writing specification and inspection.

Capturing users' engagement is crucial for gathering feedback about the features of a software product. In a market-driven context, current approaches to collect and analyze users' feedback are based on techniques leveraging information extracted from product reviews and social media. These approaches are hardly applicable in bespoke software development, or in contexts in which one needs to gather information from specific users. In such cases, companies need to resort to face-to-face interviews to get feedback on their products. In this paper, we propose to utilize biofeedback to complement interviews with information about the engagement of the user on the discussed features and topics. We evaluate our approach by interviewing users while gathering their biometric data using an Empatica E4 wristband. Our results show that we can predict users' engagement by training supervised machine learning algorithms on the biometric data. The results of our work can be used to facilitate the prioritization of product features and to guide the interview based on users' engagement.

Efficient and safe railway signalling systems, together with energy-saving infrastructures, are among the main pillars to guarantee sustainable transportation. ERTMS L3 moving block is one of the next generation railway signalling systems currently under trial deployment, with the promise of increased capacity on railway tracks, reduced costs and improved reliability. We report an experience in modelling a satellite-based ERTMS L3 moving block signalling system from the railway industry with Simulink and Uppaal and analysing the Uppaal model with Uppaal SMC. The lessons learned range from demonstrating the feasibility of applying Uppaal SMC in a moving block railway context, to the offered possibility of fine tuning communication parameters in satellite-based ERTMS L3 moving block railway signalling system models that are fundamental for the reliability of their operational behaviour.

Interviews are largely used in the practice of requirements elicitation. Nevertheless, performing an effective interview often depends on soft-skills, and on knowledge acquired through experience. When it comes to requirements engineering education and training (REET), limited resources and few well-founded pedagogical approaches are available to allow students to acquire and improve their skills as interviewers. This paper presents a novel pedagogical approach that combines role-playing, peer-review and self-assessment to enable students to reflect on their mistakes, and improve their interview skills. We evaluate the approach through a controlled quasi-experiment. The study shows that the approach significantly reduces the amount of mistakes made by the students. Feedback from the participants confirms the usefulness and easiness of the proposed training. This work contributes to the body of knowledge of REET with an empirically evaluated method for teaching inter-views. Furthermore, we share the pedagogical material used, to enable other educators to apply and possibly tailor the approach.

[Context] Interviews are the most widely used elicitation technique in requirements engineering. However, conducting effective requirements elicitation interviews is challenging, due to the combination of technical and soft skills that requirements analysts often acquire after a long period of professional practice. Empirical evidence about training the novices on conducting effective requirements elicitation interviews is scarce. [Objectives] We present a list of most common mistakes that novices make in requirements elicitation interviews. The objective is to assist the educators in teaching interviewing skills to student analysts. [Re-search Method] We conducted an empirical study involving role-playing and authentic assessment with 110 students, teamed up in 28 groups, to conduct interviews with a customer. One re-searcher made observation notes during the interview while two researchers reviewed the recordings. We qualitatively analyzed the data to identify the themes and classify the mistakes. [Results and conclusion] We identified 34 unique mistakes classified into 7 high level themes. We also give examples of the mistakes made by the novices in each theme, to assist the educationists and trainers. Our research design is a novel combination of well-known pedagogical approaches described in sufficient details to make it re-peatable for future requirements engineering education and training research.

The requirements problem consists of transforming stakeholder requirements - however informal, ambiguous, conflicting, unattainable, imprecise and incomplete – into a consistent, complete and realizable specification through a systematic process. We propose a refinement calculus for requirements engineering (CaRE) for solving this problem, which takes into account the typically dialectic nature of requirements activities. The calculus casts the requirement problem as an iterative argument between stakeholders and requirements engineers, where posited requirements are attacked for being ambiguous, incomplete, etc. and refined into new requirements that address the defect pointed out by the attack. Refinements are carried out by operators provided by CaRE that refine (e.g., strengthen, weaken, decompose) existing requirements, to build a refinement graph. The semantics of the operators is provided by means of argumentation theory. Examples are given to illustrate the elements of our proposal.

The engineering of software product lines begins with the identification of the possible variation points. To this aim, natural language (NL) requirement documents can be used as a source from which variability-relevant information can be elicited. In this paper, we propose to identify variability issues as a subset of the ambiguity defects found in NL requirement documents. To validate the proposal, we single out ambiguities using an available NL analysis tool, QuARS, and we classify the ambiguities returned by the tool by distinguishing among false positives, real ambiguities, and variation points, by independent analysis and successive agreement phase. We consider three different sets of requirements and collect the data that come from the analysis performed.

Summary European Union (EU) projects are means of the European Commission for funding research activities. Such projects address challenging research objectives by involving both academic and industrial partners, from several countries. Information and communication technologies–related projects often undertake to deliver a software system prototype. In such a context, most of the typical issues of global requirements engineering may emerge. Partners can have different background and expertise, needs are not sharply defined, and communication is hampered by linguistic and cultural differences. If these issues are not carefully taken into account from the beginning, problems frequently emerge during project execution. This paper presents the experience of applying a customized elicitation and refinement approach in the context of the Learn PAd EU project, which involved about 50 people. The approach combines collaborative elicitation and wiki‐based refinement sessions to come to a set of consolidated requirements. Lessons learnt are discussed as a guidance for researchers dealing with analogous issues in similar contexts. Some of the major observations refer to the importance of initial face‐to‐face meetings when combined with asynchronous remote interactions; the role of moderators that have to encourage collaboration and foster a shared understanding; and the definition of guidelines to select wiki‐based platforms.

Context and Motivation Natural language processing has been used since the 1980s to construct tools for performing natural language (NL) requirements engineering (RE) tasks. The RE field has often adopted information retrieval (IR) algorithms for use in implementing these NL RE tools. Problem Traditionally, the methods for evaluating an NL RE tool have been inherited from the IR field without adapting them to the requirements of the RE context in which the NL RE tool is used. Principal Ideas This panel discusses the problem and considers the evaluation of tools for a number of NL RE tasks in a number of contexts. Contribution The discussion is aimed at helping the RE field begin to consistently evaluate each of its tools according to the requirements of the tool's task.

The requirements elicitation process often starts with an interview between a customer and a requirements analyst. During these interviews, ambiguities in the dialogic discourse may reveal the presence of tacit knowledge that needs to be made explicit. It is therefore important to understand the nature of ambiguities in interviews and to provide analysts with cognitive tools to identify and alleviate ambiguities. Ambiguities perceived by analysts are sometimes triggered by specific categories of terms used by the customer such as pronouns, quantifiers, and vague or under-specified terms. However, many of the ambiguities that arise in practice cannot be rooted in single terms. Rather, entire fragments of speech and their relation to the mental state of the analyst need to be considered.In this paper, we show that particular types of ambiguities can be characterised by means of argumentation theory. Argumentation is the study of how conclusions can be reached through logical reasoning. In an argumentation theory, statements are represented as arguments, and conflict relations among statements are represented as attacks. Based on a set of ambiguous fragments extracted from interviews, we define a model of the mental state of the analyst during an interview and translate it into an argumentation theory. Then, we show that many of the ambiguities can be characterized in terms of 'attacks' on arguments. The main novelty of this work is in addressing the problem of explaining fragment-level ambiguities in requirements elicitation interviews through the formal modeling of the analyst's mental model using argumentation theory. Our contribution provides a data-grounded, theoretical basis to have a more complete understanding of the ambiguity phenomenon, and lays the foundations to design intelligent computer-based agents that are able to automatically identify ambiguities.

In requirements elicitation interviews, ambiguities identified by analysts can help to disclose the tacit knowledge of customers. Indeed, ambiguities might reveal implicit or hard to express information that needs to be elicited. The perception of ambiguity might depend on the subject who is acting as analyst, and different analysts might identify different ambiguities in the same interview. Based on this intuition, we propose to investigate the difference between ambiguities explicitly revealed by an analyst during a requirements elicitation interview, and ambiguities annotated by a reviewer who listens to the interview recording, with the objective of defining a method for interview review. We performed an exploratory study in which two subjects listened to a set of customer-analyst interviews. Only in 26% of the cases the ambiguities revealed by the analysts matched with the ambiguities found by the reviewers. In 46% of the cases, ambiguities were found by the reviewers, and were not detected by the analysts. Based on these preliminary findings, we are currently performing a controlled experiment with students of two universities, which will be followed by a real-world case study with companies. This paper discusses the current results, together with our research plan.

This paper presents PURE (PUblic REquirements dataset), a dataset of 79 publicly available natural language requirements documents collected from the Web. The dataset includes 34,268 sentences and can be used for natural language processing tasks that are typical in requirements engineering, such as model synthesis, abstraction identification and document structure assessment. It can be further annotated to work as a benchmark for other tasks, such as ambiguity detection, requirements categorisation and identification of equivalent re-quirements. In the paper, we present the dataset and we compare its language with generic English texts, showing the peculiarities of the requirements jargon, made of a restricted vocabulary of domain-specific acronyms and words, and long sentences. We also present the common XML format to which we have manually ported a subset of the documents, with the goal of facilitating replication of NLP experiments.

Requirements elicitation is the initial step of the requirements engineering process and aims at gathering all the relevant requirements through the direct or indirect interactions between requirements analysts and stakeholders. Even if the requirements elicitation problem is not new and has been approached many times over the years, it is still considered one of the most challenging of the requirements engineering process. In the proposed presentation, we aim at analyzing the journey of the research on requirements elicitation through the 25 years of the Requirements Engineering conference not only by considering the different proposed approaches and their evolution, but also by evaluating the role of requirements elicitation in the conference. Moreover, we will present the lessons learnt during this analysis and will use them as a starting point to present the current trends and outline possible future directions.

Customer-analyst interviews are considered among the most effective means to perform requirements elicitation. However, during these interviews, ambiguity can hamper communication between customer and requirements analyst. Ambiguity is particularly dangerous in those cases in which the analyst misunderstands some linguistic expression of the customer, with-out being aware of the misunderstanding. On the other hand, if the analyst is able to detect ambiguous situations, this has been shown to help him/her in disclosing tacit knowledge. Indeed, the occurrence of an ambiguity might reveal the presence of unexpressed, system-relevant knowledge that needs to be elicited. Therefore, for the requirements elicitation interview to succeed, it is important for the analyst not to overlook ambiguities. To support the ambiguity-awareness of the requirements analyst, this paper aims to provide a set of cues that can be identified in the linguistic expressions of the customer, and that typically lead to ambiguity. To this end, we performed 34 customer-analyst interviews, and we isolated the speech fragments that caused the ambiguity. Based on the analysis of these fragments, and leveraging the previous literature on ambiguity in written requirements, we identified a set of cues that can be used by requirements analysts as a reference handbook to detect ambiguities.

Interviews with stakeholders are the most commonly used elicitation technique, as they are considered one of the most effective ways to transfer knowledge between requirements analysts and customers. During these interviews, ambiguity is a major obstacle for knowledge transfer, as it can lead to incorrectly understood needs and domain aspects and may ultimately result in poorly defined requirements. To address this issue, previous work focused on how ambiguity is perceived on the analyst side, i.e., when the analyst perceives an expression of the customer as ambiguous. However, this work did not consider how ambiguity can affect customers, i.e., when questions from the analyst are perceived as ambiguous. Since customers are not in general trained to cope with ambiguity, it is important to provide analysts with techniques that can help them to identify these situations. To support the analysts in this task, we propose to explore the relation between a perceived ambiguity on the customer side, and changes in the voice and bio parameters of that customer. To realize our idea, we plan to (1) study how changes in the voice and bio parameters can be correlated to the levels of stress, confusion, and uncertainty of an interviewee and, ultimately, to ambiguity and (2) investigate the application of modern voice analyzers and wristbands in the context of customer-analyst interviews. To show the feasibility of the idea, in this paper we present the result of our first step in this direction:an overview of different voice analyzers and wristbands that can collect bio parameters and their application in similar contexts. Moreover, we propose a plan to carry our research out.

Interviews are the most common and effective means to perform requirements elicitation and support knowledge transfer between a customer and a requirements analyst. Ambiguity in communication is often perceived as a major obstacle for knowledge transfer, which could lead to unclear and incomplete requirements documents. In this paper, we analyse the role of ambiguity in requirements elicitation interviews. To this end, we have performed a set of customer-analyst interviews to observe how ambiguity occurs during requirements elicitation. From this direct experience, we have observed that ambiguity is a multi-dimensional cognitive phenomenon with a dominant pragmatic facet, and we have defined a phenomenological framework to describe the different types of ambiguity in interviews. We have also discovered that, rather than an obstacle, the occurrence of an ambiguity is often a resource for discovering tacit knowledge. Starting from this observation, we have envisioned the further steps needed in the research to exploit these findings.

A business subject who wishes to enter an established technological market is required to accurately analyse the features of the products of the different competitors. Such features are normally accessible through natural language (NL) brochures, or NL Web pages, which describe the products to potential customers. Building a feature model that hierarchically summarises the different features available in competing products can bring relevant benefits in market analysis. A company can easily visualise existing features, and reason about aspects that are not covered by the available solutions. However, designing a feature model starting from publicly available documents of existing products is a time consuming and error-prone task. In this paper, we present two tools, namely Commonality Mining Tool (CMT) and Feature Diagram Editor (FDE), which can jointly support the feature model definition process. CMT allows mining common and variant features from NL descriptions of existing products, by leveraging a natural language processing (NLP) approach based on contrastive analysis, which allows identifying domain-relevant terms from NL documents. FDE takes the commonalities and variabilities extracted by CMT, and renders them in a visual form. Moreover, FDE allows the graphical design and refinement of the final feature model, by means of an intuitive GUI.

A company who wishes to enter an established marked with a new, competitive product is required to analyse the product solutions of the competitors. Identifying and comparing the features provided by the other vendors might greatly help during the market analysis. However, mining common and variant features of from the publicly available documents of the competitors is a time consuming and error-prone task. In this paper, we suggest to employ a natural language processing approach based on contrastive analysis to identify commonalities and variabilities from the brochures of a group of vendors. We present a first step towards a practical application of the approach, in the the context of the market of Communications-Based Train Control (CBTC) systems.

This paper presents a novel approach for pragmatic ambiguity detection in natural language (NL) requirements specifications defined for a specific application domain. Starting from a requirements specification, we use a Web-search engine to retrieve a set of documents focused on the same domain of the specification. From these domain-related documents, we extract different knowledge graphs, which are employed to analyse each requirement sentence looking for potential ambiguities. To this end, an algorithm has been developed that takes the concepts expressed in the sentence and searches for corresponding “concept paths” within each graph. The paths resulting from the traversal of each graph are compared and, if their overall similarity score is lower than a given threshold, the requirements specification sentence is considered ambiguous from the pragmatic point of view. A proof of concept is given throughout the paper to illustrate the soundness of the proposed strategy.

This article presents the experience of a railway signaling manufacturer in introducing the technologies of model based testing and abstract interpretation as part of its development process. Preliminary results show the better performance of these techniques with respect to the previously employed structural coverage based testing.