Context and Motivation. Fairness in socio-technical systems is increasingly recognised as a critical requirement, especially in processes involving human-AI interaction. Fairness hazards are situations or factors that threaten the fair treatment of individuals or groups. If left unaddressed, they can accumulate into systemic bias. Therefore, ensuring fairness must be treated as a first-class requirement during system design, rather than a post-hoc fix. Question/Problem. Systematic methods for identifying fairness hazards in socio-technical workflows and translating them into requirements-level mitigations are still missing. Principal Ideas/Results. We propose Fairness Hazard Analysis (FHA), an adaptation of hazard analysis methods from the safety-critical domain to analyse fairness in socio-technical processes. FHA is demonstrated through an AI-assisted hiring case and supported by HumAInFlow, a modelling and simulation platform. The approach is preliminarily evaluated through two focus groups. The feedback from participants highlights FHA’s usefulness for structured fairness analysis, the importance of diverse expertise, and the potential for deeper integration within HumAInFlow. Contribution. This work offers a novel method for integrating fairness into requirements analysis of socio-technical workflows, and provides an LLM-based tool to automate the analysis, marking a shift from bias detection to bias prevention with fairness-by-design.

Background. LLM outputs are highly sensitive to prompt formulation: small wording changes can substantially affect output quality. This matters in software engineering, where prompts guide requirements analysis, code generation, and artefact synthesis. Poor formulations yield unreliable artefacts, yet practitioners lack principled ways to assess a prompt before inference, making selection depend on costly LLM calls and trial-and-error refinement. Aims. We investigate whether measurable linguistic properties of prompts can predict LLM performance before inference, enabling low-cost prompt selection and refinement, validated on binary requirements classification targeting F1, F2, precision, and recall. Method. We generate 9,000 linguistically controlled prompt variants from 100 initial prompts by varying 30 linguistic metrics, evaluated with five open-source LLMs on 625 annotated requirements. Regression predictors are trained via stratified 10-fold cross-validation with permutation-based significance testing; feature importance analysis identifies cross-LLM and model-specific predictors. Results. Linguistic features significantly predict prompt performance across all targets (R2 in [0.38,0.42], q<0.05). Syntactic and morphosyntactic features drive most predictive signal; cross-LLM predictors include compound dependency distribution, conjunction density, and word/sentence length, reflecting sensitivity to domain vocabulary and complex structures. Conclusions. Results suggest practical implications for prompt engineering, including overlap between linguistic patterns that reduce LLM performance and those that increase human comprehension difficulty, and the irrelevance of lexical variety as a quality dimension. More broadly, linguistic profiling combined with standard regression provides an effective, interpretable, low-cost prior before costly optimisation pipelines.

Large language models (LLMs) are increasingly applied to requirements engineering (RE) tasks, yet the prompts guiding them are typically designed manually through trial and error, yielding inconsistent and suboptimal results. Automated prompt construction remains largely unexplored in RE, leaving its effectiveness unclear. To address this, we propose a lightweight Automatic Prompt Engineering approach, Backtracking APE (BT-APE), and apply it to requirements classification. We frame prompt design as an optimization problem, iteratively refining prompts via LLM-generated candidates, backtracking search, and dynamic example selection. Evaluating BT-APE on three benchmark datasets with five instruction-tuned LLMs, we compare it against four classical prompting baselines (zero-shot, few-shot, chain-of-thought, CoT+few-shot) and a state-of-the-art but resource-intensive APE baseline (PE2). BT-APE and PE2 achieve nearly identical accuracy, both substantially outperforming the classical baselines with large effect sizes; however, BT-APE imposes a far lighter computational footprint, consuming roughly 72

Background. Requirements inspection (RI) is a well-established practice for detecting potential defects in requirements artifacts early in the software lifecycle. Recent advances in large language models (LLMs) have stimulated interest in their potential to support requirements engineering (RE) tasks. However, empirical evidence on the effects of LLMs when used as collaborative assistants in human-performed RI remains scarce. Aims. We aim to investigate the impact of LLM support on human-performed RI, considering inspection effectiveness in terms of smell identification and severity classification (i.e., nocuous vs innocuous), as well as inspection duration. Method. We conducted a controlled crossover design experiment with 34 participants, who inspected textual specifications with and without LLM support, identifying and classifying requirements smells while recording inspection time. We analyzed the data using one Bayesian regression model per outcome variable, accounting for validity threats induced by the crossover design as well as covariates and mediators. Results. Results show that LLM support negatively affects smell detection accuracy but has no significant effect on smell classification or task duration. A learning effect is present across experimental periods, but reduced when RI is first performed with LLM support. Conclusions. Our findings provide empirical evidence that LLM support does not necessarily improve performance and may, instead, hinder it for novice inspectors. Moreover, the results suggest that learning RI with LLM-support from the beginning may slow down the skill acquisition process, implying threats for LLM-supported learning.

Context and motivation. Online user feedback is a valuable resource for requirements engineering, but its volume and noise make analysis difficult. Existing tools support individual feedback analysis tasks, but their capabilities are rarely integrated into end-to-end support. Problem. The lack of end-to-end integration limits the practical adoption of existing RE tools and makes it difficult to assess their real-world usefulness. Solution. To address this challenge, we present RITA, a tool that integrates lightweight open-source large language models into a unified workflow for feedback-driven RE. RITA supports automated request classification, non-functional requirement identification, and natural-language requirements specification generation from online feedback via a user-friendly interface, and integrates with Jira for seamless transfer of requirements specifications to development tools. Results and conclusions. RITA exploits previously evaluated LLM-based RE techniques to efficiently transform raw user feedback into requirements artefacts, helping bridge the gap between research and practice. A demonstration is available at: https://youtu.be/8meCLpwQWV8.

Language workbenches are tools that enable the definition, reuse, and composition of programming languages and their ecosystems, aiming to streamline language development. To facilitate their adoption by language designers, the comprehensibility of the language used to define other languages is an important aspect to evaluate. Moreover, considering that language workbenches are relatively new tools, user acceptance emerges as a crucial factor to be accounted for during their assessment. Current literature often neglects user-centred aspects like comprehensibility and acceptance in the assessment of this breed of tools. This paper addresses this gap through a family of experiments assessing Neverlang, a modular language workbench. The study adopts a tailored version of the Method Evaluation Model (MEM) to evaluate the comprehensibility of Neverlang's meta-language and programs, as well as user acceptance in terms of perceived ease of use, perceived usefulness, and intention to use. It also investigates the relationships among these dimensions. The experiments were conducted in three iterations involving participants from academia. The results reveal that users demonstrate sufficient comprehension of Neverlang's meta-language, particularly concerning its syntax, express a favourable perception of its usefulness, and indicate their intention to use it. However, the results also indicate that Neverlang's ease of use remains a challenge. Additionally, variations in the perceived ease of use and perceived usefulness, whether low or high, influence the users' intention to use the tool. Surprisingly, no significant correlation is found between comprehensibility and user acceptance. Notably, higher comprehensibility of the meta-language does not necessarily translate into greater acceptance, underscoring the complex interplay between comprehension and adoption.

Requirements-related information is scattered across heterogeneous artefacts such as user feedback, developer discussions, and software repositories, making the extraction of actionable requirements knowledge labour-intensive and hard to scale. Large Language Models (LLMs) can support many Requirements Engineering (RE) activities, from classification and traceability identification to specification and explanation generation, but existing evidence is fragmented across tasks, artefact types, and evaluation settings, and studies rarely offer cross-task evaluations or replication packages. We present two complementary empirical studies evaluating LLMs across five RE-related activities. The first is a controlled experiment on five lightweight open-source LLMs for feedback-driven requirements classification and specification generation. The second is an exploratory industrial case study on two frontier LLMs for traceability link identification and traceability explanation generation using real project artefacts. Classification and traceability identification were assessed with quantitative metrics, and generation tasks through human evaluation. LLM performance is strongly task-dependent, ranging from moderate to high, and no single model consistently outperformed the others, indicating that effective adoption depends on selecting models and prompting strategies per task. Our contributions are: (i) the first cross-task empirical evaluation of LLMs spanning five RE-related activities, (ii) replication materials supporting reproducibility, and (iii) a broader understanding of the capabilities, limitations, and practical readiness of current LLMs for RE.

The representation of requirements plays a critical role in the accuracy of requirements inspection. While visual representations, such as UML diagrams, are widely used alongside text-based requirements, their effectiveness in supporting inspection is still debated. Cognitive abilities, such as working memory and mental rotation skills, may also influence inspection accuracy. This study aims to evaluate whether the use of UML sequence diagrams alongside text-based requirements improves the accuracy of requirements inspection compared to text-based requirements alone and to explore whether cognitive abilities are associated with differences in performance across the two treatments (text vs text with UML support). We conducted a crossover experiment with 38 participants to assess the accuracy of requirements inspection under the two treatments in terms of issues found and justifications provided. Linear mixed-effects and generalized linear models were used to analyse the effects of treatment, period, sequence, and cognitive abilities. The results indicate a significant three-way interaction between representation type, working memory capacity, and mental rotation ability. This finding suggests that the effectiveness of UML support is not uniform across individuals: participants with high scores in both cognitive abilities experienced reduced performance when using UML for violation detection. Conversely, the same cognitive profile was associated with improved justification accuracy under UML-aided inspection, indicating that higher cognitive abilities may support deeper reasoning processes when dealing with multi-modal information, i.e., diagrams and text.

[Context and Motivation] Online user feedback provides valuable information to support requirements engineering (RE). However, analysing online user feedback is challenging due to its large volume and noise. Large language models (LLMs) show strong potential to automate this process and outperform previous techniques. They can also enable new tasks, such as generating requirements specification. However, existing work largely focuses on large proprietary models. Consequently, lightweight open-source LLMs remain underexplored for feedback-driven RE. [Question/Problem] In particular, existing studies offer limited empirical evidence, lack thorough evaluation, and rarely provide replication packages, undermining validity and reproducibility. [Principal Idea/Results] We evaluate five lightweight open-source LLMs on three RE tasks: NFR classification, user request classification, and requirements specification generation. Classification performance was measured on two feedback datasets, and specification quality via human evaluation. LLMs achieved moderate-to-high classification accuracy (F1 ≈ 0.47–0.68) and moderately high specification quality (mean ≈ 3/5). [Contributions] We newly explore lightweight open-source LLMs for feedback-driven requirements development. Our contributions are: (i) an empirical evaluation of lightweight LLMs on three RE tasks, (ii) a replication package, and (iii) insights into their capabilities and limitations for RE.

The emergence of Large Language Models (LLMs) has opened new opportunities to automate software engineering activities that traditionally require substantial manual effort. Among these, class diagram generation represents a critical yet resource-intensive phase in software design. This paper investigates the capabilities of state-of-the-art LLMs, including GPT-5, Claude Sonnet 4.0, Gemini 2.5 Flash Thinking, and Llama-3.1-8B-Instruct, to generate UML class diagrams from natural language requirements automatically. To evaluate the effectiveness and reliability of LLM-based model generation, we propose a comprehensive dual-validation framework that integrates an LLM-as-a-Judge methodology with human-in-the-loop assessment. Using eight heterogeneous datasets, we apply chain-of-thought prompting to extract domain entities, attributes, and associations, generating corresponding PlantUML representations. The resulting models are evaluated across five quality dimensions: completeness, correctness, conformance to standards, comprehensibility, and terminological alignment. Two independent LLM judges (Grok and Mistral) perform structured pairwise comparisons, and their judgments are further validated against expert evaluations. Our results demonstrate that LLMs can generate structurally coherent and semantically meaningful UML diagrams, achieving substantial alignment with human evaluators. The consistency observed between LLM-based and human-based assessments highlights the potential of LLMs not only as modeling assistants but also as reliable evaluators in automated requirements engineering workflows, offering practical insights into the capabilities and limitations of LLM-driven UML class diagram automation.

The paradigm of software self-evolution enables systems to autonomously extend and reconfigure their own capabilities during execution in response to technical specifications. Yet requests for new functionality often originate from end users and are rarely expressed in technical terms. As a result, developers must translate user needs into technical specifications before the system can evolve, delaying early validation of the requested functionality by preventing users from immediately observing the resulting behaviour. To address this gap, we present ReqEvolve, a runtime code generation system that enables user-driven self-evolution by accepting high-level user requests. The system integrates automatic requirements engineering (RE) and test-driven development (TDD) to transform these requests into executable functionality through clarification, specification decomposition, test generation, and runtime integration. We evaluate ReqEvolve on 72 software evolution cases across 18 projects against two baselines: SpecFix, an RE-focused code generation approach, and an ablation variant of our system. ReqEvolve achieves 89.2% Pass@1, outperforming SpecFix by 18.8% (p < 0.01, r = 0.79, large effect) and the ablation baseline by 32.6% (p < 0.001, r = 0.88, large effect). These results provide initial evidence that user-driven self-evolution is a viable paradigm for autonomously extending software capabilities from user requests, thereby accelerating requirements validation prior to developer verification.

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.

Requirements Engineering (RE) is a critical sub-field of Software Engineering (SE), involving activities to identify, specify, model, analyze, and validate system needs and constraints. RE plays a fundamental role in the SE process. With the current advances in AI, new opportunities for RE arise to ensure the development of trustworthy AI and to utilize AI to perform various RE-related activities more effectively and efficiently. The workshop on Multi-disciplinary, Open, and IntegRatEd RE (MO2RE) successfully addresses the underrepresentation of the RE within the SE community, raises awareness of RE's diverse aspects, and fosters collaboration. The second edition took place on April 27th, 2025, co-located with the 47th IEEE/ACM International Conference on Software Engineering (ICSE) in Ottawa, Canada.

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.

Large Language Models (LLMs) are finding applications in numerous domains, and Requirements Engineering (RE) is increasingly benefiting from their capabilities to assist with complex, language-intensive tasks. This paper presents a systematic literature review of 74 primary studies published between 2023 and 2024, examining how LLMs are being applied in RE. The study categorizes the literature according to several dimensions, including publication trends, RE activities, prompting strategies, and evaluation methods. Our findings indicate notable patterns, among which we observe substantial differences compared to previous works leveraging standard Natural Language Processing (NLP) techniques. Most of the studies focus on using LLMs for requirements elicitation and validation, rather than defect detection and classification, which were dominant in the past. Researchers have also broadened their focus and addressed novel tasks, e.g., test generation, exploring the integration of RE with other software engineering (SE) disciplines. Although requirements specifications remain the primary focus, other artifacts are increasingly considered, including issues from issue tracking systems, regulations, and technical manuals. The studies mostly rely on GPT-based models, and often use Zero-shot or Few-shot prompting. They are usually evaluated in controlled environments, with limited use in industry settings and limited integration in complex workflows. Our study outlines important future directions, such as leveraging the potential to expand the influence of RE in SE, exploring less-studied tasks, improving prompting methods, and testing in real-world environments. Our contribution also helps researchers and practitioners use LLMs more effectively in RE, by providing a list of identified tools leveraging LLMs for RE, as well as datasets.

[Context and motivation] Large language models (LLMs) show notable results in natural language processing (NLP) tasks for requirements engineering (RE). However, their use is compromised by high computational cost, data sharing risks, and dependence on external services. In contrast, small language models (SLMs) offer a lightweight, locally deployable alternative. [Question/problem] It remains unclear how well SLMs perform compared to LLMs in RE tasks in terms of accuracy. [Results] Our preliminary study compares eight models, including three LLMs and five SLMs, on requirements classification tasks using the PROMISE, PROMISE Reclass, and SecReq datasets. Our results show that although LLMs achieve an average F1 score of 2

Context and motivation: End-user development focuses on enabling non-professional programmers to create or extend software applications on their own. However, before beginning the development process, software engineering best practices recommend performing requirements engineering (RE) activities, including requirements modelling. Question/problem: There is limited research on how end-users can model system requirements. Principal ideas/results: In this experience report, we investigate the problem of end-user requirements modelling in an EU-funded project about agricultural digitalisation. Specifically, a team of agronomists was directly involved in the creation of UML, iStar, and BPMN diagrams to model the transformation of socio-technical processes in four different concrete scenarios. They followed a formalisation procedure proposed within an RE method designed to help stakeholders evaluate the impact of agricultural digitalisation. Starting from textual reports including a description of the process as-is and the process-to-be, they followed step-by-step guidelines for model creation. Contribution: This paper reports insights from the experience from the viewpoint of the agronomists and software engineers involved. We identify nine key lessons that highlight the added value of end-user requirements modelling for achieving a shared and in-depth understanding of the socio-technical processes under analysis.

Empirical formal methods (FMs) aim to ground the development, evaluation, and deployment of formal techniques in software engineering through systematic observation, data collection, and analysis. In the context of safety-critical domains such as railways, the need to balance mathematical rigor with practical feasibility makes empirical approaches particularly relevant. In this keynote summary, I will outline the scope and significance of empirical FMs, and present a retrospective on my experience applying these approaches in railway system development and verification. Drawing on a series of industrial and collaborative projects, I will discuss how empirical software engineering techniques such as surveys, systematic literature reviews, case studies, judgment studies, and tool evaluations have been used to assess the applicability, usability, and impact of FMs in practice. I will conclude with a roadmap for strengthening the empirical evidence base for FMs in railways, identifying research gaps and practical strategies to improve their integration into industrial processes, toolchains, and certification workflows.