The core of the conference program are the writers’ workshops where the accepted papers will be discussed. Additionally, we offer an Open Space format for the accepted focus group proposals. The conference also contains many social activities to foster social networking.
Main room for the conference is Room 128.
14:00 – 17:30
Registration
15:00 – 15:30
Coffee
15:30 – 16:30
Pattern-Writing Dojo
16:30 – 17:30
Welcoming Session and Demo Workshop
17:30 – 19:00
Name Game
19:00 – 20:00
Dinner
20:00 – 0:00
Bar, BoFs, Sauna
07:30 – 09:00
Breakfast
09:00 – 10:00
Daily Intro and Games
10:00 – 10:30
Focus Groups Intro
10:30 – 11:00
Coffee
11:00 – 12:45
Writer’s Workshop
12:45 – 13:45
Lunch
13:45 – 14:15
Games
14:15 – 15:45
Writer’s Workshop
15:45 – 16:15
Coffee
16:15 – 18:30
Focus Groups
18:30 – 19:00
Daily Summary
19:00 – 20:00
Dinner
20:00 – 22:00
Paint & Create
22:00 – 0:00
Bar, BoFs, Sauna
07:30 – 09:00
Breakfast
09:00 – 9:30
Daily Intro
09:30 – 11:00
Writer’s Workshop
11:00 – 11:15
Coffee
11:15 – 12:45
Writer’s Workshop
12:45 – 13:45
Lunch
13:45 – 14:45
Monastery Tour
(start at main entrance)
14:45 – 15:15
Games
15:15 – 16:45
Focus Groups
16:30 – 17:00
Coffee
17:00 – 18:30
Focus Groups
18:30 – 19:00
Daily Summary
19:00 – 19:30
Prepare for Banquet
19:30 – 22:00
Banquet
22:00 – 0:00
Bar, BoFs, Sauna
07:30 – 09:00
Breakfast
09:00 – 9:30
Daily Intro
09:30 – 11:00
Writer’s Workshop
11:00 – 11:15
Coffee
11:15 – 12:45
Writer’s Workshop
12:45 – 13:45
Lunch
13:45 – 15:15
Brewery Tour
14:45 – 15:45
Hillside Meeting
15:15 – 15:45
Games
15:30 – 16:00
Coffee
16:00 – 18:00
Open Space
18:00 – 19:00
Closing Session
19:00 – 20:00
Dinner
20:00 – 0:00
Bar, BoFs, Sauna
07:30 – 11:00
Breakfast and Farewell
This is where your paper is discussed during the conference. The workshops group together the authors of up to five papers that thematically fit together. Shortly before the conference begins, we will publish the paper-to workshop assignments. If you are an author, use this page to find which workshop you are assigned to. Non-authors, please use the same page to decide which group you would like to join throughout the conference. All participants should download, print and read all the papers in their group before the conference.
Below you can find here the full list of papers accepted for the workshops. Clicking on the arrows will reveal the abstracts.
This is where your paper is discussed during the conference. The workshops group together the authors of up to five papers that thematically fit together. Shortly before the conference begins, we will publish the paper-to workshop assignments. If you are an author, use this page to find which workshop you are assigned to. Non-authors, please use the same page to decide which group you would like to join throughout the conference. All participants should download, print and read all the papers in their group before the conference. A full list of workshops and corresponding papers will be published here once the program is finalized.
Abstract: The objective of this paper is to reframe how we view the tensions between forces in patterns. I claim that the traditional approach is to view tensions as either/or options. As an alternative, I propose to adopt the perspective of paradox theory and view tensions as both/and options. The two key approaches to both/and paradox resolution are integration and differentiation. Both approaches involve a circular process in which differences are either juxtaposed to form a whole (integration) or separated over time or space (differentiation). Examples include the tension between integrated and modular systems, competition and collaboration in open source projects, and co-located and remote teams. This paper is intended for pattern researchers, authors, and users.
Abstract: “[Silicon valley is] run by engineers who are remarkable people but … they feel they should have a fix and there should be an answer to everything, … I come from the world of politics where you are always dealing with the imperfections of the world and trying to both guide and ride all the changes in society … Engineers can be very dogmatic about the way things should be. I come in as a non-engineer and say hang on a minute, what you guys are inventing that you think is entirely logical is going to collide with these issues, politic, cultural, morale, ethical issues” Nick Clegg, Deputy Prime Minister of the United Kingdom, 2010-2015; Vice President Facebook/Meta 2018-2025 (Clegg, 2024)
Abstract: Patterns describe reusable solutions to frequently encountered problems within a given context. For those interested in patterns, pattern writing is a crucial process that unfolds in several stages and involves multiple feedback loops. These feedback loops typically include guidance from pattern-writing specialists, known as shepherds, as well as writers’ workshops organized at pattern-writing conferences. The style used to describe patterns varies widely. Some writers favor narrative descriptions, while others prefer more structured formats. Similarly, descriptions may range from simple to highly detailed. To support the pattern-writing process, the patterns community relies on well-established pattern forms that writers commonly use to document their patterns. In this paper, we focus on three well-known pattern forms introduced by Christopher Alexander, the Gang of Four (GoF), and James Coplien. To help pattern writers select an appropriate pattern form and adapt it to their needs, we analyze a large-scale corpus of articles published over the past 28 years in the two major pattern-writing conferences. We evaluate both the popularity of these forms and the degree to which authors adhere to them. The outcome of this study is a set of patterns that support the selection and adaptation of pattern forms.
Abstract: This paper explores how “pattern smells”—a concept inspired (and reimagined) by software engineering’s code smells—can support more effective feedback throughout the lifecycle of pattern mining, writing, and maintaining their descriptions. Pattern smells name and classify recurring weaknesses in patterns, ranging from misconceptions, insufficient domain grounding, and gaps in analysis, to overly optimistic or incomplete “utopian” claims and lack of wholeness. Rather than focusing on fixing individual smells, the paper shows how they can guide different forms of feedback, including self-reflection by authors, iterative shepherding, writers’ workshops, and AI-generated feedback. The discussion maps specific smells to appropriate stages of feedback on patterns as they are written: early stages emphasize identifying whether real, whole solutions exist; intermediate stages refine understanding of solution details, forces, evidence, and context; and later stages improve clarity, structure, and presentation. The authors argue that explicitly naming such flaws helps participants in the pattern writing and review process articulate vague concerns, improve communication, and strengthen both pattern descriptions and pattern languages. Overall, pattern smells provide a practical framework for diagnosing weaknesses and improving the quality and communicability of patterns.
Abstract: This paper introduces the concept of Joy Patterns, which are among the elements that constitute each person’s Self, and explores a new attempt to articulate them as a pattern language. Half a century ago, Christopher Alexander identified recurrent patterns within ‘space + events’ in the external environment, through which a high quality of life can repeatedly emerge in everyday life. He then formulated these patterns as a pattern language of ‘Patterns of Space.’ Half a century later, we propose in this paper an analogous approach: identifying recurrent “patterns” within ‘self + events’ in the internal environment, through which a high quality of life can repeatedly emerge in everyday life, and articulating them as a pattern language of Patterns of Self. Patterns of Self may be understood as comprising such patterns as Joy Patterns, Sadness Patterns, and Anger Patterns. In this paper, we focus specifically on Joy Patterns and explore how they can be articulated. Joy Patterns recur in each person’s life and contribute to the generation of joy. Although such patterns are personal to each individual, similar patterns are often shared among different people; in this sense, they can also be regarded as “patterns” that may be shared beyond the individual. This paper explains five steps for discovering Joy Patterns within Patterns of Self: (1) Attraction Focusing, (2) Roots Mining, (3) Joy Pattern Seed Identification, (4) Recalling Seed-Embodied Experiences, and (5) Joy Pattern Identification & Articulation. Using the two authors’ own cases as examples, we introduce the process through which the Joy Patterns ‘Unraveling into Creation’ and ‘Willing Contribution to a Comfortable Community’ were identified. Finally, we examine Patterns of Self in light of Alexander’s theory and conclude that they can indeed be understood as patterns, insofar as they generate the “quality without a name” by resolving conflicts among forces.
Abstract: Over the past 10 years there has been a growing awareness among software business and technology leaders that a product focus (a longer-term view) instead of a project-based focus (concerned with short-term delivery of results on time and within budget) can positively impact organizations, their software development processes, and business outcomes. In this paper we explore whether considering software pattern collections and languages as products might lead to radically different approaches to developing and maintaining them over their lifetime. We consider several questions: What exactly is the value proposition for pattern collections and languages; what value do different consumers place on collections and languages; and how might we increase this value.
Abstract: This study investigates the effectiveness of three established design patterns for online education: Embracing the Virtual Class, Structured for Self-Study, and Interactive eLearning. As online learning continues to expand alongside traditional in-person instruction, educators face ongoing challenges in adapting teaching methods to environments with limited real-time interaction. These patterns emphasize asynchronous engagement, well-organized self-paced materials, and interactive learning systems to enhance student engagement and learning outcomes. Using Pathway2Code, an interactive coding education platform, this study evaluated the applicability and impact of these patterns by having students review and analyze a beta version of the site within a shared learning environment. The findings aimed to assess the strengths and limitations of these patterns and to contribute to the development of more effective and adaptable online course designs for modern learners. Overall, this paper demonstrates how established instructional design patterns function in practice within asynchronous courses and evaluates their strengths, limitations, and applicability through a case study of the Pathway2Code platform. By examining their implementation, the study provides actionable insights for designing more effective and adaptable online learning experiences.
Abstract: Misconceptions pose a significant hurdle in computing education, often stalling a student’s progress in learning complex programming concepts. Identifying these mental hurdles is crucial for effective instruction, yet researchers often struggle to select the appropriate diagnostic methodology for their specific research context. This paper introduces a collection of five methodological patterns designed to systematize the identification of student misconceptions. Going beyond mere methodological descriptions, these patterns provide actionable guidance on selecting approaches tailored to specific research settings and resource constraints. Grounded in existing computing education literature, this collection serves as both a practical guide for methodology selection and a repository of best practices to conduct research that can ultimately be foundational to improve student learning outcomes.
Abstract: In this paper we continue our work on the pattern language regarding student projects in higher education with agile aspects. Our previous work focused on the team and project setup stage of the courses running these projects. This paper targets the patterns for guidance and support that the teaching staff may apply during the projects’ execution. We present the following four patterns in this area: The Mentor-Assessor Role, Big Brother, Early Meetings Crutch, and Check Customer’s Privilege.
Abstract: This paper aims to provide action perspectives for teachers in higher education, especially in an agile software engineering context, who supervise project groups. Sometimes, the usual process of guidance, in combination with retrospectives, seems to be coming to a dead end. The teacher knows that if nothing changes, the team could fail the project, so they have to come up with something different. At the same time, they must not take over the project. In this paper, we show two patterns and five known uses that help teachers take action in a way that gives students agency over their project again. The first pattern helps the teacher decide if to intervene at all. The second focuses on how to intervene in a way that the team keep agency on the course of the project. The known uses show five frameworks that can be applied when the team either defines incoherent tasks, does not divide work equally between the team members, is sloppy in decision making, works in unintended sub-teams or has problems with giving feedback on behavior.
Abstract: Kobanashi are short comical stories in traditional Japanese rakugo storytelling performances. They are performed to warm up the audience before telling the main story. In recent years, kobanashi has become increasingly popular as a tool and catalyst in heterogeneous educational settings, for instance in Japanese language education. Providing opportunities for learners to practice and perform kobanashi themselves supports the creation and modification of interpersonal relationships and increases self-confidence.
To further promote kobanashi, we aim to create a pattern language that facilitates effective implementation and supports good practice, enabling educators to harness its full potential, not only in language education but also in other educational fields. In this paper, we propose two of a series of patterns for effective implementation of kobanashi in educational contexts. The first pattern provides criteria for selecting appropriate kobanashi stories for learners, or for allowing learners to make their own selections while minimising the risk of negative experiences related to cultural taboos or personal backgrounds. The second pattern offers a guidline for introducing a lesson using kobanashi.
Abstract: This study examines what beginners come to grasp through actual practice in Pattern Mining, as part of the Holistic Way of Pattern Language Crafting (Iba, 2024) — that is, what cannot be fully understood through methodological explanation alone, but is realized only through experiential engagement. The analysis draws on Reflection Entries written by student members of Iba Lab at Keio University between 2021 and 2025, in which they recorded their experiences after engaging in Pattern Mining activities. We analyzed the entries with respect to three core practices: Mining Interviews, Clustering, and Systematization. The analysis of Mining Interviews revealed insights into co-exploratory dialogue, attending to the contrast between good and poor states, the importance of prior knowledge, and the cycle of practice and reflection. The analysis of Clustering revealed insights into structural similarity, the quality of preceding materials, the effect of vocalization, intuitive judgment, and trust in the process. The analysis of Systematization revealed insights into the limits of bottom-up thinking, the integration of top-down and bottom-up approaches, the value of repeatedly returning to concrete examples, and aesthetic coherence.
Abstract: STEM education relies heavily on visual artifacts such as diagrams, circuit schematics, plots, tables, formulas, and graphical simulation tools. These artifacts often remain difficult or impossible to access for BVI students when they are only provided as visual representations. At the same time, long and densely structured academic materials can impose additional barriers for learners with impairments such as attention-deficit/hyperactivity disorder (ADHD) or dyslexia, who may benefit from segmented and multimodal access. This paper extends the EuroPLoP 2025 pattern collection A Pattern Collection for Generating Accessible Teaching Materials for Blind and Visually Impaired Students in Computer Science and Electrical Engineering. The earlier collection focused on transforming visual STEM artifacts into accessible textual, tactile, or auditory representations. We add two AI-mediated patterns that address accessibility during actual learning situations rather than only during material preparation. Accessibility-Aware AI Tutor describes how to provide persistent, course-specific, and accessibility-aware support that is grounded in curated course materials, explicit accessibility rules, and repeatable transformation workflows. Structured Audio Companion describes how to provide segmented, source-grounded audio orientation with transcripts, timecodes, and explicit pointers to precise accessible representations of visual artifacts. Together, the two patterns show that generative artificial intelligence can support accessibility when it is embedded in course context, stable workflows, and human-in-the-loop quality assurance. AI-generated explanations and audio summaries should not replace precise accessible representations of visual STEM artifacts; rather, they should help learners find, understand, and use them.
Abstract: The Expert Blind Spot presents a challenge in higher education, where educators’ automated and implicit knowledge leads to gaps in student learning. As expertise becomes internalized, individual mental steps required for complex tasks are often not communicated. This paper presents Applied Cognitive Task Analysis (ACTA) as a design pattern for educators and didactic professionals to systematically externalize this hidden knowledge. Based on Cognitive Task Analysis, ACTA provides a streamlined three-phase interview protocol to rapidly elicit cognitive demands, expert strategies, and typical novice errors. To demonstrate ACTAs applicability, a worked example, set in the context of computer science education, applies the full ACTA protocol to the topic of inheritance in introductory programming, producing a cognitive demands table that surfaces expert strategies and common misconceptions. The paper discusses the theoretical implications of ACTAs application for Pedagogical Content Knowledge development and cognitive load reduction in instructional design.
Abstract: Hybrid teaching requires learning environments that connect physical campus spaces with digital participation and distributed collaboration. Based on literature analysis, university visits, and pilot implementations at TH Köln, recurring spatial, technical, and pedagogical practices were identified and formulated as reusable design patterns. This paper describes COMMUNICATING GROUP LOCATIONS, SYNCHRONOUS–ASYNCHRONOUS ORCHESTRATION and VIDEO CONFERENCING ON CAMPUS. While all three practices have been observed frequently across various campuses, they have not been generalized as patterns up to this point.
Abstract: This paper presents six patterns that articulate key practices for guesthouses to cultivate ongoing relationships with local communities. The guesthouses in focus are a form of small-scale lodging — also known as hostels or backpackers — where connections with local residents arise naturally in the course of daily operations, and where the guesthouse and the local community mutually shape one another over time. The six patterns — A Role Born From Each Person, A Place Where Engagement Endures, Regional Collaboration Partner, Creating What Is Needed but Not Yet There, Local Resource Circulation Model, and Future Shopkeepers — were developed through interviews with ten practitioners across Japan, and are drawn from a broader pattern language of 36 patterns titled How to Grow Guesthouses with the Community. These patterns are intended not only for guesthouse operators but for all practitioners engaged in community-rooted place-making, offering perspectives to support ongoing relationship-building.
Abstract: Knowledge workers in collaborative settings — engineering managers, consultants, coaches, teachers, organizational developers — make consequential diagnostic determinations every day. The standard psychologicaldiagnostic apparatus was built for individual, often clinical, casework and does not scale to this everyday layer of work. Between informal intuition and full clinical diagnostics there is a wide, populated middle layer that has been methodologically underserved. AI-based language analysis dissolves the historical bottleneck: detecting recurring patterns across multiple bounded interaction episodes is now feasible at the cost and timescale of an ordinary engagement. This paper presents a pattern language of five patterns for AIsupported diagnostic work, mapped one-to-one onto the established diagnostic process formulated by Kubinger. Each pattern addresses the specific shift that LLM-based pattern detection introduces at one step of the process. A worked case — a team conflict diagnosed by an engineering manager across twelve sprint-planning transcripts — runs through the paper as a continuous example. The discussion marks the structural gap the application makes visible: a testtheoretic foundation for findings whose unit is a configuration-bound recurring pattern rather than an individual score.
Abstract: Business process management (BPM) systems and corresponding notations have shown their applicability and flexibility in modeling, automating, and mining business processes in enterprise applications and service-oriented architectures for many decades. Becoming increasingly open and accessible via software-based interfaces, cyberphysical systems (CPS)—systems interacting with the physical world via sensors and actuators—might benefit from BPM-based automation too, as they are inherently process-driven and software-intensive. We present four patterns discussing the choice of appropriate granularity levels (i.e., the level of detail) and corresponding decision forces regarding what to model in executable business processes and what to implement in traditional programming languages in CPS software. Two patterns give advice on when to model sensing in CPS at a more fine-grained or on a more coarse-grained level in business processes. Similarly, two patterns discuss these granularity choices for actuation in CPS. These four patterns guide the modeling and implementation of executable business processes that can be used for combining and automating sensing and actuation of CPS.
Abstract: Machine Learning Operations (MLOps) systems degrade when data distributions shift, even if code remains unchanged. This probabilistic drift exposes divergent stakeholder expectations about system behavior, creating coordination failures that exceed those in traditional software delivery. To address these challenges, this research proposes six pattern candidates for institutionalizing stakeholder engagement throughout the MLOps lifecycle. By identifying ten distinct stakeholder roles, we systematically map their activities across six continuous MLOps phases to clarify organizational responsibilities. Based on empirical evidence and initial expert validation, we formulate solutions as pattern candidates: Stakeholder Engagement Plan, Proactive Privacy Impact Assessment, Feedback Decision Record, Human-AI Feature Synthesis Interface, Domain Experts as Teachers, and Participatory Co-creation. Formalizing stakeholder engagement as a first-class design concern helps practitioners mitigate drift and bridge the socio-technical gap in ML governance.
Abstract: In many cases, there is a bottleneck in addressing customer requests because they represent very specific needs that not many people can solve. That can make users feel powerless to solve their problems and force them to wait a long time. This paper presents a pattern called the Problem-Inversion Principle, which aims to empower users to have autonomy to solve problems that depend on who provides the services or products. Focused on the software domain, this pattern addresses several aspects to achieve this goal, such as an extensible structure, the required knowledge and access to required resources. As expected consequences, user can increase their autonomy, and the approach used to support the maintenance and evolution of the software system can become more scalable to meet the needs of more users. On the other hand, we also discuss potential liabilities, such as the migration effort and concerns about the quality of delegating part of the solution. Additionally, this paper examines the importance of this pattern for software startups, the role of recent AI technologies in implementing it, and its application in nearshore and offshore strategies.
Abstract: While partially reflecting valid complaints, the frustrations business people express with engineers, and those that are expressed in return, are underpinned by a skewed understanding of what matters to their counterparties. Rooted in the world of data, this pattern language revisits the context of running a business, the forces driving decisions, and argues that the business environment is significantly more disorderly than engineers might prefer, exerting a price for them to pay. Illustrating designs that sit comfortably with such preconditions, it is shown that underlying these one can find patterns in their own right that are flexible, efficient, and valuable.
Abstract: Since the public emergence of conversational generative AI (GenAI) in late 2022, knowledge workers have been confronted with recurring expecta-tions of rapid substitution. This is accompanied by organizational pressure to re-alize productivity gains through rapid chatbot adoption. Yet in many cases these expectations have resulted in fragmented automation efforts and limited perfor-mance gains. This paper reframes AI transformation from a narrative of job au-tomation to one of job augmentation. Drawing on well-established job crafting theories, it proposes a practical lens to redesign knowledge work under uncer-tainty by addressing knowledge tasks within human cognitive boundaries. The paper introduces a pattern language in the form of patlets that support the recraft-ing of job profiles through outcome tracing, boundary clarification, and explicit function allocation between human and chatbot. The patlets guide users from de-hyping and misconception correction to the design of repeatable augmentation loops and quality gates that preserve human accountability. Two usage modes are outlined—rapid stabilization and deeper redesign—together with initial ap-plication scenarios and key risks, including over-delegation, accountability am-biguity, and compliance violations. The contribution is a proposal for a repeata-ble, human-centered method for sustainable human–AI job redesign in regulated knowledge-work environments.
Abstract: Agentic AI in regulated organizations fails less on output quality than on governable delegation: once agentic systems can plan and perform or trigger externally effective actions through enterprise tools and systems, auton-omy must remain bounded, attributable, and reconstructable through evidence by design. This paper presents a practical pattern language for turning agent auton-omy into accountable delegation through Delegation Contracts, Risk-Banded Au-tonomy, Execution Stage Gates, Stop Conditions, and an Agent Capability Owner. Together, these patterns make delegated autonomy reviewable before consequential action, defensible under uncertainty, and explainable over time. The paper also specifies the anti-pattern Shadow Delegation and a stepwise Re-pair Path that restores governability without collapsing utility. The contribution addresses decision-makers and practitioners in regulated sectors, especially audit, banking, and insurance who need scalable guidance for the responsible adoption of agentic capabilities under compliance constraints. A lightweight evaluation is provided through governability tests covering delegation boundaries, attribution, least privilege, pre-execution reviewability, and uncertainty handling.
Abstract: Legacy software often comes with technical debt that was built up over the years. That debt could be a lot of manual tests which all have to be executed for releasing new versions of the software and which take a lot eort to execute. It is dicult to nd the time for improving and automating the tests while being busy with executing them. This paper shows guidance on how to escape that tricky situation. The paper presents patterns on how to deal with the manual tests by tackling them bit by bit.
Abstract: AI coding agents such as Codex, GitHub Copilot, and Claude Code now operate in tens of thousands of software repositories. Practitioners report that the usefulness of these agents depends critically on the project-specific written instructions they receive [3]. Across the opensource ecosystem a convention has emerged: a structured instruction file, commonly named AGENTS.md, that shapes agent behaviour for a user, project, or folder. Analysing instruction files from repositories spanning web development, mobile engineering, database infrastructure, and platform tooling reveals recurring structural choices that address the same tensions practitioners face when briefing human collaborators: identity, boundaries, priorities, and external knowledge. This paper presents four patterns drawn from this analysis: Persona Frame establishes the agent’s identity and values; Boundary Catalog enumerates explicit do/don’t rules; Ranked Playbook introduces priority tiers and operating modes; and Knowledge Signpost links to external documentation without bloating the instruction file. The pattern where extracted from software engineering projects but should be applicable to all domains that can utilise AI agents.
Abstract: Real-time sensor-based AI pipelines are increasingly critical in industrial automation. High feedback latency remains a significant barrier to their effective deployment. System architects designing these pipelines often lack a structured design vocabulary to navigate the tradeoffs between inference speed, predictive accuracy, and infrastructure cost. This paper partitions end-to-end pipeline latency into measurable components and evaluates 12 optimization variables through controlled experiments on a YOLO-based inference testbed. These variables include protocol selection, model tiering, and hardware scaling. The study identifies critical bottlenecks in data transmission and model processing. We synthesize these findings into four Pattern Candidates: BottleneckOriented Investment, Model Size Tiering, Protocol Switching, and Resource Elasticity. These candidates provide architects with reusable decision guidance for latency optimization in sensor-based AI pipelines.
Abstract: Governance architectures for decentralized autonomous organizations (DAOs) increasingly blend on-chain transparency with off-chain flexibility to balance security, efficiency, and community participation. In this paper we collect 7 patterns covering the entire lifecycle of DAO proposal processes. The patterns include how decisions concerning proposal approvals are made (on-chain, off-chain, hybrid); how the proposal execution is safely controlled (timelock execution, multi-sig execution, emergency track); how proposals are challenged during review (proposal challenge); what are the incentives to reward reviewing potentially malicious proposals, while penalizing the submission of spam and malicious proposals (proposal bond).
Abstract: Digital twins are increasingly used as operational models of complex technical systems. In commissioning workflows they can serve as test benches before the physical system is available and as computational models for behaviours that cannot be observed directly. This paper presents three patterns for commissioning complex systems supported by digital twins. Relative Execution separates an experiment from the working point in which it is interpreted. Proposer–Policy Optimizer separates proposing actions from approving actions. Structured Read Access ensures that observations remain meaningful after transport and analysis.
Together, these patterns extend the interaction language introduced in our previous work into a commissioning language that supports observation, experimentation, and adaptation around a working point.
Abstract: Generative AI (GenAI) has had a major impact on software engineering. New tools are changing the field by using specialized, prompt-driven agents to take over many tasks traditionally performed by human developers. In this paper, we introduce a framework we call shift-up for GenAI-native development. The framework helps whole organizations focus on high-value work while relying on GenAI for support in the development. The shift-up framework is defined through a set of patterns that form the core of this paper. Towards the end of the paper, we also discuss how these patterns and the underlying framework can be evaluated.
Abstract: Serverless computing, or the Function-as-a-Service (FaaS) service model, has become a popular approach for implementing cloud-native applications. It allows developers to focus on domain and business logic without the need to manage infrastructure or even application packaging: developers simply provide function code. Another key advantage of serverless is that cloud customers only pay for the actual execution time of functions. The environments that execute these functions are shut down when not in use. This optimizes cost and resource consumption, as consumers do not have to pay for idle time of running environments. However, this approach introduces new challenges, such as the well-known cold start problem, which occurs when a function is triggered, but no execution environment is ready, leading to delayed responses. Additionally, costs depend on how long a function runs and how much memory it consumes. Depending on the use case, functions, their execution time and memory consumption, as well as their scaling behavior, must be optimized in dierent ways. This paper presents an initial set of ve serverless optimization patterns mined from industry cloud providers: Scale to Zero, Warm Pool, Snapshot Start, Delayed/Lazy Initialization, and Runtime Slimming. These patterns highlight strategies to optimize functions with regards to cold start latency and which use cases benet from specic serverless setups, congurations, and scaling strategies. The pattern collection may be extended with further optimization patterns related to right-sizing, abstraction levels, and granularity in future work.
Abstract: Industrial production systems generate vast amounts of data and require fast response to anomalies, yet expertise is scarce: a single specialist often oversees dozens of machines while the operator at the equipment lacks the diagnostic knowledge to act. The classical Blackboard Pattern coordinates heterogeneous experts around a shared workspace, but its knowledge sources are static, its controller is a fixed scheduler, and resolved incidents leave no organisational trace. This paper presents a small pattern language that updates the Blackboard for an environment in which Large Language Models can read documentation, reason about evidence, and learn from outcomes. The umbrella pattern, LLMdriven Blackboard, coordinates three individually applicable sub-patterns: Diagnostic Knowledge Source Taxonomy, Agentic Diagnostic Workflow, and Diagnostic Memory Management. We document forces, consequences, and liabilities of each, and report one known use in Surface Mount Technology manufacturing at Siemens, where a recurring vision-system anomaly was resolved by an operator instead of a technologist. The paper is intended as a reference for architects building diagnostic systems.
Abstract: The first line of defense to prevent fraudulent transactions is often implemented as automated systems. To process high-volume transaction streams these systems must be fast and scalable while staying adaptable and explainable. Rule-based approaches and single-paradigm machine-learning methods struggle when fraud patterns evolve continuously: supervised learning automates deci-sions but depends on label quality and lags emerging behavior, while unsuper-vised learning can cluster new patterns but cannot directly automate decisions.
This paper presents CLUSTER AND CATCH, a Pattern that combines supervised decision making with unsupervised clustering in a feedback loop: the supervised model handles real-time decisions, clustering highlights disagreements and novel behavior for efficient human review, and reviewer judgments improve labels for retraining.
The paper documents the context, forces, and consequences of this approach and summarizes applicability through known industry use cases.
Abstract: Combining a branchless configuration management, a trunk, with a single com-mon repository makes for a large monolithic code base. This set of patterns explores the balance of separation that is needed for comple-tion of tasks, and the integration that is mandatory for the task to become valuable within a larger system. This balance is a meeting point of engineering and man-agement. Getting it right is a key factor to complexity and risk management, and to the ability to deliver the software to its users.
This paper is part of a small series treating different topics, aspects, views, and roles needed with a monolithic code base. In this episode, we look at the value system shift that the organization faces when supporting a monolith: velocity over caution. And we supply the patterns that treat ownership and release. They base on the patterns of recent episodes on workflows, and on applicability checks.
Abstract: Automating Business Processes using work ow engines is a common task in digitization projects. However, many practical problems arise not by implementing the process- ow but the data- ow. Thus, design patterns have been mined from several industry projects to help process designers and software developers make better informed design decisions. Design patterns capture proven solutions to recurring design problems in a form that supports communication and reuse. This paper presents the rst ve design patterns covering design decisions relating to the contents of variables in executable business processes. All design patterns are presented with detailed forces as well as BPMN examples.
Abstract: Continuous Integration and Deployment (CI/CD) pipelines automate the software delivery lifecycle, making integration of changes easier, releases more consistent, and speeding up feedback from the stakeholders. This ultimately makes CI/CD pipelines an important element of modern software development. While reliability has been shown to limit the applicability and trustworthiness of CI/CD pipelines, existing works have not studied it in detail, focusing instead on generic best practices or other qualities such as speed. To address this gap, we analyze 32 grey literature sources that explore practitioners’ views and present 7 patterns that improve the reliability of CI/CD pipelines. The patterns deal with, among other things, reducing inconsistencies across pipeline runs, reducing sources of failures such as downloading and installing dependencies, and putting in place explicit resource and concurrency checks. Then, we analyze mature open-source repositories to identify evidence of the application of our patterns. Our findings indicate a strong contrast in the presence of these patterns. Atomicity and idempotency are often applied by developers, while parity between environments and pinning dependency versions are not.
Abstract: This paper presents the Branding Design Patterns, a pattern language for fostering growth by building brands through design. This pattern language was developed by analyzing the practices of Manabu Mizuno of good design company, who has been engaged in branding design that captures the unique character of clients companies and organizations and conveys it compellingly through a distinctive worldview, and by articulating the underlying essence of those practices. The results reveal that, in branding design, it is fundamentally important to shape the brand’s direction so that it becomes “clothes” that suit the client, to craft the design so that the brand becomes a beloved presence, and to accompany the process so that the brand can grow steadily over time. Furthermore, these principles have been systematized into a set of 27 patterns that show how they can be put into practice. In this paper, we present twelve patterns selected from the full set of 27. Specifically, as patterns for DESIGNING “CLOTHES” THAT FIT, we introduce Core Aspiration, A Greater Cause, What the Roots Reveal, and Story from the Past into the Future. As patterns for BECOMING A BELOVED PRESENCE, we introduce People with Hearts and Minds, Emotional Moment of Encounter, A World in a Mark, and Worldbuilding in Every Detail. As patterns for GROWING OVER TIME, we introduce Growth Companion, The Guiding Concept, Sense-Building Field Visits, and Envisioned Reality. In the final part of the paper, we further examine three companies of different types in order to show what kinds of designs were actually created through the practice of these patterns in real corporate branding projects, how those designs were received by the clients, and the extent to which they have in fact led to growth.
Abstract: Christopher Alexander’s pattern language inspired the methodological practice of using patterns to enhance the game design process. Björk and Holopainen popularized the application of pattern languages to games in 2005, and their work established a strong foundation built on by a number of scholars who have applied their principles and philosophy to many interdisciplinary design inquiries. Patterns drive game design research in diverse domains such as player experience, affective games, and serious games–nuanced and complex areas of game design that demand more rigorous methods to solve problems within. Despite the scattered local popularity of using pattern languages to solve ambitious design problems in games, widespread adoption in academia and industry remains elusive. We argue that one reason for this is the fragmentary nature of these studies. Through tracing and consolidating the evolution of pattern languages and the processes used to create them, both narratively and methodologically, in game design, we identify underlying strengths and weaknesses in these approaches and the patterns they produce. We are then able to propose concrete steps and further research that will help address these issues. We propose that removing these barriers to adoption will stimulate the use of pattern languages that address tougher and more diverse design problems. As a result, more researchers and designers will have access to pragmatic design insights and the pedagogical benefits of studying and applying pattern languages. We conclude by speculating on constructive areas for future research and invite game designers to direct their research toward centralizing pattern collections to make them more viable for general use.
Abstract: Unlike human-computer interaction, the interaction between artificial intelligence and humans requires to consider agency, trust and ethical implications. As humans and AI agents collaborate on a broad variety of tasks, both play an active role, initiating the interaction, and both can independently or together contribute to the task outcome. In this paper we collect ten patterns spanning across the design space of Human-AI interaction structures: Centaur, Reverse Centaur, Vibe Coder, Human in the Loop, Human partially in the Loop, Cognitive Offloading, Second Opinion, Supervisor, Reverse Supervisor, and Assistant.
Abstract: In contemporary society, where “wicked problems” require broad participation beyond professional designers, many non-designer-practitioner engaged in addressing social issues operate without access to expert facilitation. However, from the practitioner’s perspective, the key to tackling “Social issues” has not been articulated. To support such contexts, we developed an original Pattern Language for Social Service Design, consisting of 30 patterns derived from interviews with seven service designers engaged in social design practice. The development followed Iba’s pattern language methodology and aimed to articulate practitioners’ tacit knowledge as intermediatelevel knowledge accessible to non-experts. The resulting pattern set was analysed in two ways: (1) comparison with the seven-step framework presented in Social Design Practice Guide (Kakei, 2013), and (2) comparison with an existing Pattern Language for Design Activities that is not domain-specific. The analysis suggests that social design practice places distinctive emphasis on research depth and articulation of the meaning of action (visioning), as well as on mindset and outreach.This study contributes (1) a bottom-up conceptualisation of social design grounded in practitioners’ knowledge, and (2) provides a preliminary step toward Research through Design (RtD) method for extracting practice-level distinctions across related design domains.
Abstract: Organizations increasingly introduce digital tools under strong time pressure, strategic expectations, and growing complexity. While such tools prom-ise efficiency gains, early-stage rollouts often require users to work with evolving workflows, incomplete guidance, and unclear interaction patterns. Over repeated rollouts, these experiences can contribute to tool adoption fatigue, where employ-ees approach new technologies with reduced trust, lower enthusiasm, and limited willingness to experiment. This paper addresses tool adoption fatigue as a socio-technical challenge and investigates how structured conversational approaches using Generative AI (GenAI) can bridge the gap between tool potential and lived operational reality. It introduces two complementary patterns. The Conversa-tional Compensation Pattern supports users during early-stage tool use by apply-ing a prompt pattern language that guides them from task understanding to vali-dation, helping them navigate immature or complex tools through structured di-alogue. The Narrative Calibration Pattern supports trainers and adoption man-agers in preparing realistic, problem-based adoption stories that align expecta-tions with operational constraints. Together, these patterns form a Conversational Adoption Support approach that stabilizes both usage and expectations during tool rollouts. Rather than positioning GenAI as a replacement for tools, trainers, or change management, this paper frames it as a conversational mediation layer that helps organizations design interaction, reduce uncertainty, and foster more resilient adoption processes.
Abstract: Agile methods present a lower concern on documentation, as shown by the use of user stories, lightweight artifacts to represent requirements. Given the functional-oriented nature of these artifacts, nonfunctional requirements (NFRs) are neglected in these contexts. Despite proposals for specific artifacts to handle NFRs in agile projects, research has shown that practitioners prefer to adapt existing practices, such as user stories. However, these practices have not been formalized nor organized for their systematic reuse. In this paper, we close this gap by presenting a set of patterns for handling NFRs in agile contexts based on the results of our previous empirical studies on the topic. The mapped patterns are Tasks and Acceptance Criteria, Story Labeling, Dedicated Documents, Story Sub-sections, and Verification Rules.
These sessions allow you to dive into emerging topics or practice new skills. They are less about critique and more about collaborative discovery where you might brainstorm a brand new pattern language or experiment with a different technology. You can use the following description to decide which sessions to attend, but you can also leave that decision for the conference itself. The individual sessions will also be introduced by their organizers at the Focus Groups Intro slot.
If there is a recommended (but not mandatory) reading as preparation to participate in any particular Focus Group, it will be linked in the list below.
Abstract: What is it that truly attracts you to pattern languages?
Reflect on which aspects of pattern languages you are drawn to. Then, trace back to formative childhood experiences that may lie at the roots of those attractions. You may begin to uncover your own “Joy Patterns” – fundamental values and sources of joy that have repeatedly appeared throughout your life.
Then revisit pattern languages from a deeper perspective and articulate what you genuinely appreciate about them in connection with your own personal values.
Experience a process of deeply verbalizing what attracts you to pattern languages by using the approaches we have independently developed: “Roots Mining” and “Joy Patterns.”
The ideas and methods explored in this session are connected to papers submitted to this year’s EuroPLoP and also include new concepts we are currently developing. This Focus Group will be an opportunity to experience these emerging ideas together, discuss them collaboratively, and discover new insights.
Abstract: What if your patterns didn’t just get read—but truly lived?
We often share pattern languages through papers. It works—but does it really reach the people we hope to inspire? Could there be a more engaging, intuitive, and memorable way to communicate what patterns truly mean?
Patterns describe how people design and act in specific situations. And when we think about it, there is a natural form that brings such lived practice to life: story.
When patterns are woven into narratives—when characters actually live out those patterns—they become vivid, relatable, and emotionally engaging. Instead of being explained, they are experienced. This opens the door to reaching a much wider audience in ways that are both accessible and inspiring.
In this workshop, we invite you to explore that possibility with us. Drawing on our three-year journey of creating dozens of “Practice-Inspiring Stories,” we will share what we have learned about transforming patterns into compelling narratives that people can connect with and learn from.
Abstract: This Focus Group questions the conventional way in which patterns in a pattern language are typically used. Usually, we attempt to apply individual patterns to our own practices in order to improve what we are already doing. We have rarely questioned this way of utilizing patterns. Recently, however, I have come to think that such an approach may not be appropriate.
The idea of improving something through the additive composition of parts is closer to Christopher Alexander’s early thinking. In his later years, however, Alexander placed wholeness at the center of his theory and introduced the concept of centers within a living whole. The idea of partially and additively introducing patterns in order to make an existing design “better” remains at the stage that Alexander himself eventually criticized and sought to transcend. Then, how should we proceed?
We should begin from the great whole that we truly want to create, and continually keep that whole as the central focus. What matters is to genuinely aspire to create the great that deeply touches people, to explore how such greatness can be realized, and to shape it until its substance is fully brought to life.
Abstract: A dark pattern is a deliberately deceptive design choice in software or digital systems that nudges users into actions that benefit the provider at the user’s expense, often by exploiting cognitive biases or obscuring meaningful choices. “Dark Patterns – the Game” aims to raise awareness of these practices by creating a setting where players can discuss them in a fun way.
Abstract: As a patterns community, we strive to provide quality patterns and pattern collection to the benefit of those who apply the practices. We deeply believe that these patterns are of value to those who use them. We would like to gain a better understanding of this value.
In a similar way, the patterns community itself has a shared value system that has been carefully developed over many years. We would also like to gain a better understanding of how our community members see and experience these values.
The purpose of this event is to collect input from conference attendees about their personal relationship to patterns.
Abstract: As we all know, a fish in the water is not aware of its surroundings. How culturally specific are patterns on teaching and guiding students actually? For this Focus Group session, we decided on narrowing down the theme by choosing an exciting context: failing students. This is a high-stakes situation in which both the teacher and the student in a way are emotionally involved.
Over the years, the teaching context got (and still gets) more and more multicultural, due to an increasing number of bi-cultural and international students. Therefore, it seems to be a good idea to use the international context of the EuroPlop conference to investigate our ’default settings’. And, while doing so, get a glimpse of the intercultural aspects of the educational papers we write.
Unlike other conferences, EuroPLoP focuses on improving papers instead of only presenting them. Authors will receive high quality feedback in a constructive way during the shepherding process and during the workshops at the conference. The feedback can be used to improve submitted papers for final publication.
We hope to be able to organize the conference on-site, but this is subject to change, considering any restrictions imposed by Covid-19. EuroPLoP’s registration fee is normally 1111€ and covers all activities, accommodation, and food.