CS Paper Radar

Exploring multimodal AI--XR interactions for software development: From expert perspectives to semantic code outlines

Ilya Zakharov, Alexey Markin, Konstantin Ostroukhov, Alexey Perminov, Roham Koohestani, Ekaterina Koshchenko, Agnia Sergeyuk

cs.HC cs.SE ·submitted 2026-11-18 12:36 UTC · v1
survey

Abstract

AI tools can generate code quickly, but developers still need to understand and check the results. We explore how the enriched multimodal context derived from the extended reality (XR) workspace might help developers manage this information. We first interviewed 13 experts in AI, XR, and related fields to identify opportunities and design challenges (Study 1). These interviews informed two prototypes explored with 12 professional developers (Study 2): four tried a version where the code faded to become invisible after it was looked at, and eight tried a version that replaced viewed code sections with short descriptions of their purpose. In the second prototype, head direction helped determine when to offer a change, while manual input let developers apply it and reopen the original code. A separate open-ended demonstration with seven software-engineering researchers provided complementary feedback. Participants' accounts suggested that descriptions could provide an overview of unfamiliar code and serve as reminders when revisiting it. However, difficulty comparing descriptions with source code, headset limitations, and controls that deviates from regular IDE routines could introduce additional effort. These exploratory findings identify requirements for attention support: make summaries easy to recognize and access, preserve straightforward inspection of the original code, and retain familiar navigation and editing practices. The study contributes to human-centered multimodal systems by showing how the coordination of input channels and AI-generated representations must account for users' attention, verification needs, and established ways of working.

Comment
Submitted to HIMAS 2027: Human-Centered Interactive Multimodal AI Systems, co-located with ACM IUI 2027

Closest papers

Nearest neighbours by meaning across the whole index.

Cognitive Inclusivity in AI Coding Assistants: Evidence from GitHub Copilot

cs.SE ·2026-10-30 ·Grace Kerr, Aarushi Parasrampuria, Faith Culas, Kelly Blincoe
similarity 0.68
Abstract

AI coding assistants are commonly used for daily software development tasks. While prior research has examined their impact on productivity and code quality, less is understood about how these tools support diverse ways of thinking. Evidence from non-AI software engineering tools suggests that mismatches between tool design and users' cognitive styles can introduce inclusivity barriers, but it remains unclear whether similar issues arise in AI coding assistants. We investigate how developers with different cognitive styles experience and interact with these tools and whether they exhibit cognitive inclusivity challenges. We conducted a controlled laboratory study with 20 student participants using GitHub Copilot to complete code generation, debugging, and refactoring tasks. Using think-aloud protocols, semi-structured interviews, and inductive thematic analysis, we identify patterns in user interaction. We then apply the GenderMag framework to examine how these patterns differ across cognitive styles. We identify five inclusivity issues concerning how Copilot explains suggestions, supports validation, facilitates learning, exposes its features, and influences the problem-solving process. Our findings suggest that common interaction paradigms in AI coding assistants can privilege certain cognitive styles while introducing additional cognitive effort for others. We derive design implications for more inclusive AI coding assistants.

0.52

From Tokens to Structure: Investigating Visual Attention in Progressive Code Rendering

cs.HC ·2026-09-21 ·Xiaotian Su, Jan Brasser, David Robert Reich, Lena A. Jäger, April Yi Wang
datasetposition similarity 0.66
Abstract

AI coding interfaces present generated code either all at once or token-by-token. These rendering strategies reflect model generation rather than how programmers actually read code: selectively, non-linearly, and guided by the structure. We argue that code rendering is a first-class interaction primitive that shapes how programmers read and understand code. To explore this design space, we introduce structured rendering, a technique that reveals code in semantically meaningful chunks derived from its syntactic hierarchy, exposing high-level structure before low-level details. To isolate rendering effects on visual attention, we conducted an eye-tracking study with 53 participants comparing static, character-based, and structured rendering. Our findings show that rendering alters visual attention and reading behavior: dynamic rendering induces fewer but longer fixations and more sustained focus, while structured rendering further guides attention toward semantically meaningful units and supports high-level understanding. We release the anonymized dataset with an interactive demo (https://codegaze.vercel.app/) to support future research.

0.25

Beyond the Prompt: Linking What Developers Ask, Do, and Understand with Coding Agents

cs.SE ·2026-09-27 ·Yunhan Qiao, Summit Haque, Christopher Hundhausen
similarity 0.66
Abstract

Coding agents can now change code for developers, who describe goals, supply context, and respond to the agent's work. Yet prompts, screen activity, and task success each tell only part of this story. We present Say, Do, Understand, an end-to-end workflow for analyzing what developers write to an agent, what they do while it works, and what they can explain afterwards. The workflow has five stages (Capture, Prepare, Analyze, Integrate, and Interpret) and three instruments: a prompt codebook, a scheme for coding screen-recorded activities and events, and separate rubrics for explaining the process and the solution. We applied it in an observational study of ten experienced developers who used GitHub Copilot on an unfamiliar codebase. Crossing task performance with understanding produced four personas. The two measures agreed for eight developers but split for two: one passed most tests but could not explain the solution, and another passed few tests but explained it well. In this sample, the personas that most often asked the agent to check its work spent the least time testing on their own. These patterns are descriptive and do not generalize beyond the sample. We recommend the workflow to computing educators, industry practitioners, and researchers to adapt and evaluate human--AI communication in software engineering.

0.19

The Comprehension Gap in LLM-Assisted Software Development: An Experiment

cs.SE ·2026-11-18 ·Julian Oertel, Jil Klünder, Regina Hebig
similarity 0.65
Abstract

When LLM tools are pushed beyond their capabilities for software development tasks, developers must be able to step in. However, the comprehension that developers acquire during development might be negatively impacted due to excessive code generation. This could create a cognitive debt that needs to be resolved before tasks can be solved efficiently again. We explore how well developers understand code they generated by prompting in comparison to self-written and unseen code and how they reflect on their comprehension. We conduct a crossover experiment and interviews with 35 participants consisting of students, academic personnel, and software developers. After completing two programming tasks with or without LLM chat usage, respectively, participants are tasked with finding introduced mutations in the produced and additional unseen code. We use insights from interviews and screen recordings to supplement our results. Regarding the experiment, 36% of participants report a better understanding of self-written code compared to their generated code. On average, participants solved the comprehension task 20% faster for self-written code than for generated code and found 11% more mutations. Participants needed significantly less time to solve programming tasks when working with the LLM. However, participants with high reliance on the LLM chat and low task engagement showed significantly decreased comprehension. Our qualitative results show great variability in individual perspectives on comprehension differences and strategies to mitigate issues. Strategies for dealing with comprehension differences range from accepting comprehension differences to adapting their LLM interactions or avoiding the tools altogether. Our results warrant further investigation into cognitive debt in larger systems and developer practices to mitigate issues.

0.11

Spec-Driven Development as Prescribed and as Practiced

cs.SE cs.AI ·2026-08-25 ·Shyam Agarwal, Anmol Singhal, Andrew Macvean, Bogdan Vasilescu, Travis Breaux
highly cited similarity 0.65
Abstract

AI coding agents implement software from naturallanguage instructions, but the intent behind that work is lost when stated only in chat sessions. Spec-driven development (SDD) addresses this by having developers and agents write specifications that guide implementation, yet evidence on its use comes largely from tool vendors. We study SDD in practice through a thematic analysis of Reddit discussions, interviews with 21 practitioners, including eight tool makers, and 266 active and popular repositories in SpecMine, a corpus of 470,795 spec files from 73,030 public GitHub repositories that we release. Practitioners agreed that specs should capture intent, but many also included implementation details, used SDD selectively and mainly for new features, and reported more reliable agent behavior alongside greater planning and review costs. Our findings show that implementation detail may help agents while making specs prone to decay, that tool-specific record structures shape what people and agents see, and that participants described spec-code checks mainly when changes are completed rather than over time. We also find that individual benefits may come from clarifying intent before delegation, while team benefits may depend on SDD becoming a shared process. We derive implications for tool builders, practitioners, and researchers.

0.36

Vibe-GUIDE: A Graph-based User Interface in IDEs for Oversight in Vibe Coding

cs.HC ·2026-09-20 ·Chifang Chou, Sam Yu-Te Lee, Rudrajit Choudhuri, Kwan-Liu Ma
similarity 0.64
Abstract · 14 pages, 3 figures, 4 tables

In agentic coding, developers shift from implementing changes themselves to specifying intent, evaluating the agent's work, and making approval decisions. However, delegating implementation can introduce cognitive debt that erodes project comprehension over time, constraining developers' ability to provide oversight. In this work, we investigate the role of persistent shared representations in supporting project comprehension and oversight of coding agents. We present Vibe-GUIDE, an agentic coding interface built around a structural, live, manipulable, and adaptive graph representation organized by functional modules. We evaluated our interface in a randomized between-subjects study comparing how 16 developers completed three cumulative coding tasks using our interface or a Chat-only baseline. We found that Vibe-GUIDE can support project comprehension and sustained task performance while keeping developers cognitively involved in oversight. These findings show how persistent shared representations can complement natural-language interaction and help developers maintain the understanding needed to oversee agent-generated changes as projects evolve.

0.42

SWE-Journey: Towards More Realistic Evaluation of Coding Assistants through Long-Horizon, Multi-Turn Interaction

cs.CL cs.AIcs.MAcs.SE ·2026-10-08 ·Hexuan Deng, Yue Wang, Wenyu Jiang, Cheng Yang, Haolin Yang +11 more
benchmarkhot direction similarity 0.63
Abstract

Coding assistants such as Claude Code and Codex have become a major application of LLM agents, yet existing benchmarks remain far from real-world use, particularly in task horizon and interaction length. Code assistants require completing long chains of development work in continuously evolving repositories, while repeatedly clarifying requirements and adapting implementations through multi-turn interaction. To address these gaps, we introduce SWE-Journey, a benchmark for more realistic evaluation of coding assistants. To address the task-horizon gap, we propose a weak-to-strong synthesis pipeline that automatically constructs long-horizon coding tasks. To address the interaction gap, we mine four representative user personas from real interaction data and build a user-simulation agent to reproduce realistic code-assistance interactions. On average, models pass over 75% of tests for requested functionality with software architects, but fewer than 25% with non-coders. These results show that current coding assistants still fall short of enabling reliable coding for non-coders. We further analyze the reasons for this gap and identify asking right, finding right, and fixing right as key capabilities during interaction.

0.23

Omni-Workspace: Benchmarking Omni-Modal AI Assistants on Multimodal Workspace Tasks

cs.CL ·2026-11-14 ·Yuhao Wang, Yucheng Zhao, Yiran Zhong, Feida Zhu, Hongcheng Liu +3 more
benchmark similarity 0.63
Abstract

Omni-modal agents are expected to move beyond multimodal understanding toward completing complex tasks in realistic multimodal environments. In these settings, information is distributed across heterogeneous files and modalities, with cross-file and cross-modal dependencies requiring agents to continually coordinate perception, reasoning, and tool use as newly gathered evidence informs subsequent actions. However, existing omni-modal benchmarks primarily focus on multimodal perception and understanding, leaving complex task completion through sustained interaction with multimodal workspaces underexplored. We introduce Omni-Workspace, a benchmark comprising 380 deep-research-style tasks across four multimodal workspace types and two difficulty levels. These tasks require agents to inspect and process workspace files, retrieve external information when needed, integrate evidence across files and modalities, and produce artifacts satisfying explicit requirements. Evaluating 18 representative models reveals limitations in long-horizon multimodal workspace execution, with the best model scoring only 65.93 on average. Further analysis identifies multimodal context inflation and task-state drift as recurring challenges. Motivated by these findings, we further propose OmniFocus, a lightweight approach combining multimodal working-memory management with explicit task-state tracking to improve performance and efficiency. Omni-Workspace advances the evaluation and development of omni-modal agents toward autonomous completion of complex tasks in multimodal workspaces.

0.10