BoilerSketch - 1 Aug 2026 - arXiv:2608.00844v1

Oct 26, 2026

Upload by Angel

BoilerSketch is a TA-supervised, diagram-first GenAI practice and tablet interface designed for CS1 and early CS2 support. It reduces course support bottlenecks by pairing chat with a pen-enabled whiteboard while constraining the model to produce structured, renderable Mermaid diagrams. The system enforces academic-integrity guardrails by forbidding executable code and code-level debugging, using human-in-the-loop supervision for correction and escalation. A 45-minute evaluation with 21 instructional staff found the tool broadly helpful for conceptual understanding and routine diagram-based explanations.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

I. INTRODUCTION

Support bottlenecks in CS1 and early CS2

Program visualization and learning objectives

GenAI for computing education support

Gap: diagram-first, TA-supervised help

II. BOILERSKETCH APPROACH

BoilerSketch: A TA-Supervised, Diagram-First GenAI Practice for Structured Diagrams in CS1/Early CS2 Ethan Dickey∗, Vivan Tiwari∗, Anvit Sinha∗, and Andres Bejarano∗ ∗ Department of Computer Science Purdue University West Lafayette, IN 47907 Abstract—This innovative practice full paper presents BoilerSketch, a TA-supervised, diagram-first GenAI practice and tablet interface for providing structured visual explanations in CS1 and early CS2 support settings. Large early computing courses routinely face a support bottleneck during labs and office hours because many student questions are best answered with a diagram rather than additional text, yet most AI tutoring tools remain text-forward and unreliable at producing accurate, pedagogically useful visuals. BoilerSketch addresses this gap through a dual-pane interaction model that combines chat with a pen-enabled whiteboard for student sketches and a prompting strategy that constrains the model to generate structured, renderable Mermaid diagrams rather than free-form images. To preserve academic integrity, the system is intentionally scoped to conceptual explanation: it forbids executable code and code-level debugging and uses a human-in-the-loop workflow in which teaching assistants remain accountable supervisors who can monitor sessions and intervene when responses require correction, deeper probing, or escalation to live help. We report a 45-minute expert evaluation with 21 instructional staff from a large programming course who used BoilerSketch on representative questions and completed a post-use survey. Twothirds rated the system at least moderately helpful for conceptual understanding and at least moderately useful for typical support tasks. Staff saw the strongest value in routine diagram-based [ cs .CY] 1 Aug 2026 arXiv :2608 .00844v1 explanations and noted limits in diagram depth and applicability to more advanced topics. We conclude with practical guidance for adopting supervised, diagram-first GenAI support in early computing courses, emphasizing scope-and-escalation rules, promptas-policy guardrails, and reliable structured diagram rendering. Index Terms—Computing education, generative artificial intelligence, human-in-the-loop, program visualization, teaching assistants I. INTRODUCTION Large CS1 and early CS2 courses routinely face a support bottleneck during labs, recitations, and office hours. Students need timely, individualized help, yet many of their questions are not well served by short textual responses alone. Questions about recursion traces, tree traversals, pointer aliasing, and dynamic programming tables are often easiest to answer with a diagram that makes runtime state or structural relationships visible. At the same time, help-seeking in large courses is difficult to manage, and teaching assistants (TAs) are frequently responsible for absorbing much of this demand [1]–[4] . Computing education research has long argued that learning to program requires students to form an explicit model of how programs execute. Sorva characterizes this challenge in terms of the notional machine, arguing that understanding runtime behavior is itself a core learning objective in introductory programming [5] . The program-visualization literature complements this view by showing that diagrams and animations can support novice reasoning about state change, control flow, and data relationships, but that their educational value depends heavily on learner engagement and pedagogical integration [6], [7] . More recent work continues to show that novices use program visualizations to understand unfamiliar code and to verify their assumptions during comprehension tasks [8] . This literature suggests that visual explanation is often central to how students build and repair their mental models. GenAI appears to offer a new way to expand access to individualized support. Recent computing education work has shown that large language models can generate

FAQs

What problem does BoilerSketch target in large CS1 and early CS2 courses?

It targets support bottlenecks during labs and office hours where many student questions are best answered with diagrams rather than text alone.

How does BoilerSketch generate visual explanations, and what format does it use?

It uses a dual-pane interaction model with chat plus a pen-enabled whiteboard, and a prompting strategy that constrains outputs to structured, renderable Mermaid diagrams.

What academic-integrity and scope restrictions does BoilerSketch enforce?

The system forbids executable code and code-level debugging, and it relies on human-in-the-loop workflows where teaching assistants monitor and intervene when needed.

Oct 26, 2026

Upload by Angel