ACM COMPUTE 2025 Best Practices Track Proceedings

COMPUTE is an annual Indian conference supported by ACM India and iSIGCSE. The focus of COMPUTE is to improve the quality of computing education in the country by providing a platform for academicians and researchers to interact and share best practices in teaching, learning, and education in general. The Best Practices Track of COMPUTE 2025 invited Computer Science Educators across the country to submit an experience report for the best practices under multiple categories: 1) Novel classroom activities, 2) Imaginative assignments that promote creativity and problem-solving, 3) Diverse pedagogical approaches (e.g., flipped classrooms, peer teaching, project-based learning), 4) Designing AI-resistant or AI-integrated assessment questions, and 5) Teaching CS to students from other disciplines (e.g., business, humanities, engineering). These proceedings contain papers selected from these submissions for presentation at the conference, as well as a report (written by the editors) from the two best practices sessions where these were presented.

💡 Research Summary

The paper presents a comprehensive report on the Best Practices Ideas Session held at the ACM COMPUTE 2025 conference in Ropar, India. The Best Practices track was introduced to provide a less formal venue for computer science educators at the higher‑education level to share teaching innovations, discuss challenges, and foster a collaborative community. Sixteen submissions were received; three were selected as full Best Practices papers—“Functional Python Programming in Introductory Computer Science Courses,” “SocraticAI: Transforming LLMs into Guided CS Tutors Through Scaffolded Interaction,” and “Ancient Algorithms for Modern Curriculum.” Two additional ideas were discussed in the session and are the focus of this report.

The session was organized around three thematic areas.

1. Classroom Activities for Programming Courses – Participants advocated a shift from lecture‑centric instruction to a practice‑heavy model, suggesting roughly an 80 % practice to 20 % theory split. Concrete activities included pair programming, debugging workshops, Think‑Pair‑Share, code‑tracing exercises, algorithm‑focused games, mini‑hackathons, project‑based learning, and gamified visual tools. Frequent student‑instructor interaction, immediate feedback, and low‑stakes quizzes were emphasized to sustain engagement, especially given limited attention spans. The overarching message was to treat the classroom as an active laboratory for experimentation and collaboration rather than a passive content delivery venue.

2. AI‑Resistant Take‑Home Assignments – Recognizing the growing misuse of large language models for completing assignments, participants proposed redesigning assessments to focus on process, reasoning, and interaction rather than final code output. Strategies included: (a) minimizing traditional take‑home work in favor of in‑person activities such as group brainstorming, viva‑voce examinations, and endurance‑style tasks; (b) when take‑home work is necessary, requiring students to submit AI interaction logs, evaluate the quality of prompts, and provide reflective critiques; (c) personalizing assignments through contextualized, visual, or story‑embedded problems that are difficult for generic AI systems to solve without domain‑specific knowledge; (d) modular team projects with individual responsibility for sub‑modules, peer code review, and oral defense of design decisions. The consensus was not to ban AI outright but to promote transparent, critical, and guided use, thereby turning AI into a learning aid rather than a shortcut.

3. Strengthening the CS Education Community in India – Participants highlighted the need for robust, sustained networks of educators. Proposals included establishing regional champions, organizing regular small‑scale meet‑ups, industry‑interaction days, and practice‑sharing sessions across zones. A dedicated online platform was suggested for sharing pedagogical tools, classroom activities, and problem sets, complemented by a centralized repository of canonical CS problems (e.g., Tower of Hanoi, Traveling Salesperson) with instructional variations. Professional development was framed as continuous, practice‑oriented workshops combining conceptual grounding, hands‑on implementation, student‑led presentations, and feedback cycles. Existing structures such as ACM chapters, faculty development programs (FDPs), and online forums were identified as key enablers for long‑term capacity building.

The report concludes that the Best Practices track successfully generated a rich set of actionable ideas. iSIGCSE has committed to adopting these suggestions in year‑long activities, with the outcomes to be reported at the next COMPUTE conference. Participant feedback was overwhelmingly positive, praising the high level of interaction and the practicality of the proposals. The organizers plan to expand the track in future editions, encouraging more submissions and deeper empirical validation of the presented practices.