Guiding principles
- Think Before You Type – Don’t just jump into AI-generated responses. Understand the problem first.
- Solve Conceptually First – Before writing any code, define the problem clearly in your own words.
- Use Pseudocode – Outline the logical steps needed to solve the problem. This helps in structuring the AI’s response more effectively.
- Break Problems into Smaller Parts – Instead of asking one massive, vague question, split it into well-defined sub-problems.
- Be Specific and Structured – A vague prompt leads to generic output. Instead of “Write a Python script for XYZ,” describe the requirements, expected inputs/outputs, constraints, and any special handling needed.
- Provide Context – If applicable, mention the programming language, libraries, or constraints to ensure a more relevant response.
- Refine and Iterate – AI-generated responses aren’t always perfect. Use them as a starting point, refine your prompt, and ask follow-up questions to improve accuracy.
- Understand the Output – If you don’t understand the AI’s response, revisit your original problem definition. AI is a tool to assist thinking, not replace it.
- Focus on Learning, Not Just Answers – The goal is not just to get code that works but to understand how and why it works.
References
- Mnguni, N. M., Nkomo, N., Maguraushe, K., & Mutanga, M. B. (2024). An Experimental Study of The Efficacy of Prompting Strategies In Guiding ChatGPT for A Computer Programming Task. Journal of Information Systems and Informatics, 6(3), 1346-1359. DOI: 10.51519/journalisi.v6i3.783
- Bucaioni, A., Ekedahl, H., Helander, V., & Nguyen, P. T. (2024). Programming with ChatGPT: How far can we go?. Machine Learning with Applications, 15, 100526. DOI: https://doi.org/10.1016/j.mlwa.2024.100526
- Tian, H., Lu, W., Li, T. O., Tang, X., Cheung, S. C., Klein, J., & Bissyandé, T. F. (2023). Is ChatGPT the ultimate programming assistant--how far is it?. arXiv preprint arXiv:2304.11938. DOI: https://doi.org/10.48550/arXiv.2304.11938
- Cao, J., Li, M., Wen, M., & Cheung, S. C. (2023). A study on prompt design, advantages and limitations of chatgpt for deep learning program repair. arXiv preprint arXiv:2304.08191. DOI: https://doi.org/10.48550/arXiv.2304.08191