Exploring the ChatGPT Approach for Bidirectional Traceability Problem between Design Models and Code
Authors:
Hideyuki Kanuka,
Genta Koreki,
Ryo Soga,
Kazu Nishikawa
Abstract:
This study explores the capabilities of Large Language Models, particularly OpenAI's ChatGPT, in addressing the challenges associated with software modeling, explicitly focusing on the bidirectional traceability problem between design models and code. The objective of this study is to demonstrate the proficiency of ChatGPT in understanding and integrating specific requirements into design models a…
▽ More
This study explores the capabilities of Large Language Models, particularly OpenAI's ChatGPT, in addressing the challenges associated with software modeling, explicitly focusing on the bidirectional traceability problem between design models and code. The objective of this study is to demonstrate the proficiency of ChatGPT in understanding and integrating specific requirements into design models and code. We also explore its potential to offer solutions to the bidirectional traceability problem through a case study. The findings indicate that ChatGPT is capable of generating design models and code from natural language requirements, thereby bridging the gap between these requirements and software modeling. Despite its limitations in suggesting a specific method to resolve the problem using ChatGPT itself, it exhibited the capacity to provide corrections to be consistent between design models and code. As a result, the study concludes that achieving bidirectional traceability between design models and code is feasible using ChatGPT.
△ Less
Submitted 10 October, 2023; v1 submitted 26 September, 2023;
originally announced September 2023.
