Author: Rebekah Esmaili, Science and Technology Corp. (STC) Presented at the UMD ESSIC Seminar Series on Aug 28, 2023
Scientists commonly write their code and algorithms in legacy languages such as IDL, which is computationally fast and geared toward scientific programming. However, maintenance can be expensive due to licensing costs, and languages may not be cloud-computing optimized. As a result, agencies like NASA and NOAA are increasingly interested in converting IDL software into modern languages like Python. Converting legacy IDL code to Python can be more productive using generative AI tools like ChatGPT. This seminar introduces how ChatGPT can simplify the transition and showcase practical code conversion from IDL to Python through a live demonstration. We’ll discuss some code migration strategies, best practices, and challenges. Join us to embrace the fusion of AI and scientific computing, revolutionizing legacy code modernization with ChatGPT.
There is an interactive example, click on the Binder following link and select examples.ipynb to launch the notebook. This will walk through how to convert the code found in sample_idl_code.pro The solutions are in solutions.ipynb.
