Decomp of David Horner's Drain Storm (1997) (github.com)

A passionate decompilation effort has successfully recompiled the classic Windows 3.11 game "Drain Storm," leveraging advanced vibeslopping techniques and various GPT models to enable it to run seamlessly on modern systems. The process involved multiple passes: the initial attempt resulted in a non-functional version, but subsequent iterations improved the game's playability while revealing significant differences from the original code. A thorough examination of all decompiled functions was conducted to reconcile these differences, although only the second pass underwent extensive testing. This project holds substantial significance for the AI/ML community, demonstrating the potential of generative models to assist in complex code decompilation and software preservation. By eventually refactoring the code for maintainability and potentially expanding support to Linux and macOS, the project could foster modding and community engagement with classic games. Additionally, it highlights the importance of preserving digital heritage, prompting discussions about copyright and creator rights, as the original creator, David Horner, remains a crucial but uncertain figure in this narrative.