Are todays LLM not capable of translating that stuff into modern JS&Co?
If the LLM can’t run the original code, it can’t compare to verify the translation is accurate
If the LLM can’t run the original code, it can’t compare to verify the translation is accurate